Controversial New Strategy Causes Havoc In Trackmania!

Go to drinkag1.com/k... to get a FREE 1-year supply of Vitamin D3+K2 and 5 AG1 travel packs with your first purchase. Thanks to AG1 for sponsoring today's video!
Huge thanks to ‪@JstAnothrVirtuoso‬ for helping me learn about low input strategies.
Check out ‪@bettaTM‬'s video: • This Strat Revolutioni...
Check out ‪@Wirtual‬'s video: • World Record History o...
Please follow me on these platforms, it really helps!
Patreon: / karljobst
Twitter: / karljobstgaming
Twitch: / karljobst
Discord: / discord

[@karljobst]

25 days until trial! Huge thanks to AG1 for the support. Go to drinkag1.com/karljobst to get a FREE 1-year supply of Vitamin D3+K2 and 5 AG1 travel packs with your first purchase. Thanks to AG1 for sponsoring today's video!

[@brandonboyce6408]

looks just like mario cart now lol should probably have different leaderboards

[@lotus_flower2001]

Sponsored by Vitamins lol. If it involves a computer's help than its cheating imo.

[@lorenzomartinez8543]

I believe some tricks in Marios Bros. and Punch Out were discovered using similar brute force methods.

[@Rainos62]

I only agree with this because they are creating a separate leader board for it. if something can become competitive let it many communities adapted to weird and unusual strategies but I think making another leaderboard is the right decision to preserve the fun factor

[@IsaRat89]

I would say "good luck" in court, but I feel "good luck, have fun" fits better.

[@zeikjt]

Can't wait to see Medium Input Strategy take off, where players keep memorizing just enough inputs to not be on the LIS leaderboard

[@Fay7666]

I mean, the Uberbug record would be valid as long as you mash during the last second, right?

[@Darkabyss290]

@@Fay7666 Highly unlikely, since all those would be logged into the replay they'd be able to tell you clearly mashed right at the end to try and circumvent the rules, and imo if you do that you should be banned from the leaderboards entirely.

[@NYKevin100]

The obvious solution is, instead of having an LIS leaderboard and a "main" leaderboard, have a keyboard leaderboard and a joystick leaderboard, and just don't have any LIS-related rules whatsoever. LIS is keyboard-only, so all the serious players can just keep playing joystick and ignore whatever funny business goes on in keyboard land.

[@Thezanlynxer]

@@NYKevin100 that just excludes people who want to play normally with keyboard for no real reason. sure most people seriously competing play with joystick, but not everyone wants to go for world records anyway

[@tremen151]

@@NYKevin100issue is, there are pro players who do play on keyboard. though it is a disadvantage, it’s still possible to set world records with it, and it has been done on campaign tracks.

[@Veylon]

Any community that's about who can do X the best gradually strips away all the human parts of it until X is the most sterile, abstract version of itself bereft of the context it once had. It's a fascinating process.

[@charlesstoll744]

And that is what will keep happening, regardless of arbitrary rules imposed by the score keepers. You need at least 8 inputs? Ok, then lets optimize for 8-input solutions. I think that this should be viewed as a track design flaw rather than a player issue. If you design a track that can use less than (some reasonable value... say) 30 inputs to beat, then you made a bad track. Stop making crap tracks with shortcuts (something that is already generally disliked in cotd) and stop making short, simple tracks unless you are aware and ok with people "cheesing" the track like this. I think the community decision as Karl describes it is just going to lead to an elongation of the process of fixing this issue.

[@sven_bender]

At the highest level, most competitive endeavours just turn into people abusing the meta in order to gain that 1% advantage over your opponent

[@Bob13454]

@@charlesstoll744 It seems very easy to tell if someone is using LIS, and basically impossible to disguise it.

[@Rofel_Wodring]

based profile picture

[@lassikinnunen]

​@@charlesstoll744Yeah a lot of games are "broken" like that. Many board games too. Tic tac toe is a simplest example of such a broken "game", that becomes more interesting if you instead make unlimited size grid and require 4 in a row or 5 to win. That popular bar game where you drop plastic thingys into a vertical grid is broken too, the one who starts has control over possible outcomes.

[@ShigaonTM]

as a tm player with 3200+ hours i don't think LIS is a real issue. this type of thing is almost inevitable for those "short, so precise it's practically rng" tracks where hunting them basically constitutes trying to do the same thing over and over until you get lucky. that way playing it normally is not much different than going for LIS after enough time where it's no longer trying to make the car drive a certain way and becomes "i think i need to start steering riiiiight now". at that point you're not playing the game normally either. also tmnf official hunting or time attack (time attack is more or less the speedrunning mode in tm) in general isn't the most competitive part of trackmania. rounds based competition on tracks with high skill expression are. that way way it tests consistency and actual skill (in contrast to some of the official maps like A12 that mostly test how lucky you are and how much time you put into hunting). most competitions and specifically the most prestigious competitions are based on some kind of rounds format where time attack only ever plays a role for seeding, if even that.

[@The_Omegaman]

I disagree. Sub sections of long tracks could be optimized and be the edge someone needs. Id ban keyboards. Need natural inputs.

[@sorrowandsufferin924]

Trackmania: a racing game. you: not racing is a non-issue. A driving instructor sits next to you in the car and can perform driving inputs. But unless they do, they're not driving - they're a passenger. LIS makes you a passenger 99.99% of the time - every single micro-second except for those 6 where you're actually doing inputs. How is that a non-issue in a racing game? Make it make sense.

[@davbou]

@@The_Omegaman The only section of a long track that can be realistically LISed is the begining because that's the only place that you can know the exact position/angle/speed(0) of the car and LIS of over 15 to 20 inputs are probably impossible for a human to pull off anyways so the issue only realistically affect really short tracks

[@ThatWhichObserves]

It's at the most base level, a tool assisted run.. You do 0.001% of the work.. Inputs. The computer generates hundreds.. thousands. hundreds of thousands of input runs.. and spits out a 'success'.. This is a tool advantage.. It is as close to generative AI making a speedrun route, as it is to making a TAS.

[@mcplaya5805]

I have my fair share of competing in Trackmania. I have competed in TTCs, ESL Cups, TMM TeamCups, TMO, and a couple of Canyon team events. LIS strategies are never going to be in use for these competitions due to the complexities of the maps, thus that part of the argument is invalid. The maps which are shorter records are the most in danger. These maps have extremely competitive leaderboards and have players dedicated to them as much so as a player would dedicate to a tournament map pack. In fact, those players are dedicated even more to getting the record on that map. These are the type of players which are effected most. I sit here thinking of Trinity and Hefest just to name two of many A12 players that have dedicated in the upper hundreds even into the thousands of hours into the map. Another thought that comes to my mind about rounds based competitions is that the player base that competes in those competitions is extremely small when compared to the Nadeo hunting community. The Nadeo hunting community is extremely dedicated to their craft and individual maps, so Nadeo maps will have more people competing for those records than there are competing in rounds based competitions. TL;DR Hunting short maps is often more competitive than a rounds based competition due to player base size and individual commitment.

[@TheOneThreeSeven]

Its a super interesting philosophical question. If a track can be beaten in 6 frame perfect inputs then it reveals an underlying problem with the track. Instead of seeing this as a form of cheating, the researchers who discovered all of this should get credit for achieving a novel kind of WR and celebrated.

[@robotm99]

yup-you get it

[@Chalo122790]

100% this, like if a track can be beat with less than 8 inputs is not a track, thats just luck really in the end.

[@Leaky_Spigot]

It’s the method, not the outcome. ‘Traditional’ runners would inevitably reach the same conclusion as LIS (theoretically), but they have to put in an immeasurable amount of combined player hours to reach that kind of optimization. Even TASers spend ridiculous amounts of combined player times labbing optimizations. LIS completely eliminates that effort. It _is_ an achievement worth celebrating, but it’s not a design flaw, and the people opposed to LIS aren’t ‘wrong’ for being opposed to it. It’s an effective tool, but that comes at the cost of ideals many runners value in this hobby; shared community efforts, competitions that allow for personal expression, hard-earned achievements, ground-breaking discoveries (that aren’t the ultimate solution), personal growth, etc. LIS can share those things, sure, but it’s inevitably going to be a distilled version, or one that is completely incomparable to its predecessor. There is no comparison between the efforts of hundreds/thousands of people over time spans measured in years, to that of a computer program that simulates that whole experience in just hours, but without the human elements. Worse still, its existence will expedite the end of the ‘traditional’ experience many people are fond of.

[@bobbymcgraw2168]

No idea what you're smoking, but it's very irrelevant.

[@nomore6167]

"...the researchers who discovered all of this should get credit for achieving a novel kind of WR and celebrated" - I guarantee that if someone made a mistake and accidentally stumbled on this, rather than intentionally seeking it via TAS, then that is exactly what would have happened, and at that point, every other speedrunner would be trying to replicate the few-input strategy, and everyone would be happy.

[@EmeralBookwise]

The problem I see is that over a long enough period of time LIS is sorta what all speed runs in a deterministic game become. Using a computer to brute force the solution is just speeding up the process. Banning the use of visualization tools and requiring players to use only the onscreen information from the actual game is all well and good, but I don't see how anything more than that is enforceable in the long run.

[@EzmarVoD]

In most cases, the Low Input Strategy is not the optimal strategy, it's just the most optimal strategy with low input count, so as to be more RTA viable. It's kind of like the difference between blindfolded and regular strats for any game. The sighted strats are faster, and the blindfolded strats sacrifice some amount of speed in favor of consistency and repeatability. Aside from that, the difference between mechanically perfect inputs and human inputs is FAR different than people seem to think. Even given thousands of years, it's incredibly unlikely that a human player will approach a TAS generated set of inputs, simply because humans react to and play the game entirely differently. The time may approach the same limit, but the inputs will be very different.

[@avpguy11]

Players using an LIS strat would just use visual cues from the game rather than a timer. So player who grinded a level for 1000 hours and has perfect timing is legit, but a player who used a computer to find the optimal solution and has perfect timing is a cheater? This whole issue seems kind of silly.

[@EmeralBookwise]

@@avpguy11: is a player copying a strategy they got from a computer really any less "legit" than a player copying another speed runner? Either way they are not the one who put in the thousands of hours to develop the strategy, but copying an already established player is still how most new speed runners get their start.

[@sorrowandsufferin924]

@@avpguy11 Isn't that how it works everytime? The legitness comes from the time put into the task. Anything else is half-assing it and generally looked down on.

[@sorrowandsufferin924]

@@EmeralBookwise Here's the difference: Someone who's copying another player spends most of the time looking at the game. Someone who's copying LIS spends most of their time looking at a list of inputs. That's the difference.

[@28mancko]

Pastagrows here, thank you for covering this topic. I would like to provide some additional info. Watching the timer is the worst thing you could do while doing LIS. Players find and memorize visual cues to help them execute the inputs correctly, such as waiting for a certain block to enter your field of vision. On top of that, it is a good practice to memorize the certain outcomes of missed inputs that help you adjust your timings. 6+ inputs strategies can take a lot of time to execute. My A12 run took me around 70 hours. It is not as easy as it may seem, it mainly depends on the inputs layout. Players competing to get the same time is not entirely correct. Most of the LIS requires additional inputs that can improve or ruin your run. On harder maps, getting most optimal time or even finishing with just those frame perfect inputs, without any additional "manual" inputs is rare. Additionally, players do not have to focus on one, already found solution. Players can come out with their own strategies and execute them. If you want to beat the low inputs strategy time, you often need to add additional inputs which exponentially increase the difficulty to execute them. Finding new solution and pushing the limits is the spirit of low inputs strategies.

[@Killerabbet]

This is exactly the reason it should be it's own category. It requires a different skill set to execute, but often will completely snuff out the records using the "intended" skills to beat the track. LIS is intriguing and should be recognized and competed in by those who wish to do so, hopefully the mods will create a leaderboard and ruleset for LIS running sooner rather than later.

[@Ortorin]

"Players can come out with their own strategies" No... the computer comes up with the strategies. You just picked one to use.

[@AndrewTaylorNintyuk]

This reminds me, Did anyone Beat Deep Dip 2 yet?

[@bb010g]

​@@AndrewTaylorNintyukyes.

[@bertschi7]

@28mancko 1. This is a TAS, it doesn't make a difference if you, or the program presses the inputs, especially since a program was used to get the exact inputs to the run. 2. Go ahead, try to beat a TAS Uberbug run that beat Hefest, i want to see you try.

[@TheMegaOne1000]

I think this is a sign that the game is slowly getting optimized to the point where there are no more world records to be made. It's not good, or bad. It's inevitable.

[@fiveiron23]

Loads of people have said the same about records that have since been beaten many times over in every imaginable game and format. The questions is "what is the purpose of trackmania record competition?" Is it to see who can use the most advanced tools to prescribe inputs to achieve the fastest possible time? Or Is it a forum for exhibition of personal skill? This new software makes those mutually exclusive. It cannot be both when the method of achieving a WR is to run an analysis on a map and repeat a short set of prescribed inputs. Why not just 1 input which turns on a TAS? Whats the difference? the skill is in the software at that point.

[@meraldlag4336]

@@fiveiron23 what if AI shows us new racing lines and we copy that? Is that legitimate?

[@paulamblard3836]

@@meraldlag4336 yes. and it already append.

[@nomore6167]

@@fiveiron23 "The questions is 'what is the purpose of trackmania record competition?'" - The answer is almost always the same: ego. People are unhappy that a faster path is possible using a strategy which is achievable by a greater number of people, which means their runs may be seen by fewer people, so their egos won't be stroked as much by fawning fans. Anyone who is hunting for the fastest time by competing against themself (rather than for ego stroking) shouldn't have a problem with this.

[@braapybobby]

​@@meraldlag4336That's not feasible, otherwise in other games people would just copy computer speed runs lol

[@swilson42]

This is just silly. Every speed run ever is exactly this, just on a larger scale. How many times has Karl’s own videos told us about a player needing to make “frame perfect” inputs to achieve a particular skip or clip? Just because this one track only needs 7 frame perfect inputs doesn’t make it any different than any other map or game where the human is making hundreds of precise inputs. Every single speed run starts life as an organic artistic subjective human experience. But as it matures and “perfect” strategies for runs are found, every human slowly turns into a robot and simply replicates those strategies. What begins as very analog, turns into a digital robot performance by the end of every one of them. And a human acting like a robot for long periods of time IS HARD and deserves to be recognized. That is the skill. Becoming a robot and staying a robot for minutes or hours. This one track (or several, I don’t know) only requiring robot mode for seconds only means that track is an easy track. It says nothing about the player or strategy. Just like the duel level, remove the track if it’s too easy, but this new way of describing strategy is just that, a new way to describe it. The actions themselves are exactly the same as they have always been.

[@Verlisify]

Many speedruns even have buffer strats to achieve this perfection that are found through TASing then become humanly possible. Doesn't make sense its bad in just Trackmania

[@JayTheYggdrasil]

I think an argument against this would be that the strategy is generated entirely by a computer. Once you have the tool, it doesn't require any further innovation or creativity to take some tracks and turn them into a sequence of relatively simple inputs. If a human can't compete without this strategy & the strategy is relatively easy to reproduce, the track becomes pointless in the same way the "Duel" level was pointless. Maybe you can compete by going off script and innovating, but I see no reason for that to need to play out on the main leaderboards when it could kinda ruin the existing competitiveness of some maps with minimal human input (pun intended). *I don't even play the game lol*

[@vasheal]

This is far from anything close to a "strategy"

[@Parker--]

@@JayTheYggdrasil Well if this is the problem, that the solution was brute forced by AI to find the answer, I've got some heartbreaking news for ya; this is exactly what can and will happen to ALL speedruns.

[@artsenor254]

I think the main concern is that every track might end up becoming too easy. You can remove one track (or not remove it, by the way, just let people decide that it's not really worth considering, as it often happens when a level is kinda solved in any game), but if every track gets beaten by a LIS strat at some point, you won't be able to remove the whole game and will end up with a ton of record holders, making it a pretty much pointless category to run as a whole (as any% in Civilization V, for instance). I don't know about LIS strats enough to determine how likely it is that it might happen, but I guess this is the concern here. Also, the argument about it being "against the spirit of the game", despite looking very flawed, shouldn't be disregarded either ; this is the reason why we make glitchless% categories and this kind of stuff. Obviously, you cannot decide of an objectively correct way to play any game, but people watching speedruns will have their own expectations, nobody watches only for the competitive aspect (otherwise, they would indifferently watch speedruns on any game) ; there are things they like in the game and they might think are "missing" in what they're watching, sometimes without evident reasons. If an AI told me that you could beat Ocarina of Time by glitching through a wall and mashing A button with the right tempo for ten minutes to get to a secret end, I would probably lose interest on the category as well, despite this kind of things being a valid strategy in many other speedruns (bit of an extreme example, but still). Trying to determine who's right or wrong on the matter would only lead to never-ending arguments, so we just take people's opinion on the topic, evaluate the trends and decide what's worth addressing or not.

[@binaryrooster]

When a fast paced racer turns into Dragons Lair.

[@DeeDee-pw9pm]

I couldn't have said it better.

[@Dukephillips88]

An excellent point

[@dondom4154]

But Dragon's Lair was fun as hell! Wait...

[@syaoranli7869]

What is the context of this? Sorry I'm not familiar with the lore and why speed is controversial in Dragons Lair

[@PsychoNinjaFlea]

So like Road Blaster (1985) it's a Dragon's Lair style game but fast-paced racing instead

[@NastyMick]

10:47 A hidden monitor? Who would ever do such a thing? ( ͡° ͜ʖ ͡°)

[@goodguyrobert]

Queen pwnzalot has entered the chat

[@CB-L]

*_Karl Juggling 'Blindfolded'_* immediately comes to mind and I can't forget that scene ( ͡° ͜ʖ ͡°)

[@NickGreyden]

I got that reference!

[@Lu-db1uf]

lol did you just out yourself as a cheater

[@DarkSerris]

As long as you have a physician vouching for you ...

[@CrayonEater9845]

LIS isn't exactly novel. People use TAS all the time to find optimal routes and then try to reproduce those routes. LIS has the added consideration to make the route easier to reproduce, akin to speedrunners opting for marginally slower tricks simply because they are easier to reproduce. I think the only legitimate argument is spirit of competition, but how far do you take it? In Mario 64 do you always have to take the most optimal route/trick simply because it's faster, even if you are already ahead and on pace to set a record with slower tricks. Ultimately this just leads to the optimal tricks being required to set a new record and if LIS is optimal, that's what everyone will do. Ultimately it's up to high input strats to perform better and be required to set records

[@darkjapan]

I agree but I think the issue is people are not finding the optional route themselves, they are making the tool find the route. Even with a tas a human is still experimenting to find the right inputs.

[@Ortorin]

​@@darkjapan It's not just that the inputs are being found by the brute-force scripts, the real problem is that to validate the inputs, they use a TAS. At that point, there is no longer a human making a decision. The TAS is being used as the primary means of determining if and HOW a run will be preformed. Even when people try to reproduce a TAS for something like Mario 64, they still need to come up with human-viable alternative solutions to many parts, and they have to play all the rest of the level that isn't memorized TAS inputs. The existence of the TAS is used to INFORM the player of what is possible, and the player finds their own way to do it. With this LIS style of playing Trackmania, the TAS is not just informing the players, it is DICTATING how the run goes. Players are not driving the map under their own control at any point, nor are they ever taking a strategy and changing it to be viable for them. The inputs the TAS finds to be viable are the only inputs used by the player. The "player" no longer matters in this equation, and can be replaced by a trained monkey that will press the inputs the right way without that monkey ever needing to see the game itself.

[@andy02q]

@@Ortorin The current carpetless setup does the same. All of its inputs are dictated by TAS research and there's no deviation whatsoever, only room to use a different set of inputs entirely which is also dictated by TAS. The problem is not the method, the problem is that there are tracks which are so simple, that they can be completed as a whole on wr pace with a full LIS-setup. These tracks should be excluded from competition for most World Records.

[@zerocool6257]

just add a new catagory and be done with it

[@operator8014]

Well, this is exactly what speedrunning looks like to non-speedrunners. It's like saying they're playing the game wrong or enjoying the story incorrectly. Are people playing chess wrong by learning opening moves? This is just a far more advanced form of glitch hunting and practicing. They aren't cheating, they've just solved speedrunning in this game.

[@paulamblard3836]

and not "solved speedrunning in this game", but "solved speedrunning for few track where it can be used."

[@NabsterHax]

The problem isn't that legitimate LIS runs exist. It's that they're indistinguishable from illegitimate runs. Also, plenty of speed games have categories where the most popular and competitive categories are not the fastest. Doing an LIS run is very different from playing the game "normally," and the community is absolutely within its rights to segregate the runs so that those who do not want to participate in LIS runs can still compete.

[@ryanhodin5014]

​@@NabsterHaxI think making categories is more than fair here - Just like some games have a "no major glitches" category, or a "no credits warp" category, this gets a "High Inputs" sort of category.

[@ANTSMR_Dango]

Every speedrun community just tries to find the optimal sequence of button presses to end the game as fast as possible.

[@CB-lh4ph]

Thats why a lot of people love minecraft speedrunning. Every run is different so its all about your skill and knowledge of the game (and seed luck, of course)

[@SuperLifestream]

It's literally boiling it down to a Rubik cube. Perfecting inputs IS what speed running is

[@KetsubanSolo]

They spend a lot of time playing a game so they can spend less time playing a game

[@Bloodstar6078]

Yeah, the difference here is really just using the AI to bruteforce a route for you and then using other aids to help you time it. So there is no skill. Its bland and boring to the observers and those forced to compete this way. I'm not against it. Everything is a potential tool and can have it's own speedrunning category of people want it. I just don't find this very interesting and would not be interested in the runners who play this way. Best example I can think of is something like Pokémon credits warp. Beat the game in a couple minutes. Can be fun to watch for the novelty, but after that wears off, and speedrunners all have times that are about the same, the category becomes boring. You're not actually playing the game. People that use AI and programs this way are basically saying "how fast can we make our game less competitive and less fun to watch and play?"

[@RucadiDev]

@@Bloodstar6078 > Yeah, the difference here is really just using the AI to bruteforce a route for you and then using other aids to help you time it. Tbh this is done in a lot of speedruns, even some use external tools (like Wind Waker) during it, and a lot of speedruns sections basically consist in memorizing some inputs in a perfect-frame way.

[@drakenbakken]

Not gonna lie but most speedrunning these days are basically muscle memory inputs on frame perfect timing like rhythm games. This sounded like a breakdown of how most top runners of all games do their thing even they arent aware of it Mario 64 also did the same thing using millions of ai runs to find new strats but that was seen as genius on this channel just a few videos ago whats the difference?

[@aSmolGoth]

Jobst is just reporting. Different communities, different opinions on the same subject matter.

[@ohno5507]

​@@aSmolGothyeah but he does give his own opinion and argue for it. I think it's fair to ask why it's "genius" in one context, but "ruins competition" in another

[@Burt_Cokain]

I mean he said the reason. You give a few second run records to everyone. Combined with the fact that it’s not the gameplay wished on TrackMania, it seems obvious to take this off.

[@Stuffinathor]

I really think the difference here is the amount of effort a speedrunner has to put in. Yes AI can help with finding new strats, but it's still a difficult endeavor for the runner to perform the new strats and there is still plenty of room for skill and failure. Compare that to pressing 6 buttons frame perfect to get the same exact time as everyone else that did it before. There is no competition in that.

[@Barrowsx]

The way I see it, finding a faster strat in SM64 helps speed up a segment of a longer run, so even if you hit the faster strat there's still room for additional skillful execution of other parts of the run. With this, rather than a single strat, you're memorizing an entire speedrun since these tracks are so short, and that isn't the type of gameplay the Trackmania community wants to emphasize. It's why, at the beginning of the video, LIS wasn't seen as controversial when it was used to add consistency to an inconsistent trick at the very beginning of D10 Race, but once it was used to optimize the entirety of A12, whose record was seen as incredibly prestigious due to how inconsistent it was, it gained significantly more scrutiny from the community because it turned the entirety of that track into rote memorization.

[@fidelrivera2887]

Can't wait for your inevitable video, "How I Completely Destroyed Billy Mitchell in Court"

[@Lu-db1uf]

In Trackmania? What? You might be on the wrong video

[@AdrianGK47]

​@Lu-db1uf They might just still be hyped for it

[@pimespo]

Billy would probably sue the judge after the verdict.

[@longdongmc.johnson]

been waiting for 2 years now. it will be amazing.

[@yugimuto1639]

I need a ace attorney style reenactment when it happens

[@KrossFire330]

Brought to it's logical conclusion, all video games will eventually be distilled down to some fucked up version of Dance Dance Revolution - where only your inputs matter because everything else happening on screen is background fluff to ensuring the correct sequence of button presses happen at the correct time.

[@RanialsAwesome]

You oculd say that about any game at any time.

[@Tuubasd]

"Moderation won't be difficult" might be THE most naive thing I've ever heard on this channel

[@juliushibbert1091]

we all need to start shaming gaming script-kiddies again like everyone used to 10-15 years ago

[@Alexstarfire]

@@juliushibbert1091 Also 20+ years ago. CS has been around a long time, and I doubt that's the first game with script kiddies.

[@ThatsPety]

how would seeing what runs are LIS be difficult? genuinely asking

[@Mortifix]

@@ThatsPety Because it will evolve. Like why not just add 2 or 3 extra inputs into your run in an area that you know won't affect it whatsoever.

[@ThatsPety]

@Mortifix I cant see how that would do a very good job obscuring the fact that it's an LIS run. I would think a run consisting of a small number of very robotic inputs, plus a few natural looking but inconsequential inputs, would still look extremely suspicious, no? Or maybe not, idk much about the game honestly

[@1Raptor85]

The high jump comparison is flawed, it's not like putting springs in your shoes, this is like having a planned spot on the ground and a specific timing on the jump pre-memorized helped by planning tools (which is something that most sports actually have). It's not "jumping" for you, it's planning your strategy, you still have to be good enough to follow it. Plenty of other games use TAS tools to plan their routes as well, and it's never been a problem before. I think it's mostly just seen as a problem here because the tracks are so short that winning strategies will be fairly simple to pull off.

[@brennantmi5063]

My favorite part of the high jump analogy is this exact same thing happened in high jumping. Dick Fosbury won the 1968 gold metal by doing that silly backwards jump, everyone else at the time was just leaping upright. Many people wanted the backwards flop banned because it was an entirely new and better way of completing the competition within the criteria of the competition. Thus, making a high jump analogy for this would more accurately be making a comparison of the joystick runners are the high jumpers who leap upright trying to get the low imput runners aka the Fosbury flop banned because that tactic was better than the norm even though it is fully within the rules and spirit of the event.

[@vikkysenpai]

What's even more crazy about the comparison is that even when showing a visual indicator of the inputs (which are all frame perfect) the person attempting them couldn't even get 4 of the 7 inputs done correctly. So even with the tools it doesn't mean you're going to get it done correctly, you still have to put in the practice for the timing.

[@autumnleaves1213]

Agreed. This feels like the old guard being unwilling to accept changes. If there's an optimal way to do a level with a minimal number of outputs, then that's just the pinnacle of what can be reached on that level. And if LIS could be used on most levels of the game, that just means that the game's skill factor was artificial to begin with. Having everyone be able to get a record sucks of course, but you can't just artificially stop that by preventing the most effective strat from being used. If you do that, then what is even the point of "fastest time" when it's just "fastest time outside of optimal strategy"?

[@NN-ge2wf]

Yeah the high jump thing is the worst possible analogy for this, and if anything just illustrates the idiocy behind this whole thing.

[@GraveFable25]

​​@@autumnleaves1213Most maps don't have a LIS. That's not really the issue. The problem is that the more the leaderboard is populated with these runs that are impossible to verify the less people will care about it anyway. The non LIS leaderboard would be created organically anyway and would quickly become the one everyone cares about anyway. The ruling just cuts out the middle step.

[@krzysztofzaborowski8132]

"memorizing a few inputs and using the game's timer to execute them" this is how speedrunning works when the game is not randomized like in this case

[@SoulSin]

Agreed. This is the price that trackmania pays for being like this. You could forbid the tool, doesn't change the fact that people would indirectly use it.

[@sage5296]

Red ball and Super monkey ball come to mind here, both of them have level by level strats where you just go for basically a human executable tas

[@isocle]

Yeah, imagine if SMB1 was just a completely deterministic speed game, I bet no one would even speedrun it...

[@nomore6167]

Exactly. Didn't Karl recently have a video about Castlevania where you had to get hit by an enemy at a very precise point in order to walk up stairs and reach a glitch point? This LIS issue is basically speedrunners saying, "you're finishing too fast, so it doesn't count". Or, alternatively, "your time is so far below our times that we simply won't allow it".

[@NabsterHax]

@@nomore6167 The problem is that LIS runs can't be verified as legitimate. They're indistinguishable from a TAS. Also, a speed game that has been "solved" with a completely optimal solution isn't interesting or competitive. In other games we have categories to prevent new glitches/techniques from destroying beloved speedruns even if they're technically faster.

[@Churahm]

Not sure I like the idea that using "brute force" isn't a good way to play. Tracks like the one displayed in the video is literally just a bunch of people making a left turn and bumping into a wall thousands of times until it works that one time. It's pretty much brute force in its own right. IMO the issue is with the track itself. That glitch that allows to skip the whole thing is so whack, like what do you expect.

[@aquilegia9734]

Sounds to me like this whole situation can be solved if people who are against LIS instead focus on more "normal" tracks that don't have weird out-of-bounds bugs that glitch your car from start to finish within seconds of starting the race? Just accept that some tracks require LIS to get a WR, and some tracks require "classical" racing skills. Sooner or later any speedrun of a fully deterministic time-trial game will always come down to humans trying to most accurately reproduce a TAS. And if a track/level in the game can be optimally completed by executing just six (frame-perfect) inputs, it might just means that particular track is not very interesting.

[@nomore6167]

If the strategy was originally found by mistake rather than found using TAS, you know the title of this video would be something like, "Brilliant gamer stumbles upon new world-record strategy requiring just six inputs".

[@meraldlag4336]

People act like the old a12 wr holders weren’t already doing the same thing over and over again in order to try and get a better time

[@ThatWhichObserves]

The problem is now, they've done none of the work.. They're in a very broad sense of the concept of artificial intelligence.. using AI to generate 'optimal' runs..

[@Needformadness2]

Holy disingenuousness

[@roevhaal578]

Yea it's really only a problem on braindead luck maps.

[@ForeverLaxx]

I don't see any reason to disallow these types of runs. When you boil it down, it's just like any other any% speedrun of any other game. If they don't like that, they can just make a specific category for it. There's no need to ban it outright.

[@JeanKP14]

That's why they said they're thinking of making another category for it lol

[@NabsterHax]

The issue is that it's impossible to verify if they're run legitimately. And they're not being banned, they're getting their own category.

[@S0ULESSB0NES]

MFW a deterministic game has it's speedrunning competition boiled down to "push this button at this exact time"

[@joebykaeby]

All forms of video game competition inevitably end up becoming an arbitrary sequence of inhumanly-precise inputs that's completely divorced from anything the player was ever intended to do in normal gameplay. This is literally no different than any trick setup for other games where you need to line your character's shoe up with the right pixel and then make the character randomly jump, crouch, flip, and twirl around so that they end up at the right sub-pixel to pull off unintended movement.

[@Obi-WanKannabis]

Agreed. I still think it should be banned though, because it makes cheating trivial.

[@Freddie7191]

Except in those games that trick is just one part of a bigger run, whereas in this game, that trick would be the entire run.

[@chris-m1g5r]

That's my first thought: "How is this any different from setups that many games speed-running communities use? On the topic of tools, how is it different from the tools OoT runners use to find ROM warps" but the thing that makes this different in my mind is the time of the run. Many games have setups for things like cannonless, but that isn't the whole game. Imagine if the entirety of SM64 was spawning next to the wooden bridge and grabbing the cannonless star and every level was just a different variation of that with a different star that had a similar setup. They would ban these kinds of strats too because even though the setup exists, there's far more things that happen outside of the setup in those games that still test runners skills, they aren't just that setup, whereas Trackmania is a game of levels that can be solved with a single setup.

[@ShiningDarknes]

I only disagree with the ban because they did not already have a category for this 100% legitimate strategy set up before deciding to ban it, only the promise of setting up a category in the future. This is ultimately the end point of any solo driving game with zero variables outside the player’s control. At some point someone figures out the absolute optimal path and how to make that path happen then suddenly everyone is chasing tenth or even hundredths of a second for the top slot.

[@weare2iq376]

@@Obi-WanKannabis Only on certain maps though, they'll just have to be ignored in the future. Question becomes at what point do you make the cut off? I think it should be whenever 100 players make the mark, after that you grind the next level... At some point in the future someone will become 100th player to complete the last track in x moves. Game over, on to the next humans beat this one 😂

[@kaspikey9553]

Trackmania can’t catch a break huh.

[@progamerjack01]

Nope... :)

[@cheese1627]

Cheese

[@Buchaus]

If you brake, you slow down. And that's not gonna get you a world record

[@progamerjack01]

@@cheese1627 😋

[@marcosdly]

The game itself wants a break by now. It's the players that never do.

[@bruhmaxxer]

The big issue here is that even if it's dishonest to calculate an optimal run with that software, you can't really un-know the optimal way to play. You can't ban information from spreading. I guess, in a sense, that it's a cognitihazard.

[@Travixty]

knowing the strat exists doesnt help unless you specifically go for it

[@bruhmaxxer]

@@Travixty it can help though, as you can apply knowledge found from the strat in a real run.

[@user-gi7vi9gm4t]

just don't execute that run

[@bruhmaxxer]

@@user-gi7vi9gm4t There's also the question of wether the player was being legitimately just efficient, or wether they were cheating. Of course there's the "ban less than 8 inputs thing, but what if the player uses a minimum input strat for only a segment of a run?

[@BrooksMoses]

Thing is, these aren't actually optimal runs. They are optimal runs within the low-input constraint, but it's almost certain that there are faster possible runs with more inputs in nearly all cases. Also, to partly answer @bruhmaxxer's question: The only place a minimum-input strategy works is when you know exactly where the car is, so it would have to be at the start of the run. Still a bit of a problem there, though. (And, really, that's what's beeb happening with that 1.83-second-left-turn wallclimb for ages, though it's a very very short strategy that only applies to the first two inputs.)

[@Rbrxrbrx]

This is no different from the doom speedrun that used a computer to calculate a setup for the void. Completely fine, as long as it's a human doing the inputs.

[@Zantonny]

Memorizing inputs and then repeating that memory is literally how Trackmania is played though. It isn't going to win on more complex maps, but if it's a short simple map... that's all anyone is doing

[@SquidGains]

I guess it's just a difference in where those memories come from. Actual experience playing or a tool telling you what sequence to input. In a sense not that different but the latter does seem terribly boring imo.

[@The_Big_Jay]

​@@SquidGainsWelcome to the natural progression of speed running. It goes from people showing genuine skill at the game to a sweaty guy of stream hitting 4 buttons and skipping to the final boss and mumbling, "Th.. that was a gwimble skip.. it happens because of stack overflow.." and then beating the game without showing proficiency in anything but the ability to hit buttons in the right order at the right time.

[@jakobjas4212]

@@The_Big_Jay And that's where speedrunning kills a game in my opinion. Sucks any of the fun out of it.

[@The_Big_Jay]

@@jakobjas4212 Totally agree. I also fail to see how those people actually enjoy the game. They insist that they do, but when you skip 90% of it, I don't really believe them.

[@Churahm]

@@The_Big_Jay My favorite is when some runners know exactly how to glitch speedrun, but wouldn't be able to even finish the entire game quickly because they only know their "route"

[@Johnnyb3g00d]

It's always curious to see when a speedrunning community goes "you can cheat" but "not that kind of cheating". Every community is different. My thought is that this is, unfortunately, the inevitable conclusion to the game. Why wouldn't it be? No longer are people beating the levels in the fastest time possible. They are now beating the levels in the fastest... arbitrarily decided 9-or-more-input runs.

[@selenasilverstep7981]

karl even points out a good solution in his golden dark example. they banned the LEVEL, because "the fastest time" OF THAT LEVEL is no longer interesting for competition.

[@peekachugaming3231]

A map with 9 inputs is hardly a worthy map. People should stop complaining. They are mad because some idiot pressing by time beats them lifeless souls. What if the pro's just would play maps with more than 8 inputs mhhhhhh... What if every map below 12 Inputs is LIS / Normal, and everything above 12 is LIS banned.

[@p00ya_]

They didn't ban look-down in the end though

[@Matfridt]

​@@peekachugaming3231They do that already, look up Mudda's run on Deep Fear, it's arguably the greatest run in Trackmania history and you literally could not use LIS to beat that run. Only tying I can think of that could beat it is probably some highly advanced AI TAS script

[@Swarm_]

@@peekachugaming3231 Most people grinding for these WRs are not pros. The best players do not spend a 100 hours on one 10 second track. Wr hunting and being an overall top pro is very different(although the people grinding for WRs are obviously insanely good, but mostly a level below the absolute top). You won't see Carljr hunting a uberbug map for 10 hours.

[@mrphlip]

I do feel like there's a meaningful difference between using tools to _play_ the same fast, vs using tools to _route_ the game, and then executing that route by hand. The latter really doesn't qualify as "tas" to my mind. There are plenty of speedruns which have individual tricks and setups that involve executing a specific pattern of inputs, found by computer. It's just that usually that's just one trick in a run, while here it's able to be the entire run. As is usually the case in these sorts of debates, my default stance is "split the category, have leaderboards for both, wait and see which one ends up seeing activity long term"... in this case I feel like the non-LIS category would probably see more consistent activity, while the LIS category would at best see occasional bursts of activity whenever people find a better route (like, maybe currently the leaderboard is a bunch of people tied first with the best 6-press route, maybe in a couple months it'll be a bunch all tied first with the best 8-press route, and there's an even better 11-press route that a couple of people are grinding trying to pull off, idk)

[@Euclase]

The problem with separation is that it happens not because it's different, but because people are angry. This category will be perceived as full of cheaters no matter how you found or executed these strategies

[@ruolbu]

@Euclase eh its not an issue, people are always inherently judgement, this is the kettle calling the pot Black. for outsiders speedrunners are cheaters and weirdos in general, what does it matter if one in group dislikes another ingroup

[@tkienjoyer]

The problem isn't that people consider it tas, it's the following: A. There is no way to differentiate yourself by being "better", unlike in normal trackmania, where no two runs are the same, in LIS runs everyone has the same time, which would ruin the leaderboards. b. LIS runs are really easy to cheat, due to the nature of the runs, so you'd either need too be extremely excessive with proof standards, otherwise people will just upload a tas and (without excessive proof standards) there is no way to know if it's legit. People aren't saying it's tas, no one has. People just don't want the a12 leaderboards to have 100 people tying the world record in a few years. I feel this is a good compromise, recognizing the efforts of some players, but not dooming the leaderboards to everyone having the same time except for maybe one guy who did 9 inputs to save 3 hundredths.

[@mrphlip]

I mean, the fact that these runs make cheaters harder to detect is a real problem they'd have to solve... it'll require new rules and more thorough verification. I'm not in that community so I can't speculate about specifics. I don't think that it being hard is a reason to throw them out forever. But I can understand restricting the main category first while figuring out the details for the new one.

[@mrphlip]

@@tkienjoyer Maybe no one who actually runs the game considers this tas, I don't know, I'm not in that community. There are definitely some people in the comments on this video who are suggesting it, though.

[@chickensofdeath490]

This is the natural endpoint of speed running, sucking out all of the fun of playing the actual game and reducing it to a couple button presses. Obviously this shouldn’t be banned, the whole point of speed running is finding out the fastest and most optimal path to victory.

[@randomjunk1977]

I like to jokingly say the goal of speed running is to have the least amount of fun possible.

[@ZBostOneR]

Thats not true, the goal is to have fun and compete thats why there are categories. How they manage them leaderboards is what sucks the fun I guess.

[@DarkGerson]

just for Tod Rodgers to find out about this game and come up with an even crazier strategy

[@SayAhh]

Toddmania! He will claim to be able to start in 2nd gear...in a CVT car!

[@DarkstarArchangel]

His name is Todd Todgers!

[@guiltytrain3689]

Todd Rodgers would already have every record on every track by at least 5 seconds. He would shift into second right when the race starts. 😂😂

[@marcosdly]

His records will be whatever the legit WR is, minus 1 millisecond.

[@bostonrailfan2427]

@@marcosdlythen will lie about people attacking him for being too good until a programmer exposes that it’s impossible

[@fluke1]

8:23 i think it's silly to say that the point of competition is to distinguish players based on their skill level. the level with the 6-input uber bug is a completely luck based competition for the WR because no two ubers will be alike without low input strategies (LIS). does this level test the players' skill level? no, this map is just a test of their luck and a little bit of their determination. similarly, the maps that are most impacted by LIS are those that are short and have one special cut/trick. that's why i think LIS are morally okay. the cheating issue is a lot more serious though.

[@ruolbu]

this to the T. the example level is not a test of skill, its a test of the grind or die mentality. And any level that can be manually optimized to 6 inputs will be manually optimized, I don't see people screaming about that. If a level can be optimize that much it deserves to be optimised, whether it's with a tool or not. Manual execution is where its at and if the optimal way to play a level is trivial then its just a bad level just as Karl said at 9:25

[@seanrrr]

My thoughts exactly. A12 in particular is just a grindfest. Whoever has the free time to spend 100+ hours hitting the same wall over and over again will eventually win. The only skill here is being able to do the setup consistently.

[@xchronox0]

@@seanrrr You can just not run the track though. Let the wall rammers have their glory for their patience, instead of everyone sharing a participation prize.

[@irmofs]

How is this different than “carpet less” on SM64? A TAS tool was used to brute force that section and gave multiple low input solutions that players then tried and successfully used..

[@TrixyTrixter]

I guess the problem is that LIS don't force someone to write the inputs painstakingly frame by frame. Its like the only thing I can think of here.

[@siosilvar]

The only difference is that TM records are all individual levels, while carpetless is just one trick. LIS completely obviates a handful of campaign maps, and is completely useless for basically any map with checkpoints.

[@coconuts7960]

fundamentally comparing two different types of games and then acting like theyrethe same

[@HyperMario64]

​@@coconuts7960 I think the comparison is actually fair, the difference is pretty slim if we look at the trick itself and how it got discovered. The difference is that SM64 runs are quite long and runners need to manage risk to improve their times. At the trick level runners compete in consistency first. It's very common for speedrunners to use setups for hard tricks, it was just not really happening in TM before because setupless is usually faster. Quite a weird meta.

[@ZwhatshQ]

Because carpetless is just a tiny fraction of a huge speedrun, you still need amazing skill to complete the rest of the run, while LIS is the entire fraction of the entire speedrun, if you just get those inputs right, you're done.

[@iterative3605]

"[...] play the game as intended [...]" define intended, because I could argue that doing 'flying backflip 360 no scope' to the finish line isn't intended. At the end of the day it's not "intended" it's more following a set of arbitrary rules to ensure fairness between competitors.

[@Xeractyll]

As someone who's been around the Trackmania scene for a while, I'll give my take on LIS: The whole point of TM is to speedrun. To get the fastest possible time. I feel as though since other games let you use and deliberately go for these strategies to compete for top times, why shouldn't TM? I understand it's awkward and weird having a hunt for specific frame perfect inputs, I understand that kinda goes against the spirit of the game, but it's part of the spirit of speedrunning, and that's what I like to focus on. Regardless, I completely understand the ban on the usage of it due to the fact that its very tough to properly detect when someone is cheating using the method. Anyone could be using a program to help them out externally that can't be detected by the built-in anticheat in Donadigo's Competitive Patch for the game. I will say that I also find it odd for metronomes and other visual cues to be banned too. Again, goes against the spirit of TM, but I feel like its part of the spirit of speedrunning. Guess its another community choice thing. And one final thing: As someone who's friends with that guy at the end who cheated, I'll say that while he's been handling a lot of shit from what happened, he's actually a nice guy and imo, doesn't deserve the hate. What he actually did was use an external program for visual cues on when to time the inputs. Though this occurred after the ban on LIS anyways, so he wasn't even planning on going for an official record with it. Imo it's barely even considered cheating from a speedrunning standpoint, but it's against the rules on the leaderboard page (even though since it's a LIS run it can't be uploaded to leaderboards in the first place?) My point is just please don't put hate on him for it. It was something small, and he's already lost so much within the community because of it. He's not a bad person by any means. Also, Trackmania Nations Forever is an awesome game guys!! Please come check it out if you haven't already

[@AtomicArtumas]

tbqh, as someone who's a challenge runner and minor speedrunner, but has been a Pianist for quite a while and am also a professional composer, I find it hilarious whenever a community tries to ban metronomes. Like... it's so obvious they're not used to being around people that have used and been around metronomes for extremely extended periods of time. I can accurately count time without one with absolutely no problem regardless of what I'm doing, even when doing stuff like... "various physical activities" if you know what I mean, etc. etc. - Unless I'm overly tired or possibly if I was drunk or high (never done drugs or gotten drunk before so couldn't tell you.), I can keep a basic time rhythm without any form of external stimulus with absolutely no problem. So all banning the use of a metronome would do would be to give me a massive advantage that it would take most other people over 10 years to accurately replicate.

[@Xeractyll]

@@AtomicArtumas it's just banned cuz it's a community thing, all speedrunning communities have set up what they feel is appropriate for speedrunning each game. Imo it's weird from an outside view, but it makes a lot of sense from an inside view.

[@Squatch24]

This is the future for every single game. Look at Ocarina of Time, once a top speedrunning game that had its' coding figured out by TAS then easily replicated by humans to beat the game in under 4 mins. It's just the logical conclusion to almost all games. No real solution besides making 100 BS categories that confuse and lose interest by everyone.

[@ohla3644]

exactly, it sucks but they kind of just figured out how to absolutely maximize track mania. i mean he brings up that "when does a racing game just become a rhythm game" and i mean if you think about it that's literally speed running at its perfection right?

[@Darkota122]

You need to keep in mind that most tracks in trackmania are not beatable by LIS and will never be due to the checkpoint system. Yes there are computer solutions exploiting nose bugs but they are not humanly viable.

[@Fay7666]

​@@Darkota122 It's likely that every standard track will have an LIS setup. S4D is already kinda rhythmic so you can add that in, then add whenever you need to make a turn and there we go. And I'd argue, not any different from a non-LIS run.

[@dopaminecloud]

the categories making people lose interest is kind of intended, to take prestige and excitement away from these methods

[@lilthreadd]

@@Fay7666 how did you judge that its very likely?

[@WuRuYi]

Luckily this is only viable on shorter maps where inevitable input errors don't snowball. Many of the most prestigious maps are so maxed out that being even a couple frames off in a single corner will crash instantly, so I don't see LIS touching most of them any time soon. However, never underestimate the sheer autistic grinding power of speed runners. Also, it doesn't affect maps where the optimal strategy requires analog steering (basically all full speed and non-stadium tracks), where hitting an exact and changing steering value through a corner is not humanly viable.

[@sealiusmaximus2330]

Yeah, unfortunately those short stadium tracks are some of the more played tracks

[@rotciv1492]

So basically, LIS is turning short Trackmania tracks into Punch Out Glass Joe WRs.

[@user-vu9yq3kj1w]

I can see having different leader boards for it, but I don't see why it's controversial though. Didn't the Mario 64 Carpet skip video also use a program to brute force it and help give basic input setups, how is that any different?

[@karlvyh]

I would say the main difference is that "carpet skip" on its own isn't a prestigious record that this method reduces to specific inputs where everyone who can copy them would tie for WR.

[@facecat1]

Carpetless is only a small part of the run but it was very controversial too

[@Yttriumble]

Runs that have carpetless as part of them can be distinguished from TAS. That is the difference.

[@harstar12345]

​@@karlvyhthis is any% If there were a way to skip checkpoints and still finish TM community would embrace it. It's just elitism of the current best players not wanting to lose their edge

[@EzmarVoD]

Because the standards for different games are community-specific. It's not down to whether or not a strategy is inherently positive or negative for a game, but rather down to if the community is in favor of it or not. It's all subjective.

[@weare2iq376]

Sorry Karl, but Trackmania mod's insight into how cheating will be ruled upon if they have 8 or less inputs just isn't going to work. They're just going to get a bunch of 9 input records... I know nothing of this game other than what I've seen here, and in some other channels where WR's are a thing. Here's my hot take though... How are you supposed to moderate against LIS now that it's a well known thing in the community? It's still human input, just with extra knowledge now (minus the cheaters). If mods apply a higher input limit to be human it'll just be met with LIS at that limit. 9 inputs, woooo. 10 inputs, wooooo. They'll be wasting their time, and at what point do you stop applying a limit? Isn't this just what all gamers are looking for? The ultimate shortcut to perfection... This is certainly going to spread beyond TM into things like Mario Kart, etc. Now that I think about it Mario Bros. is basically now at the point of emulating the simplest TAS to match the fastest speed possible. Is it surprising this is the way game world records are going? Just trying to stop the inevitable... Real human competition has always been, and always will the way forward (Still can't stop cheats there either 😅). tl;dr This is inevitable, face to face competitive gaming is the way forward...

[@lilthreadd]

"They're just going to get a bunch of 9 input records..." i wish i could explain just how fucking wrong this is, but i legit can't, because you have probably never touched this game. 9 input records sound as doable as human noseboosts lmao. possible in practice, absolutely impossible in reality. the mods aren't retards, they chose the number for a reason bro

[@ElAkrayNOS2378]

Moderators: 8 inputs or less, please. Me: *wiggles furiously in the air on my way to a LIS wr*

[@solomassive6233]

Every time I hear about a situation like this it always seems like it’s just the current leaders and top players that fight the new way. But after a while usually the new way is begrudgingly accepted and becomes the new norm. I think they just don’t want to feel like all the time they spent doing it the slower way was wasted.

[@NorrwichGreen]

Interesting historical fact about the high jump example: the technique in the clip is called the Fosbury Flop and when it was used at the 1968 Olympics it basically replaced every other technique because of how effective it is. While an interesting parallel to LIS, I think the important distinction here is that there is a difference between a clearly superior technique vs trivializing the competition as a whole. Thought it would be fun to mention, good luck at the trial!

[@SnowWolf2150]

Reading this kind of reminds me of the Javelin Throw as well. They had to start changing the Javelin a bunch of times because throwers were getting too good at it and throwing it father and farther by changing up their techniques.

[@knoxiegb1782]

tbh that's exactly where I thought where Karl was gonna go with that

[@sysop073]

That would have been a much better thing to discuss instead of a truly terrible analogy about modded shoes

[@jakobjas4212]

I think it depends how game breaking the new strategy is as you say. If we're sticking with the high jump analogy, the Fosbury Flop is a great optimisation created within the framework of the sport. LIS would be equivalent to being able to use a trampoline in the high jump. You're still jumping over the bar, but for all intents and purposes it's now a different sport.

[@miachen2635]

Yes, and it was also dependent on a change in equipment. Scissors kick is so you can land on your feet when there's no crash pad.

[@eluhmentssheer8664]

I think knowing there is a legit faster solution and not being able to do it is just as demoralizing

[@onyx5902]

I get why it's banned, but at the same time it's really dumb. If players just found a way to find lis without a computer, would it still be banned? Lis is just finding a good setup to a trick, but with a computer and let's say the trackmania speedrunning community would continue optimising runs normally just as they are today for 100 years or so, they would eventually come to these exact lis strategies, that's just how run optimisation works.

[@WizardofTruth]

Well yeah but runs would still vary and not be optimized down to be consistent like this

[@jeezuhskriste5759]

You’re right. It’s exactly the same, except the computer did the hard part of figuring out when to do things for you. If you want to play a game where frame perfect execution gets you a spot on the leaderboard, go play Geometry Dash.

[@SomniaCE]

I feel like the best solution would be having a standard and Low Input category for every track, but then you run into the myriad of moderation issues that come from Low Input and the difficulty of deciding what the cutoff for standard vs Low Input even is. There's also just the obvious outlined issue of what the point of creating a separate category is if it is less of a leaderboard full of diverse times and moreso just a single time repeating infinitely because you either got the perfect run that everyone else got, or you didn't. It just does not seem conducive to healthy competition for a racing game. I think about the racing games I run time trials in and the thought of them becoming frame-perfect solvable is the most depressing and anti-fun thing I can possibly imagine. Stuff like that is fun to see from a theoretical perspective in the same way a TAS for any game is super fun to watch and learn about, but I don't want TAS-style play to become the standard way to play. For rhythm games this style of plat and optimization makes sense, but again those are games that will have hundreds if not thousands of inputs over the course of a single track/chart that also allow you to hit at various timings to get various scores (Good/Great/Perfect etc) and so while a perfect run is possible, at the highest level it will rarely ever happen for the hardest charts let alone for all of them.

[@alfred0231]

I disagree that players would eventually find these perfect setups given the extreme amount of chaos involved here. And if they did find it they wouldn't find the lowest input variant and it would only be a handful of maps. With this approach it's any map like this is instantly solved. This way of playing the game is different enough to warrant a different category.

[@mysock351C]

Without the computer is literally the raison d’etre of speed running, to find the most optimal strategy and then execute it. Even simulations don’t seek to actually get the moves directly, just explore what further optimizations are possible. However memorizing inputs blindly is a much different thing since there is no actual strategy, or even thought, on the player’s part. Just a monkey pressing the right buttons. It’s really no different than just having the computer play for you.

[@danielcomings5872]

I mean, if Zelda speed running can evolve into "how to trigger a glitch to make the credits play as fast as possible", I'm not sure why "how the game was meant to be played" and "the spirit of the competition" is more important for Trackmania. If they created a separate leaderboard (equivalent to an any% category) that seems to answer the question already.

[@angrytigermpc]

This is why I don't even really like watching speedrunning - by design, the intent is to skip as much of the gameplay as possible, with an ideal speedrun literally allowing the runner to start the game, do something to screw with the code, and have the credits playing within 5 seconds of starting. The less engagement with the "traditional gameplay experience", the better, in their opinion. At that point, why even play the game? Why does "beating" Demon Puncher 2023 in a credits skip that takes 5.2 seconds to cause somehow carry more prestige than a 5.3 seconds-to-credits run of Dudebro Adventure? A person doesn't even have to care about either game nor how different they are from one another, the entire experience is just a vessel for them to execute an ACE credit skip.

[@mikereynolds1368]

Using ACE and/or other glitches is a different beast than just playing a game as fast as possible IMHO It's the difference between those that want to play DDR with odd graphics and those that want to play a racing sim.

[@luipaardprint]

@@angrytigermpcnow I really feel like playing Dudebro adventure.

[@arthurbrandonnielsen]

So, how does this differ from Mario speedrunners attempting to mimic TAS?

[@Linvael]

Scale? A Mario speedrunner at the top level does many more frame perfect tricks in a run, as well as a lot of non-frame-parfect inputs they do in between.

[@liborkundrat185]

There are few differences, but none which I can see as being an argument for LIS to be deemed illegitimate. As for the differences: 1. Mario TAS is optimal and unbeatable (until a new set of inputs is found) which can't be said for the Low Input Strategy. LIS doesn't get you the best time - it merely gets you a good time consistently. There are TASes of A12 which are faster than Pastagrows' LIS WR. (Just search for "a12 tas" on KZbin and you'll see many examples.) It's just that the bug-over-the-wall is so lucky and unpredictable that no one's ever managed to get it as well as the suboptimal (LIS) way. 2. If I recall correctly, Mario updates its position and variables 60 times per second. Trackmania does so 100 times per second, making each input more difficult and precise to execute. So in Mario, you have to match inputs which are less precise, but you have to execute a ton of them instead of just few. 3. Trackmania players aren't used to require a video of their attempt, as that has never been a requirement.

[@sorrowandsufferin924]

We already HAVE people mimicking TAS - for the line. Which means you're driving the run yourself. LIS isn't driving, LIS is memorizing. Trackmania is an arcade racer, not a memorizing game.

[@joehigh82]

EXACTLY! This is the most low effort ill informed and thoughtless video I have ever seen him put out. I hope this is a one off and not indicative of future releases.

[@noobzerg1990]

@@sorrowandsufferin924That is the inevitable point of any speed run in any game. Memorization. Any game any speed run is about memorizing certain routes and optimal solutions to problems.

[@ryan_ford522]

10:13 - in 2021 at a local SSBM tournament, someone was asking for my opinion on his genius idea (his words) he came up with; if he could make an app that displays the inputs need to beat SMB1 optimally but with a DDR style screen and if I think it would be possible for people to grind out for a WR run

[@RunicSigils]

If it were any other game, everyone would hail it as the new best thing for getting times down. I don't see how this is any different from finding out a new thing that lets you skip to the credits. Just make it it's own category.

[@appendix3812]

The issue for trackmania is that you can't make it it's own category. The leaderboards are shared in the game itself, not a separate website. It'd be fine otherwise, but players dedicating themselves to a track to learn the muscle memory + skill of carrying out all the inputs just to get beaten by someone staring at a timer and not even looking at the game feels awful.

[@luminica_]

@@appendix3812 As if hitting multiple frame perfect inputs in a row didn't require any skill. Trackmania isn't the first game where someone found how a precise route and went to execute it, I don't recall any other game banning them. In fact, it is common to expliot a sequence of frame perfect inputs to do a skip or kill a boss.

[@bostonrailfan2427]

it’s literally what’s being done in every other speedrun game, it’s only an issue here because of a few demanding certain arbitrary rules that never existed before

[@LucianDevine]

@@appendix3812 And how do you prove that somebodies weird looking run was LIS instead of something they stumbled upon? How are you supposed to catch a run that was discovered solely because of LIS, memorized, and performed, versus a run that was stumbled upon, mostly memorized, and performed?

[@lgbtthefeministgamer4039]

@@luminica_ it is, yes, but these other frame perfect inputs you talk about that are allowed are just one of the many things you need to do in order to pull off a speedrun. i think the main issue here is not that the hunting program exists and gets used to plan live runs, but the specific strategy here where you can boil a speedrun down into 6 frame perfect inputs in order to get a world record in a way that means many people will get any world record gained using this trick

[@Shivaxi]

I feel overall this is a bit silly, both the LIS method and the decision to ban it and only allow runs with more than 8 inputs. People have been using the in-game timer to time certain inputs forever, so unless Nadeo updates the game and removes the timer, this is just always going to be a thing. On top of that, the majority of official maps already require more than 8 inputs, with some maps being quite long. That would be a LOT to memorize, and even if you had that fancy half pie chart with a sun dial on your screen to show you when to hit the perfect inputs, you only have so much screen space as those lines get more and more condensed, making it harder to tell when you need to hit the key and being far less accurate, especially on maps over a minute or 2 minutes or even sometimes 3 minutes long. So LIS doesn't become that effective for the longer maps, the in-game timer would actually be more accurate...and then it's just a LOT to memorize, which I s'pose is doable but like...can you ban someone for having really good memory on when they need to hit certain inputs? Even if referencing the in-game timer? That's still a skill in itself. I understand wanting the game to be played the way it was meant to be and all, and outside tools of course are a big no no, but this feels like a losing battle if the community tries to ban ANY kind of memorization (again 8 input minimum is just really funny to me)...which is a huge part of any game and speedrunning I feel like. Dunno, just food for thought.

[@kape2978]

This feels the same as MvM purists in TF2, maximize every strategy to get your loot faster but if you ever try using the gas passer because it is more optimal, you get marked for death basically. Crazy.

[@memereview1337]

Didn't realize how a degenerate Anti-Game TF2 has really become until I saw this comment.

[@callumbreton8930]

Nah, that’s just MvM, hence why no one plays it

[@kwanzan61]

The only problem I have with banning LIS is it seems to demote the use of a keyboard vs a controller because of how the keyboard is on or off inputs it I feel like over time those strategies could have been found yes the TAS makes these strategies appear much faster and should be limited by just outright banning the LIS you are also banning people who hunted maps in that style without the use of a TAS, to me banning this style of play would be like if Portal banned a speed run of a map because they didn't use enough portals to solve the puzzle because there was a simpler answer

[@facecat1]

Bear in mind, the 8 input bound wasn't decided arbitrarily, no legit human (non-LIS) record on an official map is likely to have less than that, even on keyboard. The maps are just too complex, top players have to constantly make micro-adjustments to be as precise as possible. The point is to preserve the spirit of the game, and if someone does somehow get a new legit wr with less than 8 inputs and can prove it, I'm sure it would be accepted. And there are still insane non-LIS achievements by keyboard players in this game. Just the other day, a trial map that went unfinished for 6 years was beaten by a player using kb.

[@innertuber4049]

​@@facecat1the problem isn't the 8 cap, it's that a person who has practiced a track enough times on a keyboard *will* start to do the same inputs over and over, with the same timing. How can we differentiate a person who found this through AI brute force from somebody who just experimented until they found the right method?

[@14mwh014m]

@@facecat1 "if someone does somehow get a new legit wr with less than 8 inputs and can prove it, I'm sure it would be accepted." 🤣

[@etranablala9843]

​@@innertuber4049 there's no difference, if the player did a run with 8 or less inputs, it's a LIS run, no matter the method used to find the inputs. I believe the first LIS were found years ago by trial and error on some kacky (very difficult/luck based) maps

[@prow46]

​@@innertuber4049 Just to clarify: 1. LIS are meant to be human viable, hence the lowest input possible, however the average keyboard run on A12 would take around 20 inputs, which is 3 times more, so it's preatty easy to distinguish which one is which as it is unnatural to drive without small corrections, which increase the inputs' count, even on very short maps. 2. Even an offset of 0.01 seconds of a single input will give you a completely different run. For example on A12 any offset from those 6 input will result in a did not finish, like you wouldn't even get the uber bug. If you're experimenting you'll likely have more than just 6 inputs, it's basically impossible to get two runs with the exact same inputs to the hundreths of a second. 3. LIS was not banned because it's unfair or anything (D10 start as shown in the video is an example as it's a 2 input LIS and it's accepted without any controversy), but because there's currently no way to detect cheaters. Replay analysis was used multiple times to find cheaters in the past (most notably Riolu, but also more recently on A15), because most inputs didn't make sense, but with such a low input counts it's basically impossible to tell if the run was made by a human or not. 4. One thing Karl didn't mention is that leaderboard's replays are public, so anyone can download them (see the previous point). Anyone could download it, extract the input, and copy the run. Usually this would be noticed immediately, even on shorter maps, but here it wouldn't be the case. Of course you could just do the 6 inputs with TAS and call it a day, but there are already checks automatically performed to prevent the upload of TAS runs, but this will just provide another, parallel way, to upload cheated runs. 5. Video proof still doesn't address the issue as shown by Karl. TL;DR: despite still being remarkable to execute 6 frame perfect inputs (100 fps), banning this strategy is the only choice as there's no way to say if a run was actually driven by a human or not and this would possibly intact the integrity leaderboard, not about "the way the game should be played" but regarding cheaters. Plus the same thing might go undetected in community maps, where people still spend hundreths of hours.

[@therealsashagray]

Speedrunners use computer programs in this exact same way on other games yet those aren't banned. I think it's the simplicity of timing the LIS strokes and thus potentially thousands of people having the exact same world record and them fearing it will make the speedrunning community look like a joke as to the only REAL reason why they're banning it with this game when they allow it for others. Banning it makes no sense if you think about it, unless you ban using computer programs like this for ALL games, not just this one.

[@jacquelineking5783]

Devils advocate but there isn't an over arching speed running body making rules for every game. So scattershot in Mario 64 being okay was decided by its community and this was decided by the TM community. That said hating on a program discovering the lines is hypocritical when you put out a video praising scattershot for finding a human feasible carpetless. That is my problem with this video.

[@xchronox0]

@@jacquelineking5783 If doing that star was it's own segmented leaderboard (time starting when you select the star, ending when you get it) then yeah, it'd be an issue. But it's just a small part of a larger run. In TM, leaderboards are per track. It's easy to see it's the same technique but also easy to forget the context of how it affects the run itself. A tool making it easier after over an hour of manual playing is definitely different from making it easier from start to finish.

[@TheO416]

2:22 unfortunately it's both. speedrunners use timers/timing everywhere in the game the difference is before we had to find the right timing on feeling and you never 100% knew that what you did was optimal, now with this tool you do and it becomes so much easier. personally for games like this (meaning shorter runs) it takes the fun out of it because you're just doing as you're told and getting a WR. so yes it hurts this type of game

[@Chocomint_Queen]

My viewpoint is that trackmania is ALREADY a game full of frame-perfect bugs and tricks, and it seems pretty hypocritical to talk about how things like frame-perfect uberbug are important parts of the game but then going "but you can't do _these_ frame-perfect tricks". For many tracks, the game's already barely about who's the best at driving, and more about who's got the patience to ram into a wall enough times to hit a frame-perfect miracle.

[@labaffue608]

Trackmania is clearly not a frame perfect type of game. If you can't drive well it's impossible to get a World record without LIS

[@FlameStick]

​​​@@labaffue608 Oh boy, may I introduce you to some kacky maps?

[@Dakkalazy]

'wow I love it when speedrunners optimize game routes to achieve the fastest time' >speedrunners use computer to optimize game routes to achieve the fastest time 'no, wait, not like that!'

[@xchronox0]

"I love it when speedrunners save 10 seconds in an hour and a half long speedrun by using a set of discovered inputs at a very specific and late part of the run" vs "I hate it when speedrunners use a set of discovered inputs to trivialize a 10 second track" If the star were it's own leaderboard that started from the moment you select the star and ended when you collected it, it would definitely be an issue.

[@Uragaan1]

you mock that concept but there is truth in that. Sports and competition is inherently about human achievement and human struggles versus another human. It's the storylines that are interesting. Why have spelling competitions, math competitions, or any board game (chess, Go, shogi, etc) when computers do that stuff infinitely better than humans? There are people that do like watching computers fight itself to see what is possible, but that sort of thing is just not interesting to the majority. Sport inherently needs to function within a set of accepted rules to keep things interesting first and foremost for the viewers.

[@michaelstythax137]

When your racing game goes from racing the car to bouncing around a map using physics logic you're not really a car game anymore

[@Churahm]

Right? They use the argument that it isn't the spirit of the game. So you're telling me, intentionally bumping into a wall to be launched into the air, skipping the entire course and falling into the finish line, that's within the spirit of the game??

[@NabsterHax]

@@Churahm This is exactly the same in any speed game. Categories have "arbitrary" rules all the time that are designed to promote the most popular, fun and competitive components of the game, and diminish boring, unfun anti-competitive things.

[@radio100jogosdeemacs2]

I would've agreed about this previously, but now I know F-Zero GX is the GOAT and bumping to go off track is cool sorry.

[@awesker002]

Honestly is this really all that different from someone building a TAS to show the fastest way to beat a game like Super Mario Bros then attempting to replicate it?

[@lilthreadd]

yes, because super mario bros has a single input device.

[@Architector_4]

@@lilthreadd i'm unsure what you're saying? yeah, there's exactly one controller you play the game with lol

[@dkosmari]

Pretty much. People are salty because some tracks were too exploitable by glitches. If a Mario game had an individual level record that require a single trick to wall-clip to the goal after a few frame-perfect inputs, everyone would grind to match the TAS, and the level would be dropped out of the leader boards. Let these tryhards grind just to tie with the WR on this boring track, go play a different track yourself. When everyone has the WR, nobody gets to brag about it.

[@lilthreadd]

@@Architector_4 it's a bit different to make a tas in a game with digital inputs and in a game with analog inputs

[@Architector_4]

​@@lilthreadd sure? trackmania on keyboard and mario on nes controller are games on digital inputs, seems comparable

[@donoteatmikezila]

This is the classic any% vs 100% debate just in a different wrapper.

[@coredeep3424]

Agreed, though closer to Any% vs Any% Glitchless

[@charlescox290]

How is this different from players looking at a TAS in a Mario game and seeing how to reproduce it in live game okay?

[@RunicSigils]

It's not, but they cope.

[@xchronox0]

Looking at a 5-10 second TAS segment near the end of an hour and a half run is a lot different than looking at a TAS that has more seconds than it does inputs.

[@Kosmicd12]

It's funny, this is exactly how things work in most other game speedruns and finding setups is celebrated in that context. Funny how relative everything is to the specific game/community

[@ChicCanyon]

Speedrunning communities are so weird. If there are distinctive strategies only possible with input method A or B then they clearly should be separate categories. Any% will always be whatever is the fastest.

[@paroxysm6437]

I dont think you watched the video. Computers find the most optimal path and simplify it to where it's just execution based. Can you just un-know the most optimal strategy? Should we force people to find less optimal strategies?

[@dlericx1876]

Learning from TAS are not a new thing. I don't know where the controversy is

[@meikaishi]

Different categories seems the right answer every time, it's the same reason why Ganon% was created for OoT, sure it was super fast to skip to the credits after following very specific inputs, but it got to a point where the fun was all gone

[@brennantmi5063]

@@dlericx1876 This is the first time I have ever seen a community sad that TASers are finding faster and more constant strats. Any other game and these runners would be heroes. "oh no, people are playing a fresh file with no modifications in a way that is not normal to go fast, oh the humanity!" I don't understand how one could even have a category effectually called "Any% glitches unless the glitch has a consistent set up"

[@ComFurt]

@@brennantmi5063 If every map has the world record be literally the perfect possible time, competition vanishes. When the practice needed to become a WR holder stops being dedication to the game and starts being looking up a button sequence all the fun is lost.

[@smathlax]

But surely this is only a problem for short levels? The level shown was only 10 seconds long, so people can do 6 frame-perfect inputs. But ain't no way anyone will do 100 or more frame-perfect inputs in a longer level. I personally don't see any problem with this. I don't think the analogy with the high jumper putting springs in their shoes is valid. Rather, the analogy should be: is the high jumper allowed to analyse the materials of the floor from which they jump, so that they can develop a better technique? Of course they can. You can use all the tools you want to do analysis _before_ the competition. The shoe spring is a tool used _during_ the competition, which is why it's banned. Likewise a robot that presses keys for you _during_ the competition should also be banned. But the keyboard itself isn't banned, and studying the level _before_ the competition to learn inputs shouldn't be either.

[@seanrrr]

Yes, there are only 4 or 5 tracks I can think of in TMNF where LIS could be used. And those tracks are so short that the top records are almost identical anyway. I love your point about analyzing the floor! That's exactly what a lot of athletes do to standardize their techniques. Football punters take measured steps away from the ball so their run-up is identical every time.

[@davidbrinton1526]

Haha, I remember as a young child playing Mario Brothers in the 80s thinking, I could just write inputs down on a piece of paper, not look at the screen and just press the buttons on the paper and beat the game without dying (speed running wasn't even a thought then, just surviving). Of course that never worked because I couldn't incorporate the timing piece of it, but this totally reminds me of that.

[@Patashu]

Yeah, I love this too. I especially love the visualizer tool that tells you how early or late you were on each input, since that's the feedback most games don't give for even intended tricks.

[@darkjapan]

I bet you could had pulled it off using music cues.

[@ThisisCitrus]

But the child didn't use machine learning to figure out the most optimal inputs, they learned it by playing and testing.

[@corvetteee1]

Technology has been used to theory craft speedruns for a very long time. The infamous barrier glitch in Wind Waker was reliably cracked using TAS to measure the angle necessary to do the skip. Tons of other examples exist where technology was used to create speedrun routes. But at the end of the day, it's about beating the game/level as fast as possible. These people aren't cheating, they've discovered an optimized way to get the fastest time, using ingame tools and their own skills to do so after using other tech to test their ideas. Anyone who says that they're cheating, unskilled, or ruining the spirit of the game is probably just salty. While these players can tie speedruns on some levels, they'll never reach the same success on other tracks that can't be memorized with LIS.

[@idontwantahandlethough]

I actually really enjoy this "phase" of speedrunning: it allows the runners to take a back seat and allows the researchers/tech to really shine (since it's been distilled down to pure inputs)! I think that's cool, the people behind the scenes don't get enough credit :) P.S. also, "we won't be able to tell if people are cheating anymore" is an insanely stupid argument.

[@sunxnes]

I love that Karl gave his sentiment about this specific phenomenon in Trackmania, because it closely mirrors the “look down” phenomenon in Goldeneye that made him quit it for years.

[@d.kk.d8667]

i m playing since 2008 and never undestood why they must use LIS... play the game as it is, its simple. Thats what makes it competitive. (its fine to use TAS etc to find new ways-tricks etc solo and then come back without and try make it.)

[@sebastianprzygoda1866]

How is it different from memorizing hundreds of chess openings and games across history and using that knowledge against weaker players? It just engages different regions of your brain to achieve the same goal, and as long as inputs are human made, records should count

[@zergu-yf3gd]

the analogy works but that sucks in chess too.. becomes a game of memory instead of thought. that's why fischer random chess was invented, they shuffle the starting pieces so you can't just memorise solutions.

[@plexxarbiitch]

There’s a big difference between memorizing something and having the inputs you need right in front of you. No one would have a problem if they tried to memorize the inputs without a third party program telling them the inputs they need to make. This analogy doesn’t really work. Maybe it would make sense if chess players had a computer open telling them all opening moves

[@carctct6435]

Idk if you have seen competitive chess but memorized stuff gets thrown out the window after 5-10 moves. The rest of the game is played with calculation and tactics. Your analogy is more akin to "Learning surfaces in trackmania gives you an advantage against new players" which yeah it should reward you knowing the game. But competitive chess is rarely extremely rarely won by openings alone, skill is used almost every game. I making a point for casual/competitive divide as no one is using this in casual play and doesn't apply.

[@zergu-yf3gd]

@@carctct6435 nah it really doesn't, sometimes they go 20 moves deep in what they call prep aka prepared openings.. obviously depends what your opponent does, some top players are known to throw in suboptimal moves just to get their opponents out of their prep. but yeah it's not good for the game, and that's all my point is.

[@AeroPicsou]

A12 hunter here. (I must have 450h on A12, and I currently have a 10.85.) I'll be honest, what's most unfortunate about this story is that we're more talking about lis, than about the people behind it who actually hunt the map. And I'm sure that the day the map is actually beaten, everyone will be talking about lis even more, and put aside the insane new world record (because trust me, beating 10.61 is absolutely insanity) All I can say is that the map isn't just a game of chance, as some people say. Without going into too much detail, it requires a lot more skill than people might imagine. But of course, there's always a bit of luck involved. For example, I missed the wr twice to within 0.01s input (if I'd pressed +/- 0.01s key, i could have wr) and once to within 0.02s. And I know I'm not the only one (Viun, another fellow hunter also have two 0.01 fail). Anyway, lis won't demotivate real players to take down this record.

[@HMZeroTwo]

Personally I don't see a problem with this, it's just players doing a frame-perfect shift into second gear

[@Lucas.Blevins]

HA

[@CallomTM]

It was banned because there wasn’t any way to know if people were cheating

[@Operation_Bagel]

@@CallomTM You missed his joke

[@zikilemini2]

Talking about "playing the way it's intended" while showing something litterally called "uber bug" is beyond me. Either it's quicker, either it's not. If a track can't distinguish a real player from a computer, maybe the track is bad and shouldn't be prestigious.

[@timonix2]

this is absolutely valid. People have been using TAS to route speedruns for a long time. Often doing exactly this. Finding a a sequence that a human can replicate. Hell, you could even do this without a TAS. Just trial an error.

[@Bodharas]

I think that's the point. One takes you, the individual, to put in the work and learn and get the perfect run. While other does it for you until you do it yourself. I'm not arguing whether it's acceptable or not. I don't have skin in the fight. As an outsider looking in this is how it looks.

[@joehigh82]

@@Bodharas Once a run is optimized, TAS is what always comes in and helps push whats humanly possible to the next level. This has been done in Super mario 3, Ninja Gaiden etc.. all covered on this channel as positives in the past btw.

[@noobzerg1990]

@@Bodharasthis is ridiculous I hate the analogy in this video of having springs in your shoes for the high jump TAS is way closer to seeing a robot perform the jump and replicating the exact movements of the robot. At the end of the day the goal is to jump over the pole. If the athlete jumps over the pole after copying the technique of a robot. It doesn’t matter they jumped the pole. It really doesn’t matter who came up with the technique. Whether you copied it from another speedrunner or a TAS. Saying that “it’s the individual” is dumb. Should speedrunners not be able to look at each others runs?

[@Bodharas]

@@noobzerg1990 you're still skipping putting in the work. No matter which way you spin it. You're having a computer do the work.

[@Bodharas]

@@joehigh82 doesn't matter if it was done before. From an outside perspective it looks like you're skipping in actually doing the work. Whether it's cheating I don't know, but it definitely is a shortcut and diminishes the feat of a run.

[@wooloobutbrave9151]

Karl Jobs is the only creator who can always be a month late to every piece of drama

[@snowstorm9310]

I don't really see this as anything other than the logical end point of speedrunning as a concept. Your entire hobby is built off finding the fastest ways to finish a game, be it exploits, sequence breaks, arbitrary code execution and what have you.... This is the eventual end state of speed running games. Achieving the perfect run, with the fewest inputs as to make it as accessible to you as possible. All this really sounds like to me is people are mad Trackmania is simple enough that the end state came as fast as it did. You've got 2 options. Ban LIS for as long as you can until you eventually give in and accept that the future is coming whether you want it too or not. Make it it's own stand-alone category so you can feel better about the time you put into the game.

[@timothye.2902]

agreed. It sounds like Trackmania is essentially solved, or in the process of being solved, as a game, and the community is panicking because they want the game to be harder to solve than it is.

[@josef4852]

@@timothye.2902 I don't rly think that's true. This strat mainly applies to "luck" based skips and the times it gives you aren't achieving a "perfect" run, it just makes certain strats way more consistent. Also I think you are kinda missing the point of the times being easily cheated bcs you can easily fake these types of runs. But ye personally I don't rly see the issue of these strats in speedruns but ultimately it's for the community to decide what's best for their game...

[@snowstorm9310]

@josef4852 You can apply that logic to basically any skip that isn't 100% consistent in any game and still arrive at the same conclusion you came to, which ultimately is my point.

[@josef4852]

​@@snowstorm9310 I don't rly undestand what your point is ngl. Is your point that people will eventually get to the level of tas? No they will not. The thing is that most of these setus are not "perfect". I mean look at cannonless. People used to do setup but nowadays that is just too slow and they go for strats that are way less consistent but faster. I think in general these setups aren't an issue in the games bcs the whole runs aren't pressing 7 buttons like these short tm runs so you still have to "play" the game after performing the setup...

[@snowstorm9310]

@josef4852 No, that's exactly my point. They absolutely will, or they'll break the game in other ways as to make those lengthy setups unnecessary. That's just the forward progress of every game. They will all eventually be broken and solved equations that all follow the same inputs (with human error accounted for obviously) to achieve the solved fastest route. Generally, these always trend towards options that are consistent (either mechanical or trained to become so) for the player to achieve. Trackmania turned out to be *Very* easy to break, and now it's basically solved.

[@libertyjones3799]

A game community when a game that promotes precise gameplay is played extremely precisely: *shocked Pikachu* and *bans being precise*

[@DaijDjan]

I'll be honest, I don't really get the issue - trying to strip away unnecessary inputs is completely normal in speed running, including learning from TAS gameplay. I kind of get the fear of cheaters to some degree, but other than that it sounds like pro elites being mad to be shown how little imputs are really required to beat them.

[@BlackCatRedScarf]

Yeah. I guess that's part of the natural speedrun progression. Some may say that a lot of speedruns may become not valid as soon people begin to inspect the game code, use tools and debuggers to see how things can be optimized, since the game wasn't meant to be experienced like that, but if that's the case, a good amount of newer records would be invalidate and records would be able to be generated only until the first "cheating" occurs while all newer attempts would not be possible.

[@phaolo6]

You should watch the full video, there's a more reasonable argument for excluding or separating these runs in the second part.

[@DaijDjan]

@@phaolo6 I have watched the full video, none of the other arguments really convince me much though. I kind of get the potential cheating one - but everything else just boils down to "that's how it's in every speed run without a random component".

[@BlackCatRedScarf]

@@phaolo6 like... Someone can use a hidden screen, a sheet of paper or even memorize the inputs and timing (just like folks memorize getting around dark caves in Pokemon). Optimization is the heart of any speedrun and just because it is a run made with very low inputs, it doesn't necessarily mean that the person is indeed using third party tools to execute the inputs. While there are some losers who are going to be cheating with a lower amount of inputs, there are also those who naturally reached the optimization after studying the game and now they are operating pretty much at their margin of error and trying to get all inputs aligned as much as humanely possible. If there are doubts about the category and records, just create a new category that suits a certain audience and there we go. I really do not see the point of removing those records at present moment, but if people really want to see ... Perhaps... More legit attempts, then let's transform this into the "Guinness world record" style where judges need to be present then and see what happens. A lot of speedrunning scene runs pretty much as a hobby and based on trust, but as soon as things being to be taken too seriously or gets really sweaty and plagued by cheaters... I don't know... I suggest forgetting it and going touch grass.

[@NamaeAidiaNai]

The main argument that I can get behind is that this fundamentally changes the way you play the game. Like Jobst said, it's not even Trackmania anymore at that point because you can get the record fairly easily without looking at the screen. It doesn't require an excess of skill or even game knowledge to do. While speedrunning is invariably about beating the game fast, it is also generally in favor of actually playing the game as well.

[@jamez6398]

The problem is that they are literally banning people just from memorising a strat and then implementing that strat. You have to use the exact same controls deliberately less optimally because if you play as optimally as possible, you will be literally banned just for being too good. That is just incredibly stupid. I couldn't deliberately kneecap myself like that, if I was a top runner and wasn't willing to use LIS because I thought it was bad or whatever I would have to quit rather than play suboptimally even though I would rather just do LIS on the LIS leaderboards when they make those.

[@phir0002]

I don't buy the argument "this isn't how people would really play the game"... so you are telling me that wallclips and flagpole glitches in SMB1 are ways that "people would really play the game" or are they extremely precise actions whose only purpose is speed-running the game? How many of those glitches were found using TAS? How many optimizations to SMB1 speedruns were discovered using TAS? What is the difference from a practical perspective?

[@RunicSigils]

There isn't one, obviously.

[@hiramesensei3112]

the whole argument about it not being the "correct" way to play the game is almost the exact same argument that people who don't like speed running use to denigrate speed running in general, "because it's not the right way to play," and it's insane that Karl and these TM guys can't see that

[@Fay7666]

"AI bad"

[@ChrisRWitt]

The way I read that argument is "using brute force tools to solve a run, then simply trying to copy its execution is not how humans play." The vast majority of the work that goes into a speedrun is the routing and theory crafting which is entirely removed by LIS. You're just plugging answers into a box. Maybe this isn't the best analogy because I've always hated writing essays for school, but language model AIs have ruined essay writing in schools, and now it's basically impossible to use them as a learning tool for students because everyone is just getting ChatGPT to write for them and spending the rest of their time tailoring the essay to force it past AI detection tools. LIS is a sapling version of this.

[@Fataha22]

Like fr any% Super Mario 3 is including arbitrary code and casual gameplay can't even find that without TAS

[@TOBYSHERIDANWHITEPOWER]

I know some of these maps are absolute classics, but maps where so few inputs allow you to even finish are just bad maps in my opinion. The fix should be that maps are made better, with scenery blocking massive shortcuts while also being longer and more complex

[@BrooksMoses]

Yeah, I'd agree with that. The problem with Trackmania is indeed that some of these maps (especially the TMNF campaign maps) are classics, which means that there is a line of speed records going back 20 years, and it's hard to throw away that tradition by saying "with today's techniques these are now bad maps for speedrunning competition" even though it's arguably true. People still want to speedrun the old maps. We can't just replace them with good maps now and have the records on the new maps be as meaningful.

[@R3BootYourMind]

In some more complex maps this could be used to achieve a shortcut consistently in the start before driving the rest of the map normally.

[@gwentarinokripperinolkjdsf683]

@@BrooksMoses IMO it's just another step in the story of those mapos

[@Sypaka]

This is the reason, why I use checkpoints as much as possible. Even a simple ring checkpoint can do wonders, because it forces the player to jump into a direction and angle I want them to go. And yeah, I build scenery as well. And hidden eastereggs. Because.. why not.

[@SimonVaIe]

​@@R3BootYourMind frame perfect input tricks with a consistent setup as part of a speed run aren't exactly new or rare, and some have been found through a tas as well, so I'm not sure I'd consider that as a big problem tbh

[@Xeractyll]

Another TM video from Karl. It was inevitable. As someone in the TM community, this whole LIS situation has made me lose my mind for how much people have been chatting about it. I've never seen an issue with it, but the reason of "it's harder to detect cheating" make perfect sense with it tbh. Just really sucks. Though I hate the rule of "no metronomes" that TM has.

[@luis241661]

interesting. as a musician i’ve always thought of game inputs as a rhythm. i’ve taken the time to use a metronome on some games as i just couldn’t figure out certain sequences in games. playing a song accurately and identically every time requires repeating “inputs” repeatedly until its muscle memory. i play games in the same fashion. practice certain inputs until its muscle memory, put it all together and finally play it all out as a “song” without mistakes.

[@adorablerainbowfist]

Reading this title my first thought is "What did Wirtual do now?" 😂

[@d.kk.d8667]

Trying to fin dd2 🤣

[@Fataha22]

He finally remember his password 🤣

[@Orinn000]

At the highest levels, any deterministic game becomes a rhythm game. There are 2 skills that any game needs to speedrun it: plotting the route, and executing the inputs. The "skill" involved is in both, but both are mainly memorization, followed by muscle memory. This just seems like the same argument we see over and over when a new strategy proves to be better than the one people have spent years perfecting: it always "crosses a line" and "goes against the spirit of" something. The spirit of speedrunning is always at odds with the game "as it was meant to be played," and while new categories are normal, this seems like a nebulous distinction. Whatever line you draw, people are going to go right up against it. 8 inputs or more? Expect all the WRs to be 8 inputs. Go to 10? Yeah, that's exactly where the WR will sit.

[@ciCCapROSTi]

Speedrunning inevitably marches towards removing the fun out of a game.

[@jakobjas4212]

Agreed. Speedruns that are someone just playing the game incredibly well and quickly will always be more entertaining and fun.

[@ciCCapROSTi]

@@jakobjas4212 hence why I adore Quake speedruns so much. You can break the game physics, but not much on the official campaign. It's pure skill, and what skill! So much so, even TAS runs don't look entirely alien. Also Elasto Mania. Though TAS breaks the game completely, the human competition is crazy.

[@AvariceNight]

I wish them all the best, but I fear the Trackmania community is committing to an impossible to win battle against this. It's the fate of all deterministic games that fewer and fewer inputs will be needed as records are perfected for the simple reason that the amount of time they have to hit those inputs can only go down. If the best way to play a map is with 6 inputs, the record will gradually trend towards it no matter what preventions are put in place. The preventions themselves rarely ever stop the actual issue the community is targeting but instead targets actual players and burns out moderators on the constant drawn out audits to ensure the record was legit. I hope I'm wrong, but I can't list how many times I've seen knee-jerk reactions to things like this among these communities cause them to completely implode.

[@bipolarprobe]

I disagree with the approach taken to moderating this. Memorizing frame perfect inputs is a totally normal and valid approach to short difficult tricks in any speed game. The records should be valid and you can have a separate leaderboard for non LIS records. It's not at all unusual for degenerate strategies to alter the way speedrunning a game works until it's basically unrecognizable and communities to then say, yeah that sucks lets make an alternate category that bans that. Glitchless/no major glitches become the default category for some games because people prefer to run them, but they don't replace the bare any% leaderboard.

[@Temaran]

I don't really see the problem. This is just speedrunning for any highly deterministic game. The closest parallel would be something like mario kart on the Snes, or many retro games in general.

[@Uragaan1]

The problem is that it will no longer be "competition" or "interesting". Like if you want to see the best chess moves possible played every single move u just let stockfish play against itself. It's certainly interesting to a very small set of ppl, but that's not why people play or watch chess in general. It's the human factor that's interesting. You minimize that to this degree and the community will just die out.

[@GsrItalia]

During F1 history, sequential gearboxes were introduced (1989, if I recall correclty). For someone, was not real "shifting", because the actuators were inserting gears while driving. But was faster than manual shifting, so stayed. Then arrived pre-progammed downshift: at that turn, a driver could simply press shift down, and the gearbox and engine will automatically shift down to correct one as fast as possible (first separate buttons, after time-maps for knowin where you were at that laptime more or less). Pre-programmed downshift disappeared, seamless/quickshift gearboxes are still there. So. As F1 describe: if you're not cheating you're not really trying.

[@SylveonSimp]

F1 is not competitive anyways considering his definition, because the drivers each have different equipment.

[@lairdofgarscaddon]

This process has happened to all sufficiently old enough video games. It's the only way to optimize. What the devs should do is what all the best games have done: release regular updates that make the game more complicated and interesting.

[@taliker]

Isn't this just a Loki's wager problem? If it's 8 input minimum people will do 8 then if you change it to 10 they'll do 10 etc

[@itscutaru]

Pretty much, yes. And on top of that, there's no guarantee less inputs would translate into shorter times. There might be a 12 input route that gets a shorter time than the minimum required 8, and you'd have no idea whether it was a LiS or the player manually grinded the runs. Effectively, solving nothing.

[@codeyamos]

10:45 "Players could use a hidden extra monitor to show the visual inputs, and you would never know." Yeah, unless you were to say, read something off of it. Or repeatedly, and blatently, look at it.

[@charlie2105]

If this isn't allowed should we not allow the use of TAS to learn new techniques at all like we already do with other games? Maybe all that's needed is a better way to verify a legitimate speedrun without the assisted tool. Or have a separate category for third party verification.

[@eessndjd]

Using tas tools to find new tricks =/= using a brute forcer to make a racing game a rhythm game You don't use tas tools rta if that's what you're saying

[@brennantmi5063]

Remember glitches are allowed in any% runs of trackmania under the condition the glitches don't have consistent set ups and require grinding for rng as that is the spirit of the game smh

[@YourFavouriteDraugr]

All that's needed are better designed tracks and new installments into the franchise.

[@siosilvar]

​@@eessndjd You can't just say those aren't the same thing when they are. The only real issue here is the plugin that displays the input timing.

[@noobzerg1990]

@@eessndjd”making a racing game a rhythm game” I don’t see how this is at all a problem. There are optimal ways of playing just turns out that the optimal way of playing this racing game is by making it a rhythm game. Thats the point of speedruns. It always has been finding the easiest solution to get the fastest time. If it means 7 frame perfect key presses over a 10 second run. Then good. That’s the optimal solution. Simply banning a way of playing because “I don’t like it” is ridiculous.

[@asdt560]

It's a track design problem. If a level is so simple it can be beaten by single digit inputs that's not the player's fault. It's the designer's. If you want world records to be hard, then the challenge must adjust to the skills and strategies players have.

[@themidro98]

In the end of all. All speedruns are just trying to find the most optimal button presses for the fastest time. I think the LIS should count as a WR.

[@BeefinOut]

That's assuming the joy of speed running is seeing a number go down. For some people, that may be true, but for most people, the joy is in the routing and execution and community. Tons of huge games have a super fast category that's largely ignored in favor of longer, make restricted categories, because the fast ones just aren't fun or competitive. Look at Pokemon Yellow; the any% category beats the game in a minute using save corruption glitches. It's extremely fast and wildly uninteresting, so no one plays it. Allowing LIS records to count for official Trackmania times is akin to making the Any% pokemon yellow category the only category, eliminating the need to learn the things that make the game fun.

[@siosilvar]

​@@BeefinOut That's an incredibly skewed comparison. LIS doesn't eliminate any of the learning or fun of TM when there are dozens of campaign maps and tens of thousands of community ones where it's completely non-viable. It obviates... A12 which is basically just grinding for the lucky uberbug, the first few seconds of D10, and not much else.

[@themidro98]

@@BeefinOut I agree with you, but in the end LIS is the fastest way a human has compelted the track, so it is the any% WR for the track. Of course they should make a seperate leaderboard with different ruleset which rules out current LIS practises.

[@cloaker2375]

"so beautifully well made" the 450,873,124,512 bugs, glitches, and exploits that help make competitive TMNF both interesting and sometimes infuriating (I HATE RAMM BUGS):

[@JaqueefiusBrown]

teenage mutant ninja turtles

[@JMonkey222]

​@@JaqueefiusBrown*Teenage Mutant Ninja Furtles.

[@JaqueefiusBrown]

@@JMonkey222 mb

[@sukasuka7292]

@@JMonkey222 Teenage Mutant Ninja Furries

[@svk_zoltan]

It should be allowed. LIS is not a game breaking strat, its a track breaking strat and tracks that are broken by LIS are already broken. Great example is A12-Speed, you don't need much skill to smash into a wall over and over hoping it will shoot you into the finish.

[@Verlisify]

As long as nothing is modding or externally affecting the run it isn't cheating. Many speedruns even have buffer strats to achieve this perfection that are found through TASing then become humanly possible. Some levels in games are trivialized to a couple of frame perfect inputs. Dunno why trackmania gets treated differently