What is the Highest Throughput Possible in RollerCoaster Tycoon 2?

[@kABUSE1]

9:56 lmao Marcel you will hate me for this but after staring at the different coasters a good 10 minutes with full brain capacity I am finally pretty sure to have your answer: The only logical conclusion is that the on-ride photo section counts as a customer :D Let me know if this is correct.

[@MarcelVos]

Yup.... This checks out. 18 guests on a train, of which 7 buy a photo and the total customers increases from 57018 to 57043. Wow. If I put a photo section on the test park vertical drop and force the guests to be happy, it can briefly hit 7k an hour before most guests already have a photo, so this definitely is the reason. It looks like the thing with a big park causing lag isn't true and me removing the park just caused guests to buy fewer fotos, which is why the guests/hour went down. That also explains why I couldn't increase the test park throughput by building a park around it.

[@kABUSE1]

@@MarcelVos The follow-up question is: Do you know if guests will buy photos of the same coaster multiple times or will be less likely to buy a photo if they already have one or something?

[@bitemytail]

@@kABUSE1 They will only buy 1 photo per "category", so 1 rollercoaster, 1 water ride, etc.

[@MarcelVos]

@@kABUSE1 There are four different on-ride photo categories (steel coaster/wooden coaster/inverted coaster/water ride) and guests can only ever buy one of each category.

[@thedoormatt6998]

I thought the queue screens were the cause lmao

[@funx24X7]

The lift exit being a sudden drop back to ground level is an incredible touch

[@daninda1665]

No guests were harmed in the making of this video.... I'm disappointed in Marcel... 😝

[@ducatisti]

I literally stopped the video at that point because I had to savor the brilliance for a little longer.

[@jefffinkbonner9551]

Question: could a mechanic always fix the ride via the bottom station exit or are there repairs that require him to access the top station?!

[@AllThoughts3rased]

​​@@jefffinkbonner9551 mechanics always enter the first station with an exit. If you have four stations, A B C D, and B and D have exits but A and C do not, the mechanic will fix the ride by heading to station B. In this case, A and B both have exits, the mechanic will go to station A to repair the ride.

[@dawndefender]

Given that RCT peeps can survive fall from any height, that makes it pretty interesting (and intense) ride. Given that you don't ride it from the top station lol.

[@theplasmawolf]

7:55 the contempt in your voice when you mention guests checking watches, it's just perfect.

[@JLock_17]

12:08 Leave it to the Swastikoaster to be the most efficient vanilla design.

[@MarcelVos]

As much as I hate people commenting on that as it was an accident I can't deny that swastikoaster is some good wordplay.

[@seanmeadows4620]

​@@MarcelVos Rotational symmetry spits that shape out naturally. There was an old Starcraft Brood War pro map that was shaped like one.

[@jefffinkbonner9551]

@@MarcelVos All I see is a lovely Dutch windmill.

[@straightpipediesel]

And all the entertainers are in white costumes 😂

[@JLock_17]

@@MarcelVos Eh, accidents happen.

[@BrendanCS]

3:56 Imagine walking up stairs to get on a ride that just takes you back down to ground level

[@1onehuman89]

This also describes all water slides and toboggan / alpine rides (And probably a few other types I'm forgetting)

[@HappyBeezerStudios]

just enter at the bottom and get pulled up

[@Nessus875]

The elevator ride has the "please don't break my legs line" and the "please break my legs line" for those guests who want some freedom to choose their form of mandatory suffering while in Marcel's park.

[@MarcelVos]

All throughput tests were performed with rain and breakdowns disabled, so that definitely can't be the cause of unexplained difference. Also, even if I max out the guest walking speed by constantly spamming the cheat to give them maximum energy I only get to about 5600 guests/hour with the test park vertical drop, while I get to 5950 with the one in Bumbly Bazaar. It has reduced the gap, but there is still a gap.

[@doc.rankin577]

I could only postulate as to what the cause is but do you think the map/scenario itself might effect the throughput? That's the only real change I could see.

[@waffleweave]

Does the speed of guest walking into the ride entrance change if they are on a sloped up/down queue line?

[@Aurakeks]

I could only imagine it has to do with the intervall of the guests per hour calculation. Maybe the time those coasters started operating in the save games just hit a sweet spot, where the calculation is always triggered at the most optimal point?

[@MarcelVos]

@@waffleweave Yes, which is why all the tests have a flat queue line.

[@MarcelVos]

@@Aurakeks The guests per hour calculation the game does doesn't matter, as I just look at the number of guests that have ridden the ride. We have probably found the reason though, see the new pinned comment.

[@Sythgirla]

4:20 "This is going to be a bit more complicated" >Shows absolute spaghetti of a park

[@user-et3xn2jm1u]

Instant classic from Marcel. The twists, the turns, the surprises. Overall banger of a video.

[@ayrtonpavot3096]

couldn't have said it better myself

[@Davtwan]

Seeing the thought processes for these tests are fascinating, but hearing Marcel call a guest “idiot” was the best part of the whole video.

[@Scream162]

“It’s the guests that are the problem!” Marcel was likely a fellow ride operator at some point in his life!

[@Atoran95]

I seriously love these "how can we improve it" videos, it's like watching someone do a puzzle

[@SirPembertonS.Crevalius]

The 3 tile maze getting 7k+ guests was incredibly unexpected.

[@MarcelVos]

@@shockwaverc1369 Then don't read the comments before watching the video lol

[@pierreo33]

@@shockwaverc1369 yeah well what do you expect? the comment section exists to, well, comment on the video...

[@frankherrewijn2508]

Not anymore because you spoiled the end

[@kingkuma4112]

Don't get lost!

[@jessehunter362]

@@frankherrewijn2508 the comments section exists to steal comments on the video!

[@StriKe_jk]

The most obvious difference between your test map and a real map is, well the map. The test map is empty, while the real map is packed full. Now this is pure speculation but a full map maybe cause slowdowns to the game which in turns means that it is more forgiving to tiny waiting periods, similar how a TAS bot can perform tricks a human can not by slowing down the game to crawl. And because the game runs on an internal game clock this affects the guests per hour

[@MarcelVos]

This seems to be a reason! I removed all the rides and stalls on the bumbly bazaar map and the throughput decreased to about the same as the vertical drop coaster in the test park. What's interesting is that almost all the rides in the park were actually closed during those tests. Only the ones in the shot are opened.

[@MarcelVos]

Nevermind, it was on-ride photos...

[@HappyBeezerStudios]

Either all the stuff simply means more to render, or because more stuff means guests have more to check out.

[@TmOnlineMapper]

"It was on-ride photos" Does that mean that the guests that pay for the photos are then counted twice, @MarcelVos?

[@Destroyahx2]

​@@TmOnlineMapper There is another comment where the mechanic is explained. But yes.

[@DrTarte]

The inner machinations of rct2 are an enigma.

[@KroutonCing]

7:39 “No, it is the guests that are the problem”. Sounds like something you’d say before drowning them.

[@baystated]

Entrance placement at mid-train was the best unexpected self-taugh pre-youtube lesson for myself when designing a coaster with throughput in mind. Any guest never has to walk more than half the distance of a train to find a seat.

[@TheTdw2000]

12:06 now that's what I call German efficiency!

[@Skalatsosse]

German efficiency as performed by a Dutch (aka. a swamp german)

[@eagsh6651]

@@Skalatsosse Er meint das Kreuz :D

[@Skalatsosse]

@@eagsh6651 Ohje, jetzt seh ichs erst 😂 Na wenn das dem ganzen "highest possible throughput" mal nicht nen etwas anderen Ton gibt.

[@MA-naconitor]

Thought I was the only one seing it. Marcel did change it though in the thumbnail

[@shrimpboom8]

Something about Dutch people and windmills. The joke writes itself really

[@danielmckenzie5659]

12:06 The Swastika is the most efficient roller-coaster design

[@theholysoup2691]

Thank you marcel.

[@JustenHarper]

Like, I only played Theme Park for the Sega Genesis about 25 years ago, but I'm mesmerized by these videos. I think it's just a passionate person who is an expert in their field making well-concieved and well-produced videos with an enjoyable voice & delivery. It's nice to just completely immerse for just a minute.

[@Parenthasees]

I just recently discovered your channel, and rediscovered my fascination with Roller Coaster Tycoon. I haven't played since I was a kid and after watching your videos, I haven't stopped thinking about how much I want to play RCT again! I'm going to try to download it on Steam this weekend, unless there are better ways to play it. I never was that good at Roller Coaster Tycoon, but now that I've found your channel, I'm sure your expertise will help me do better! Thank you so much Marcel!

[@Apiem1]

The steam version is the best one. I say dont get RCT Classic just because I think the new interface is rather ugly and clashes with the games visual style... Or maybe thats just me. Anyway, three simple tips for doing better at the game for casual play; make coasters semi-realistic and you wont run into many intensity issues, keep coasters relatively short unless flush with cash, and charge $5 dollars per coaster and $10 for the high excitement (8+) ones. You wont need to do much price management that way but still turn a tidy profit. Cheers!

[@SemiHypercube]

I like how the three-tile maze still ends up being one of the best rides for this, I feel like that design keeps showing up in these kinds of questions (maybe that's just me though)

[@HappyBeezerStudios]

Another thing visible was how bad the space rings are

[@tappajaav]

@@HappyBeezerStudios SPIIIIIIN

[@owenw1068]

Hi Marcel, I think guests with hats or balloons always walk at maximum speed so that might explain the difference in throughput between the test coaster a the park coaster.

[@MarcelVos]

I added some hat stalls and balloon stalls (both in seperate and combined tests) and it made no difference. I can also still see guests walking at different speeds with them on.

[@owenw1068]

It would be interesting to see an experiment with sick and tied guests given hats to compare with happy guests

[@daninda1665]

How is walking speed calculated?

[@upgradeplans777]

@@daninda1665 You might want to take a look at the video "How fast do guests walk in Roller Coaster Tycoon 2?" from this channel. I also think that the walking speed could be the difference here (I have no particularly good evidence for it though). So it could be that optimizing energy en nausea levels helps with the throughput.

[@DarkWarchieff]

Interesting queue path pattern on the final one, Marcel.

[@thekornreeper]

Indeed

[@gigabytemon]

Thank you for continuing to make these videos of your experiments. I have never played any RCT games before, but hearing you say "Hello everyone, and welcome to another video" before you start doing the sciences always makes me happy.

[@voodootrois]

I am always slightly amused by how silly the guests walk on the Dodgems or flying saucers towards the vehicle.

[@hooverkinz]

Same!! They zig zag so fast it’s hilarious

[@jefffinkbonner9551]

I like to think they’re getting electrocuted and are twitching across the plaza 😂

[@hooverkinz]

@@jefffinkbonner9551 omg I will never unsee that now

[@StefanVeenstra]

Also 20 seconds dodgems. By the time everyone is seated they can get up and leave again.

[@Kain8719]

I just stumbled on this out of boredom. I just got to say, i really am loving these deep dives. Work days are gonna be so much better for a while.

[@isaactierney5132]

The sass towards the guests checking their watches got a laugh out of me.

[@XiCynx]

Those queue lines at 12:15 are almost not safe to use for the public eye. xD

[@timgoldberg3333]

Marcel Vos videos are like an episode of Regular Show, it starts out normal and then slowly gets more and more insane until by the end you're wondering how they came up with this madness.

[@noel3700]

12:10 Germans approve this queue design 😂

[@justinshin2279]

12:12 interesting choice of design 😅

[@HECKproductions]

marcel: this is the highest throughput in RTC2 me: really? marcel: ...vanilla me: aaahhhhhh

[@commiecomrade2644]

New sub here. Been binging your videos. You're like the Spirit of the Law for RCT! Really enjoying the content.

[@jefffinkbonner9551]

That comparison has not only been noticed by others, but was actually mentioned by Marcel as an inspiration.

[@El_Presidente_5337]

12:07 AUF DER HEIDE BLÜHT EIN KLEINES BLÜMELEIN

[@DitzyNizzy2009]

3:15 - One of the guests using the loft jumps for joy about being where he was a few seconds earlier but now with a broken leg.

[@Kimera92]

8:02 "it only takes even longer now, you IDIOT" Dude, I laughed so hard at this lmao

[@vincentbaillargeon9139]

I like how far you go in every video. Always very fun to watch!

[@katinabianca]

God damn I found your account my accident and I’ve never wanted to play RCT2 again more for years. Praying I’ll be able to download it to my iMac

[@raw391]

Try a log fume with two stations. station 1 (1 tile long) at maximum ride height using supports and terrain with entrance and exit, then station 2 (as long as possible) under ground with exit only and no entrance, has incredible throughput. Love your video's, thanks man.

[@AlBundy319]

I think it's fair to assume that Marcel know more about RCT1 and 2 then the team who actually worked on the game back in the days. Marcel, you're awesome as always!

[@CheesecakeMilitia]

I always love placing ride entrances at the center of the station platform for throughput reasons. Seeing creations where the guests have to walk down an absurdly long station platform to get to their seats was always a pet peeve of mine.

[@davidromeroblaya7920]

At the end of Deus Ex intro, Bob Page says: "Aquinas spoke of the mythical City on the Hill. Soon that city will be a reality, and we will be crowned its kings. Or better than kings. Gods...". I think it describes perfectly the mentality of a RCT2 player.

[@Zorf96]

Good boy Marcel having the queue lines in the left-facing orientation 12:23. Legit glad you paid attention to that issue haha

[@MarcelVos]

I paid no attention to it, I didn't notice until the video was almost done. I would've redone that bit if it was the other direction though.

[@Zorf96]

@@MarcelVos I appreciate that very much :) it's a silly thing in a way, such a common figure to make by accident, but it's nice to know you care

[@blobskin]

That last bit gives me an idea. A "one roller coaster for an entire park" idea. Whenever you want to build another roller coaster, instead you have to expand the current one, adding a new station or two. I wonder how crazy that could get.

[@JustinSaneTube]

love that in his innocent chase for an optimum design for visitor throughput, one design stood out to others in a completely different manner. but Marcel didn't see this. Marcel saw visitors moving in an optimal pattern. bless him. also, I would've been so disappointed by visiting that maze! I gave you these money and you make me walk between 2 bushes basically?! :P

[@agargaragara]

Excellent video as always. I was a little surprised to not see Splash Boats get an honorable mention. Using a four station design, you can get not only a very high throughput, but also charge a decent amount because only one of the four sections needs to meet the minimum stat requirements. Even without optimizing for throughput, I've seen around 15,000 guests per hour (in-game numbers). Not as good as coasters, but probably the best water ride?

[@MrAmanox]

Death maze and giant swastika processing over 20000 people per hour. Oh God, what kind of a park is this?!

[@RagnarokiaNG]

The 64 guest maze is one of my favourite little things about OpenRCT2, its just nice making a free maze and seeing it full of guests whenever I go over to it.

[@gijstennekes1678]

12:07 weird svastika queue-line.😂

[@pablocau3792]

"In África, every 60 seconds a minute passes. Together we can change this." Marcel Vos: _"I'm gonna rock this man's world"_

[@PeterMetzger]

"No. It's the guests that are the problem" reminded me of Principal Skinner saying "No. It's the children who are wrong."

[@askapk]

Such a classic game. I've put in hundreds of hours as a kid... but I feel like I haven't even touched the surface of what you've done in even a single video. Lol.

[@bula-zi1if]

Everytime i watch Marcel I get into this game again. All these videos make me want to hop back on open rct.

[@DrMedMueller]

Ah yes, the overpowered svastika coaster reigns surpreme. At least in vanilla

[@marcelbelzer587]

Because of you I startet playing rtc2 vanilla again and currently cleaned all Szenarios till "höllische aussischten" im from germany and don't knows its english Name.. The Hell one. Thx for the great content from one Marcel to another :)

[@CadenceCube]

its called “Infernal Views” in English :) Alles Gute!

[@dustinmorrison6315]

This game is so amazingly intricate.

[@Apiem1]

Oh hey! Was this the result of my topic suggestion in the last video? Keep up the fantastic work! Edit: somehow I figured the death maze would make an appearance. Also the guest discrepancy is very interesting, no clue what that could be about. Its fascinating how these kinds of questions always seem to highlight often unnoticed guest behavior or little quirks in the game design or hillarious bugs.

[@Anubis7169]

I got the 4-wide water boats ride up to 5500guests/hr or like 45,000$/hr (IIRC) shortly after being built. 0 wait time, and, even in a packed park, the guests shortly thereafter could not queue on it fast enough to fill it. The ride also had no lift hills for the passengers, and the ride time was only like 15 seconds - 2 drops and then a turn into the exit station. Once the "ride just built" excitement died down, the guests got on it even slower though. Literally can't get them to queue fast enough, even with constant ad campaigns and/or ride coupons. Instant financial win once the ride is installed. Ridiculously profitable.

[@ZlueTheDragon]

Thinking about why the rides have different throughputs in the testing area vs the real parks, you should look at the variables that are different between them. Are the guests more energised? Do the guests have higher nausea resistance? What about food stalls?

[@stevenmorrison4692]

That's what I was thinking because I was reading park tutorials online and certain parks the guests entering have higher or lower nausea ratings, harder to attract, guests are cheaper than other guests for a different park etc.

[@samachu09]

I've never heard of this game before, let alone played it but yet, I keep coming back because you make interesting videos. I might even get the game some day cuz it looks fun!

[@rice2meatu]

Kid me would have vomited at the thought of combining math problems with a cool game like this, but decades later I find it fascinating. Thanks for calculating all of this

[@theykeeponrising]

Marcel: Hello and welcome to another video. It is important to give your guests the sweet release of death, but it can get exhausting drowning all your guests individually, so today we are looking at the maximum number of guests you can drown automatically per hour.

[@WackoMcGoose]

"DON'T GET LOST was great! Kinda crowded, though..."

[@TS_Mind_Swept]

Wait...the DeathMaze™️ didn't have any guests falling into the void... is this really a Marcel video? 👀

[@IllyasArt]

On a roll with the uploads lately :D

[@Hokiebird428]

So I guess the questions is: What ride type and design is the biggest possible money maker in pay for ride parks? You have a massive throughput of 20,000 on that coaster, but probably can’t charge very much for it due to low stats. On the other hand, a coaster with 10+ excitement maybe doesn’t have as high of a throughput. So, where is the balance between the two attributes, and what is the maximum possible without cheats?

[@EthanTheWerewolf]

My brain has decided it's time to binge watch rollercoaster tycoon stuff, so here I am, watching this video

[@esotericVideos]

4:10 lol The lift casually dropping guests off the second floor exit falling back to the ground. It's an elevator that takes you to the first floor or the first floor.

[@JNDeaux]

_"It is the guests who are the problem."_ I've always thought so. That's why the fish at my parks are so well-fed.

[@TheThewulasvegas]

You drew a thingy…you’re park is a war crime! Lol!

[@odenwaldquelle8228]

A comparison between the tower rides, as you did it, in short VS maximum hight wouldve been interresting to see.

[@eenstukjespoorweg55]

Zoals altijd, zeer interessant Marcel. Vooral ook dat je altijd net even weer verder gaat dan realistisch is voor game play.

[@emmata98]

10:34 maybe the guests walk faster or the in-game clock is manipulated due to lag?

[@rdfox76]

And now, just to put these numbers into perspective... in real life theme parks, rides that can swallow 1500-2000 riders per hour are seen as "capacity monsters" and "people eaters" on a par with some of the top roller coasters in the world. Yet in RCT2, you have no problem getting that level of ridership out of *flat rides*.

[@neyvie]

Super fun video. Loved this game so much back then

[@jgfjfgjfhjf]

I've never played this game, never will, and I still enjoy your mad science

[@InternetTAB]

it's not often I see a 1 minute ago upload. your vids are always entertaining thanks for your work

[@antarean]

I feel bad for the people who get on the elevator at the bottom station only to fall back to where they started from when they exit.

[@HappyBeezerStudios]

The roller coaster variations show how important it is to have a sufficient data size in statistics. In that case, doing multiple passes and removing the most extreme results on either end. Getting rid of any possible outliers while also ending with an average.

[@drago939393]

What started as a game about making cute rollercoasters and parks turns into crazy math and non-euclidian geometry lol

[@Songbearer]

Between Marcel and Decino, when time machines are invented I am going to blow 10 year old me's mind

[@wariowario9739]

"It's the guests that are the problem" Open the drowning pits!

[@ducttapeanddreams]

I want to see now "efficiency park" where you get the highest rating you can using the most efficient throughput, and use one ride of each type/category. One tile death maze, super efficient coaster, maybe a coaster where one of the entrances has to use the elevator

[@StriKe_jk]

Try running the same coaster on an otherwise empty map vs a map packed full with other rides, attractions, guests etc to cause load to your machine, the code, the engine etc. The more realistic the better to stress all systems.

[@borisvokladski5844]

With all these high throughput roller coasters, you will become a RCT2 Millionaire in no time.

[@HannibalGraham]

You can make longer queue lines if you put TV's and entertainers. It's an exploit you can use to keep people in the park. You make all your lines able to have maximum capacity with rides with LONG ass times so these people stay in line as long as possible. With the TV's and entertainers, they won't get upset and want to leave the park, so your queues can have I believe 510 people and keep it that way effectively permanently. My point is, they won't be looking at their watches. Your maze is also built wrong. Need the entrance next to the exit and use I believe 4 squares to the exit. It's a straight line and they never go in a circle or back/forth etc. They go straight to the exit.

[@RennacBTW]

"this is the most..." Man I cant wait for the video in 2 years where you beat your own record

[@chrispunkt407]

as the good german i am i emediatly recognize my history here 12:08

[@In.New.York.I.Milly.Rock.]

Who would've thought that the industrial-level-death-iconography park would excel at processing a critical mass of living beings! You're one accident away from a Nuremberg tie

[@dewaard3301]

This reminds me of a recent Doom speedrun that required thousands upon thousands of rockets and enemies to be in the game. No PC could run it at playable speeds, but the in-game timer doesn't care. Sounds like some dedicated fan might actually be able to show this 1M+ number.

[@WaynesStrangeBrain]

I like how in an alternate universe Marcel could be a wealthy quantitative analyst with absolutely no change in skills or personality

[@karremania]

I gues a way to measure is to time each departure of the station. Though we talking maybe 3-5 seconds per train to move in, fill up and start, it has a bit of RNG how it exactly be filled (back to front, left to right) that in a worse case scenario you might end up at the higher 10% RNG then in the lower 10%. No idea if there is any tool out there that can accurate measure the exit time of each train, but it might be worth investing, aside you can more accurate check if the throughput calculated by the game is accurate at all.

[@leonardorocha2244]

Could the unexplained difference between throughput be because of guest speed? If they have different speeds based on energy, or something else, this could make the guests take longer on average to arrive at the station, as the faster guests would be stuck walking at the speed of the slowest ones.

[@MarcelVos]

Cheating the guest energy (and thus speed) to be constantly at the max closed the gap a bit, but there's still a difference (see pinned comment).

[@Wtdtd]

1:57 Haaaang onnn. In game one hour is 48 minutes irl you say. One in-game hour gives 527 guests. This means 48 IRL minutes gives 527 guests. Therefore one IRL hour should give more than 527 guests because one hour is longer than 48 minutes. Instead you arrive at 422. I think instead of multiplying by 0.8 you should have divided. Now you are going from 100% to 80% instead of the other way around. 527:0.8=658.75 (I guess that one guy is missing an arm. 659 if you don't wanna be rude to him ;) ) Right. Now to watch the rest of the video. (I hope this mistake doesn't repeat. Or that I'm mistaken instead.)

[@mm3shadowman982]

"Don't Get Lost" is a meta design I guess

[@PianoWindowYoutube]

Idea: please calculate what coaster type has the most possibilities for lay-out (kind of track pieces): statistical number of options on a specific piece of land.

[@MarcelVos]

I've toyed with that idea before but I'm afraid my maths skills aren't good enough for that.

[@demrandom]

@@MarcelVos i could probably code a program do it with a list of parts and what their xyz movement is in tiles/valid followups (banked piece into unbanked piece) if such a list exists anywhere do point me towards it

[@TyroRNG]

My theory for the higher throughput in normal parks compared to test world might be lag. Perhaps the game slows down slightly but different elements get slowed down at different rates. Perhaps the in game timer runs slower because of lag but the guests entering the ride don't or to a lesser extent causing more guests in the ride per in game second.

[@PhilipMurphy8Extra]

Marcel is the man of RCT2 for sure.