Hearing Gavin talk about board games, and how he’s passionate about making a game for people “like him who like the same things.” I love it! Great to put a name and face to some of the brilliant minds behind the scenes. Wish more devs would do this!
@Silver_Brush17 минут бұрын
W human
@eldenmeow512344 минут бұрын
Fucking idiot devs. Just balance your matchmaking. Clearly no normal person likes the way you’re doing it.
@ovidiocorchero2795Сағат бұрын
my dawg chillin in the basement
@asymmetricfuzion970Сағат бұрын
I teach on a creative media/event management course that focuses on esports. We have a huge overwatch community. We have 2 teams entered into a national tournament in Overwatch called British Esports student champs. We do an entire unit on esports skills and strategy analysis. We have top 500 players on our team, we have top 50 players who coach. We all play a LOT of overwatch, know a lot about the game and we will put together in class balanced scrims and it’ll still sometimes be a stomp in one direction. Sometimes the strongest team will be stomped too. Sometimes it’ll be a stomp in one direction then next map it’ll be a stomp in the other direction. There’s so much that goes into a “win” in overwatch it’s next to impossible to balance for any match making system. Just kind of is what it is. I get frustrated by it too
@sirrslipp2 сағат бұрын
What about 5 stacking? Even in narrow groups sometimes it feels like they put us against a group of 3 and 2 that are higher elo to balance out the “lack of teamwork”. It feels way easier to duo or 3 stack than it does to 5 stack for some reason idk maybe it’s just me.
@Spluge_McDucktv2 сағат бұрын
Morgan turned me into a Newt!
@ALBEverything2 сағат бұрын
1:05:18 "This commenter doesn't trust us" Uh, ya, its a pretty widely consistent stance to not trust Blizzard. Didn't use to be that way.
@EchoYoutube2 сағат бұрын
Something that was really messing with me was understanding her positioning.. that advice about not pushing any engagements on corners or 1v1s is a godsend.
@CSM-I0I2 сағат бұрын
I feel like I've heard Gavin's voice in one of my games before
@RowanokeWoW3 сағат бұрын
I’ve never once thought the match maker was forcing 50/50s, if it was then how are people climbing? The only issue I have with the matchmaker is I wish they get rid of the modifiers. Go back to a flat amount gained or lost. There’s nothing more discouraging than going from a game you’ve lost and you lose let’s say 22% you win the next game but it’s expected and you only gain 18%. The discussion of player psychology and player ranked anxiety is something they’ve tried to tackle before. That’s how we got the whole debacle at the start of OW2 with the 7wins rank update. The Ow team tried and I commend them for it but it didn’t work. This is the same deal.
@Demystificator3 сағат бұрын
I wish I've been there cause its an interesting interview. Very insightful. Thanks to them for explaining their work. I do believe them when they tell the Matchmaker is summing up the MMR of players and try to guess the outcome. Yet, on the game screen, you sometimes see like a whole division like Gold 5 - Plat 5 game. Does it mean one Team is G5 and the other P5 ? Or is it a min-max players MMR showing up and the teams sum is actually closer. (I assume from that interview, its min max players but I'm always afraid to not getting it right). I also wonder why winning a "40/60" game would only grants 20% bonus in MMR. I feel it could go up to 50% (winning 30% max instead of 24%). I feel like 4% max is too little. For 2 reasons : the number of games to actually rank would be lower and smurfs would climb faster. The other reason is, that if you win a 48/52 game, you don't even win 1% bonus even if the opposite team has actually 10% higher MMR than yours (52 ~= 1.1x48)... In terms of OW1 MMR, coming above 2000 MMR, it would mean beating 200 higher MMR isn't worth more than beating a 200 lower MMR... Maybe the problem might come from players roles rather than Matchmaking and maybe Matchmaking could take in consideration the "carry power" of Supports or DPS cause sometimes, Support players feel helpless if they have worse Damage players. It asks to be more aggressive than supportive, which is kinda weird imo. I'm curious if they have data about that : putting Better DPS with lower Supports and the opposite, and same with tank ; in order to find a way where matchmaking would ask for more specific role goals. Cause if the point is just about winning, that would mean, Zenyatta would be better pick than Mercy cause it has more carry-bility, right ? I'm also wondering if they run some sort "simulation" with just numbers and some simple "behaviour" like "tilting players" / "focusing as losing", etc... ; and just try their ideas on that simulation or if they try directly on the live game.
@kidpresident_14753 сағат бұрын
I guess I'm just a gold player afterall. No more copium Edit:although, if I didn't have such bad social anxiety and I could find a group of people on Discord to grind with every night, I bet you I could hit Platinum again.
@tjtj0123 сағат бұрын
43:23 tbh though while kda doesnt affect the match much, your number of deaths do. What if there was a contribution modifier based on number of deaths to the average of the team. As the better player will probably die less. Not to mention if youre dead you arent contributing to the win
@Anewlevel00914 сағат бұрын
The fact that they are willing to come hang out and talk about match making is proof enough for me they care. Especially compared to games like cod and their ptsd conspiracy theory inducing match making.
@DannyMertens-u7c4 сағат бұрын
If matches so tight.. Why so many games that feel unwinnable and completely mismatched? games you end up with in one team a Rein using good use of corners and firestrikes etc... and in the other team a rein that charges all over the map and dies constantly alone in the enemy backline... Shouldnt they be both tanks with about the same MR? Especialy since tanks have so much impact in the outcome of the game.. I understand that sometimes someone else playing someone else account even in ranked.. but it just to many games like one tank does not even know the basics of the hero he is playing and it is another easy win or fast loose game.
@tylerbacoka42484 сағат бұрын
Why didn't you ask morgan why he looks dead inside?
@kollayder65545 сағат бұрын
"80% of the matches we make tanks are 1 division apart" 30-40% of my games in mid diamond is against master tanks, a lot of wich gm\top500 players who get into masters rank somehow. I literaly played agains top80-100 gm1 tanks in previous seasons, who where in my diamond lobby. How come they fall so low? Maybe i dont play good enough to win more games, but jesus, wining washed gm players and still being diamond is kinda ass to be honest.
@TheSassyman5 сағат бұрын
Bring in a "Diff" modifier that IS based on stats. If you get a certain percentage of stat difference from the role on the enemy team, you get additional MMR increase. If you are performing WAY better, from a stat standpoint, you should get something. I understand that you can't solely rely on stats, but in a large diff situation, you definitely can.
5 сағат бұрын
Stats are always dependend on Hero choice and Team comp. Look at Moira for example compared to a Lucio in a Brawl comp. Both can have insane value, but moira stats will always be better. Moreover if a Match is not a complete stomp, the winning team has in most cases more damage Overall, while the losing team has more healing (because the enemy team does more damage). So therefore the loosing support can look better based on stats but the winning support just could not heal more then the actual damage dealt
@Strange_Heroes6 сағат бұрын
My beef with ranked isn’t with matchmaking. It’s with how ranked is expressed in the UI. It’s not a fun experience. Worse, it’s designed for engagement… like slot machines. I’m don’t in a competition against other players so much as watching a progress bar go up down. Didn’t win? Just pull the lever one more time to win your quarters back. I hate this paradigm in modern ladder play.
@AdamJC88Zeitgeist6 сағат бұрын
I love this! Super helpful. I think a lot of this philosophy applies to some other ranked FPS games as well.
@leipomon6 сағат бұрын
I ended season 12 on a 69% (nice) win rate as a venture one trick. I spent time actually calculating my play style and understanding positioning which allowed me to climb from gold 5 to plat 3. I think I can keep going to diamond honestly
@RayJames-mk3yq7 сағат бұрын
For a game as complicated as OW, performance based MMR adjustments wouldn't be a simple damage/healing thing. It would be a deeplearning AI approach using all the data stored in a replay (which is a lot) across their entire playerbase and having a deeplearning model that can predict your likelihood of going 50/50 in a particular rank, with particular confidence. You would get a big SR boost for every win you're far below the rank the system thinks you can still win it. This is how Valorant's anti-smurf system works btw, and why an Immortal smurf would never spend more than 2-3 games in Bronze, no matter how deranked the account is.
@ВасилийПупкин-з4е7е7 сағат бұрын
I’m already gave up on 5v5 I just waiting when 6v6 be permanent
@ctcpcp7 сағат бұрын
You devs have made the competitive so bad it’s unreal. Especially playing tank. Overwatch 1 had so much better ranked experience even when balancing was off high q time goats. This feels like they have 0 clue what they are doing or how to make the ranked experience fun.
@ctcpcp7 сағат бұрын
Making heroes like Mauga and Sym on top of that, the devs are the ones that should be banned from touching this game.
@SnoozeSSB6 сағат бұрын
Cry in a corner it’s a skill issue
@ctcpcp6 сағат бұрын
@SnoozeSSB Your mum had a skill issue when she created you
@GarySnail-o3j7 сағат бұрын
Easy non-matchmaking ways to stop stomps... bland identical heroes, lots of healing and immortalities, softer counters. Not saying any of those are good options.
@air2data7 сағат бұрын
Let’s be real, this interview was BS. Matchmaking is bad to keep people playing the game with the hopes of spending money.
@SnoozeSSB6 сағат бұрын
It wasn’t BS at all, they gave facts , skill issue all day baby get better
@air2data5 сағат бұрын
@@SnoozeSSB nah it was BS. Matchmaking bad even in higher ranks. Thats why most people Smurf. But I do agree on the skill being a factor part. To climb you have to improve your play to the point where your individual play can greatly affect the team.
@LilTuba8 сағат бұрын
Incoming TED talk: Individual MMR in a team game is already inherently imperfect, but when it is also almost entirely based on win/loss, it is even more low resolution for the individual. That is until you put massive amounts of time into it (they even said 100 matches is where ranks start to look accurate), which I am not a fan of. I think this is a contributing factor to people burning out/not playing ranked. People don't want to have losses pretty much completely cancel out wins, 1.1 steps forward, 1 step back... (numbers don't matter, only the tiny differential between numbers) This win/loss system, coupled with bad RNG, doesn't rob you of your rank, but your time, your time needed to get to a desired rank/maintain your rank. And I would say that it is also is more mentally draining.
@LilTuba8 сағат бұрын
There are two arguments against performance based MMR they mention in this talk. Neither of these are bad arguments but I think they are not compelling enough for me to say that performance based MMR shouldn't be at least tried as an experiment. Argument one is the "Mercy scoreboard problem", to which I would say that the system should be not held hostage by outlier characters, but besides that, I am pretty sure they have tools to compare heroes to other players at that rank, as well as the fact that their is a crap load of info that is not on the scoreboard that they have to work with to judge performance. I have had games where my scoreboard looked like crap by I had the lions share of my teams final blows, constantly making game winning trades: scoreboard problem, not a stats problem, scoreboard is cosmetic. Argument two is how they don't want to have scenarios where the game gives out an extremely harsh punishment or make people feel bad on wins. But to me, having more variation of how a player will feel at the end of a match is actually a good thing. Currently, every loss feels bad, and every win feels good, with little variation. But with performance based MMR, I think having the occasionally super bad feeling loss, will make other losses not feel so bad in comparison. Also I think having this amount of variation means people can't get carried to an inaccurately high rank either because once someone is hitting their limit performance-wise, they would stop having their rank shift much from match to match, with the current system being forced to give a player a certain amount of SR per win(right?). Same would go for the inverse with smurfing.
@LilTuba8 сағат бұрын
I have played a game with performance based MMR, I basically never lost rank, either gained or held in place, and it was crazy how I felt rewarded both on wins AND losses. It felt like the game was telling me I was doing good/improving, even if I wasn't able to win every game. Meanwhile my duo in that game was having, partially, the opposite thing happening, which is a sucky thing about this system but you know what? That meant I couldn't carry that player as easily to a rank they aren't meant to be at, "ruining" matches they play solo in the future. I really wonder how much that sort of dynamic is at play in OW. Our current ranked system does not give a player that kind of direct, "you're improving/maintaining/slipping" narrative, because you either win or you don't, showing modifiers is a good change but I don't think it is enough of a consolation and may even make the player feel even more out of control. It is incredibly easy in OW to ruminate over specifics of a lossed match, kicking yourself over what you could have done differently, because all that matters is the win, which makes sense in something like Chess, but OW? I don't really think so. I have never felt confident in OW's ability to accurately rank people, but in my opinion, more data to work with to determine the rank is better than less data in the chaotic world of online matchmaking, in a game that can have outcomes come down to one properly placed cooldown. I am by no means saying performance based MMR is an inherently better system but I for damn sure think it is worth trying as an experiment.
@swerleybird3 сағат бұрын
@@LilTuba You're perhaps right that pure win/loss gets clouded and slowed down by the other players on your team, but ultimately if you play enough games, this will all cancel out. The only constant in each game is you, the rest of the lobby changes around you every game. I think even with advanced statistics, beyond the scoreboard stats, performance based stats would be a really bad idea. It's really hard to fairly compare all the different maps, players, comps, and heroes which by the way are also changing during the course of a match. No stat can perfectly capture everything you did or didn't do to contribute to the outcome of the match. The only stat that can purely capture win/loss value is the win/loss itself, and with enough games the noise/variation of all the other randos in your lobby fades away.
@ZeroCodeG8 сағат бұрын
Seems like a simple solution is to simply let both teams know the probability of a win for each team, and show total MMR average per team beforehand. Additionally the team with lower probability to win if its lower then 46%, they should lose proportionally less mmr in that match they had the disadvantage. Also, if there is a leaver in the match, it should keep that into account, and have the remaining 4 players lose half of what they would usually lose, meanwhile the leaver gets a 50% extra penalty, that derank amount will be removed automatically, if he hasnt been a leaver in the next 20 games (bc sometimes internet does crash, and encourages toxic ppl to be normal human beings.)
@Jamunjii8 сағат бұрын
I also love getting out into the wrong server and being on 130 ping. small indie game company btw.
@Joshuak2k48 сағат бұрын
I think what you’re talking about is not just the identity of the beginning of OW1 but the entire cycle until OW2 where in 2 everything has more of rounded edges to be able to fit everywhere the entirety of 1 is a bout sharp edges I feel the only disagreement of that rule is sigma and echo but if you think about their introduction into the game it made the game broken which is why I wonder why 6v6 is wanted so much but that’s a different matter
@lknight12668 сағат бұрын
Will there ever be an instance that if someone in your game throws/griefs and is reported (and punished/banned for that), that mmr will be retroactively given back from that single match? Obv the downside would be the winning team loses that mmr afterwards, but I wonder how the community would feel about that
@galelululu9 сағат бұрын
Devastating revelations: Overwatch system developers CONFESS that they’ve RIGGED the system so that in EVERY match, ONE TEAM will LOSE!
@shiekahfan019 сағат бұрын
i played maybe 3 games of OW classic, how can a 6v6 game feel so slow? You would think that adding in a 6th character would help speed up the game but its the exact opposite. And don't get me started on the bs of "you lost because you didn't secure both points", that has always bothered me and its why i refused to play OW till OW2 came out and fixed that issue.
@fractalelf77609 сағат бұрын
The matchmaking is utter shite. Only game I know of where one can easily go on a 6-8 game loss streak, but never ever get the same on wins. Total BS. I tracked 500 matches a few years back, in OW1. What was glaring is the loss streaks far outweighed win streaks in frequency and length. Tired of this rolling nonsense in games in OW2. I’ve noticed one clear tendency in OW2 at least in QP is that if you are a solo player near the top of a tier, you will get saddled repeatedly with a stack including console players vs stacks of PC and expected to carry these guys play. I check every match and it’s so obvious it’s unequivocal. They make it all sound so even and it’s anything but 70% of the time.
@Rando32107 сағат бұрын
maybe if your on a losing you should stop playing cause your tilted.
@fractalelf77607 сағат бұрын
@ Maybe I’ve already accounted for that. You’re also assuming those were all contiguous play, I assure you they were not.
@Rando32106 сағат бұрын
@@fractalelf7760 its not abnormal to go on 6-8 game losing streaks, if it happens consistently the is the only reasonable cause is the player. I doubt you never getting the same amount of wins because if you do consistently lose that many games in a row, without ever going on a winstreak, you would rank down forever.
@fractalelf77606 сағат бұрын
@ What’s abnormal statistically is not getting equal frequency or ranges of wins equally by a large margin. That shows the system is flawed, it is statistically biased. I’m sorry but I know enough statistics to know better, and I am certain if I ran the test again, and others have also reported it, something fishy is going on. I am not saying one cannot climb, but it’s very clear the matchmaking is flawed to say the least.
@Rando32106 сағат бұрын
@@fractalelf7760 What test are you talking about, if you have proof of a fucked up matchmaker post it somewhere. Also if its just 1 account the data is useless. If you want to prove the matchmaker is flawed you need at least a few people and accounts.
@KingQuags9 сағат бұрын
1:16:53 I love the idea of putting toxic players with high block rates in a queue with a bunch of bots
@hearts_drown64659 сағат бұрын
Bro when he said : 60/49 or even 6*/3* would be fair and competitive SAID IT ALL- IF ITS NOT 51-49 it’s not fair and is heavily uneven
@s1lveralkaro9009 сағат бұрын
It's funny they say all this with clear evidence against it
@shawnskyeviews9 сағат бұрын
1:01:19 Just so you all know…they consider telling someone that they shouldn’t do mauga into dva disruptive behavior and is a bannable offense…I wish I was kidding
@mistressfannypack91849 сағат бұрын
Im not sure the comments understand that this book is not an opinion piece lol
@Shnesity9 сағат бұрын
We are so lucky as a community to have somebody like you Spilo, and a dev team that is this open and transparent. What other game has the BACK END DEVS talking to a content creator for over an HOUR!? Insane stuff, super awesome. Keep up the great work guys :)
@marvynjeanbaptiste32069 сағат бұрын
I know this has nothing to do with matchmaking but can you ask the devs if it’s possible to allow Doomfist to actually fight back when he’s in mauga cage. Expecting me to do anything with 4 shotgun shots seems unreasonable. And expecting my supports to help in solo q seems even more unreasonable
@ArielHal90008 сағат бұрын
Just play around it. It's an ult so it's a pretty easy to track and manage your engages so you don't jump him when he has it. It's just one ultimate ability, if it's so overbearing for you at that point just swap onto a different character.
@kennethbontempo103110 сағат бұрын
They should add a double or nothing matchmaking feature. If your team wins you can all opt to play one more round as a team against new opponents. If you win then double the SR. If you lose then back to square one or something.
@libertyprime930710 сағат бұрын
15:35 Idk how he thinks a coin flip in real life is less certain than 51/49. That would mean the coin is pretty heavily biased to one side due to an engraving or whatever. I think the stat he's citing is that the chance of getting the same side again is about 51% just because it's likely a person will use similar flipping technique again which slightly favors repeating the outcome.
@Ron742_10 сағат бұрын
I still think individual performance should matter in how much rank you lose or gain. Why should someone on the losing team who performed so much better than their teammates lose a similar percentage of rank? Punish people who under perform.
@walker561410 сағат бұрын
What is "good" performance? High healing numbers for support means you just healbotted the entire time. High damage numbers for DPS means you just shot the tank the entire game. Sometimes you literally can't take enough space on tank because your supports are Mercy/Lucio. There are too many variables that are out of an individual's control to fairly estimate who was underperforming.
@yuno43739 сағат бұрын
@@walker5614 I think elims/d*aths ratio is something to consider... if support is 2-20, they CLEARLY underperform, because support is supposed to have least d*aths on the team. It means this person doesn't know what they're doing. I've had games where I had 2 d*aths as support (not healbotting) and the other person had like 10-12 d*aths. Why punish us equally for losing? I've had games where one dps in my team did really good job, had good game sense, target priority etc. and second one was acting so clueless it hurt to watch.
@gardenvariety48188 сағат бұрын
41 different characters with different play styles would all need different carefully measured and designed performance indicators, that’s a ridiculous amount of work for something that probably wouldn’t even massively effect people’s rank. If you consistently perform better than your opponent’s you will increase in rank. Get good.
@Ron742_8 сағат бұрын
@@walker5614 I could see it being subjective, but also consider when the devs use metrics to nerf or buff a hero, (when it's not based upon pick rate). Specifically, metrics mentioned here by yuno4373. The devs know/have access to the numbers we don't see. They know what the average stats are for each rank and hero, and they can tell who is under or over performing. They could shape their system to consider when some over performs in a match but still loses.
@Ron742_7 сағат бұрын
@@gardenvariety4818 They already nerf and buff based upon hero performance. They have the metrics. They probably already have target stats for each hero at a specific rank. Would it be more work? Yes. But it would be a more fair, better system imo. If they want to keep individual rank tied to a team's weakest link, then sure, keep it the same, I guess. "If you consistently perform better than your opponent’s you will increase in rank." But that's not necessarily the point here. The discussion is about noticeably outperforming your teammates in a loss, but still losing a similar amount of rank when you have good teammates. Individual performance should matter for rank, even if only a slight amount.
@gregoryj54710 сағат бұрын
That Suzu idea could be the main principle on a new support 👏
@Ron742_10 сағат бұрын
Nothing wrong with them designing a system to make it harder for people to get out of a certain rank. It's just where they think you belong! Just go with it.
@whiteshadowzx10 сағат бұрын
I think hero selection and comp makes the match making feel unfair. The match maker can create a perfect 50/50 based on MMR, but if my mains are countered by the opponents mains, the match will be much more difficult than a 50/50. I might be forced to switch to a hero that I can't play at the same level as my main.
@walker56149 сағат бұрын
Yeah it's kind of why you need to juggle at least 3 mains per role if you want to climb.