Does these bug fixes include new ones? Or are you actually starting to look at her? And by “look” I mean “discuss buffs for her” - I do not mean “laugh at her winrate dropping” like you’ve been doing the past 2 years.
There are two things that we say a lot and that I’m not sure you understand.
-
Your predictive algorithm is bunk. Just calculating expectations of a win based on SR is not grounded in reality in Overwatch. It creates a lot of volatility in the results. Take one example: A team of Mercy mains at 2900 SR vs. a team with 2-2-2 (Tanks, Supports, DPS) at 2900 SR. Your algorithm predicts 50% chance of win for either team. It’s wrong. That’s just one of many, many variables. Your algorithm does not take into account that there is a margin for error and that the results themselves might not be meaningful.
-
50% isn’t “perfectly fair”. What’s fair is the better team winning. As long as you think 50% is the goal, you will code things with this incorrect view of reality. Your job is to stay out of the way, not dictate results. It doesn’t mean you have to allow the best players to beat up on the worst players for 100 games in a row, but deciding that everyone deserves an equal share of wins is making a statement about those players’ skill that is unjustified and unsupported. It’s also what’s ruining the game for many. You have to find a better way of differentiating the tiers.
If the premise of your entire ranking system is about fighting back against an algorithm which is trying to make you share the wins equally among everyone (50%), then how can you be surprised people get angry? That literally means that when you rank up and beyond each tier, you are not just winning against the players of that tier, but against the algorithm which for a while at least will have been trying to give you more losses so there’s enough wins to go around for everyone. Your quality of matches, as a Diamond player who just got out of Platinum, will have been tougher than the average Platinum match.
I’m not saying you are coding it so it thinks “10 game win streak? Let’s hit them with 10 losses!”, but you are tuning everything to equal out to 50%, and they are equaling out to 50%, so that is the end result, however it is being done.
2 Likes
I tried this, and I ended up losing 850 points. Thanks, Scott Mercer. /s
Don’t worry too much about SR. Have fun, play to learn and improve, and your SR will rise over time.
If you meant this, you would have made it like StarCraft II or something similar where you can only see your tier (Bronze, Silver, Gold, etc) and MMR (not ‘SR’) is visible, but only to you.
If it’s just a meaningless number, why is it so big and shiny and prominent with this extravagant animation to show it going up? Come on now, be honest.
Also, if you had separate MMR per hero per distinct group, equivalent to StarCraft II, things would be less toxic.
You might have heard of StarCraft II. It was Blizzard’s main esports game before OW. /s
You guys are moving backwards.
Continuing the discussion from Overwatch PTR Patch Notes – June 5, 2018:
direct quote from the notes
FYI, the new Looking for Group features do allow the option to specify the requirement to use voice comms or not to.
There is no direct influence, but you will likely be playing as the highest rank player, so there is a change that the opposing team will have an average of less than your own personal skill rating, which means you will gain less when you win, and will lose more when you lose.
That is actually a very good idea to have “Come here for healing” instead of “group up” for healers. I hate having to point at a moving target and tap the heal button.
Just clarify, if a group leader creating a group in the LFG turns on the voice comm requirement, this does not disqualify you if you don’t have a mic or simply don’t use your mic. It simply requires voice comms to be turned on.
My only concern is lack keys or space on the command wheel for additional game directive voice lines.
PTR patches are usually on Thursdays, but live patches are usually on Tuesdays (but not always for either).
That feature has nothing to do with this discussion technically, but the last offical answer on this is here:
Greetings Scott! Thank you for posting to your community!
I would like to address a couple of issues here on Match Making System(s). Both Comp and QP.
Quick play match making is needing a repair. We are mainly talking about Account time experience being the primary factor here. I have read threads on here about players that feel they are good at FPS games, yet they are not entirely familiar with how to play the game Overwatch. My friend and some of the other friends I tried to get to play or come back to Overwatch have left. Because they are good at FPS games, but do not understand quite how to play Overwatch, yet they are being put in with 400s, 600s and 700s at times. To them, this is intimidating and other players that are of the higher levels are flaming them for being there in the first place. He now will not ever turn on Overwatch again.
My next concern is on Competitive SR. Something that you may already address or not address here. That matches which turn into landslides should have little to no loss on SR when it comes to comp match. I just had two comp matches in a row last night, which were both landslide loses. I switched to anything possible to help my team and the match was done shortly after it started. I THINK you really should consider the clock to be a factor when it comes to the rate of decay/earning for SR points. So that way it is not something that promotes TOXIC repercussions for a bad match. We are all humans yes, so with that being said. Not every time, can even the world’s fastest and elite computers could ever factor in a few variables. Your system may have given myself and my team that 51% chance to win. But it turned into a 10% chance of hope.
Thank you for reading this post and for the rest of the members of this community. Thank you too!
While snippets of data are good, you’ve only posted means, which are kind of useless without the corresponding variances/standard deviations. It doesn’t really matter if your stats are telling you the win rate from the matchmaker is 50% if the stddev is +/- 35%.
Put another way, it doesn’t matter what the mean success% is if the perception of the player-base is that the matchmaker sucks in the sense that it creates an equal number of steamroll wins and steamroll losses, which is my (and others’) anecdotal experience.
Indeed I would argue the primary metric for determining the quality (or lack thereof) of the matchmaker is the variance, not the mean – the majority of games should feel like both sides had an equal chance of winning but as a player this is not what I experience - there are just way too many steamrolls, which I believe is what is one of the major causes of the rampantly toxic nature of your game.
So your analysis, while appreciated, is fundamentally flawed.
Regarding SR gains/losses, there is an obvious solution I’ve never heard discussed, and that is just to be transparent and show the players the breakdown of the SR gain/loss. eg: player loses game, loses 31SR, but the game tells you: 25SR base loss + 6SR loss because you underperformed vs your past performances.
I don’t think you can “blame” the player-base for rampant misconceptions when you don’t provide us with the info we need to make sense of the outcome.
3 Likes
Quick play has its own hidden MMR system which is separate from Competitive, but by definition, it is still a type of skill rating. In all game modes experience level is not a true indication of skill (look at my profile as an example). So there really is no need to be intimidated by higher XP levels, because it really does not mean anything. Besides quick play is a very friendly and causal game mode.
Scott Mercer has address this in the past on the older forums in this thread.
“Fair” matches doesn’t always mean that every Ilios match goes 3-2 and 100-99 on the final point, or each team gets the payload to the end in overtime on Dorado, etc. Sometimes when two evenly matches teams play, the result can be one-sided. It just means that at that single moment in time the enemy team played better. It’s not always the matchmaker’s, your’s, or your team’s(!!!) fault that you got stomped.
1 Like
This post is WRONG at gm.
6 player groups are impossible to find balanced matchmaking at higher levels, leading to a consistent “advantage” for the 6 group, meaning they gain less and lose more, but have a higher win rate, “52%” doesn’t matter if you’re also only gaining 2/3rds as much, however the win rate is also much higher than 52%.
Duos are the most overpowered from the matchmaking perspective at high ranks, this has been known forever. Attempting to group with any more than 2 is just lowering your overall odds.
BTW, solo is also a ton worse at higher levels, 49.94% only applies when everyone you’re playing around isn’t constantly abusing duo queue.
It was a nice numbers fluff piece for grouping, and it’s true if you’re below gm you should be grouping. And also, AT gm, you should be duoing/trioing over solo queueing.
But grouping and matchmaking are the antithesis of each other, especially when you don’t have personal impact performance. SIGH. We’re going in the exact opposite direction of good balance, but it’s been terrible for so long it doesn’t really matter at this point, i guess.
Hopefully you guys follow this up with a special queue for top 500s/GMs. Grouping at GM is the exact opposite of what you want happening.
Though it’s not even close to priority compared to removing hanzo.
First of all, I apologize in advance for the long red that’s coming next and for my grammar, English is not my native language, so with that said, I really need to understand something and based on my personal conclusion, I can say this:
The SR system is flawed and inconsistent, or the matchmaking is purposely stopping people from climbing.
I understand that by this point people would be thinking things like “you deserve your rank” “git gud”, hear me out for a moment, I’m not saying I deserve I higher rank and I understand that no matter what your rank is, you should be able to get into a certain SR ranking where you would need to improve your skkills or game knowledge in order to continue climbing, according to the Blizzards system, however, this is not the case and here are some data that I’ve gathered through 4 seasons of ranked play, let me show you.
I started playing 4 seasons ago, and after the first 3 seasons, stuck in low gold and having the thoughts of “I feel that no matter how hard I try, I dont seem to be able to climb out of here” , that’s when I decided to purchase an alt account and start from scratch.
After purchasing my alt acc and leveling up to rank 25 (playing again masters and high diamons on qp) I started my placements and the results were, to my surprise and joy, way better than I could have expected.
On placements, there was a point where I was playing against mid diamonds, but was not able to win those 2 game so, that would tell me that Im nor Diamond material yet and I thought it was fair, so aat the end, I got placed in high plat, couple hundred points away from diamond. I strted to think, well, this is where I truly belong and where I need to start getting better in order to climb, but as of today, I have been consisntently playing around the same SR rating, around 2700-2800.
However, here is where the tricky part starts, when I go into my alt acc (already mentioned it is around 2100 SR) one would think that I should be able to climb my way up to kinda the same SR I have in my high plat acc, however, there are 3 or 4 games where I can totally destroy the other team and then, all of the sudden, after those 3 or 4 wins in a row, I sget on a loosing streak that gets me into the same SR of where I started playing, and I onder why, that’s when I started noticing some peculiar things and here is when the “Perfect 50% mtchmaking system” comes to mind.
I have to clarify, I’m a pharah main and thats what I use the most to try and climb, but after a win streak, I’ve noticed the same pattern that decreases my chances of winning a game, despite me being able to maintain a 50% wr on a higher acc, which are these:
- The other team has obviously a smurf, someone who is <50 lv and has some kind of skills totally outside the current rank (doesnt matter if it is gold or plat)
*I get a thrower or a leaver
*I get a team full of people who only plays a certain class
*People who’s seasong high was diamond or higher, but for some reason are in gold playing
But as if that as not enough, that happens on both sides, fter loosing 3 or 4 games and a row and end up being with pretty much the same SR as I started, then thing seem to flip to my side and
*That thrower or troll that was on my team for the last couple games, suddenly appears into the opposing team
*I get a team where we have players able to fullfill a 2-2-2 comp
*As a pharah main, I get people on the other team that has never gotten close to a heroe that is able to counter me (Hitscans, mccree, soldier, widow, etc) you can see them trying, but you can tell they are not skilled enough to harrass me enough to stop me from destroying their team, let aside if I have a pocket mercy.
So doesnt matter how hard I try, after winning 3 or 4 games in a row, I know its my time to lose and end up being unable to climb to where the system has placed me in my alt acc, high plat.
Thins get even worse if I try to group up, it doesnt matter what kind of q I jump into, wheter its duo, trio or even a 6 stack, things get more obvious. After winning some matches in a row we get:
*Teams with 1 silver and 2 diamonds, low diamonds, but at the end of the day, still diamonds.
*People who’s season high was way higher than high plat
*People on my team tht disconnects or throw
And the same scenario applies to both accs, but I decided to pick my gold account as an example and focus on that one so that people can’t say things like " you deserve your rank, git gut and you’ll climb" which is obviously not the case considering I can play consisntently in high plat using the same heroes.
My conclusion and I now I might be speculating here, but I really think this “50% winrate matchmaking” is not fair at all and you guys are not only using SR into account to measure the odds of winning a game, and other couple of reason you brought up on this thread, such as the odds of teams and other things like that.
I do not understand why you need to force a 50% wr on people within the same SR (lets average this at 100 SR) where if, your sytem worked properly, you would not need to do that because people within the same 100 or even 200 SR difference should have the same odds of winning or loosing a game, at the end of the dy, that is where your system placed them and if you are not in that rank, you should be able to consisntently get wins and eventually get to a point where skill is beyond you and you are back to 50% wr, because you need to improve in order to continue climbing.
This is obviously not the case and I think the system is forcing a 50% wr based on other factors such as your opponent’s wr, the heroes they play, hell, even considering putting a thrower or a troll on your team, I dont think its very hard for Blizzard to identify someone who has been throwing last 10 games and then place it on your team.
So yes, I do not think this is fair, to me or to ny other person who is willing to climb, because no matter how good you are, you will get to a point where the system will forc you to lose using external factors and not only your skill and SR, in that case, I do not see the point of having an SR system.
Just to wrap this up, I apologize once again for the long text and I undertand that there are things that may go over my head and I havent really thought about yet, so I did this in the hopes for someone to be able to explain exactly what is the reason behind this happening and trust me, I would love if someone could make me change my mind regarding this.
First off I want to say that I will be using LFG. I have very excited for this feature. The quality of games have been great on PTR (when there was a population). And despite being closer to 50%, I will be doing it.
From what I can gather:
- You say group size does not affect SR gain/loss.
- A solo group, while it tries to make it 50%, is likely to be less than 50%, making gains higher and losses lower.
- Groups are closer to a 50% chance and never dip below 49%.
Which tells me that while you are not directly making groups gain less and lose more, it can happen indirectly with win chance being “solo < group”. Which hopefully will happen less with LFG because there will be more groups.
My biggest question is: How is win chance determined? SR? MMR? Other factors? I like other posters here have experienced great win streaks only to be beaten down by odd circumstances. On the win streak you have people willing to work together, while on the losing streak you have people who can’t aim, won’t cooperate, and will blame each other… Which kind of tells me win chance is determine by much more than SR…
However, having LFG I’ll be able to avoid all the bad stuff, even if it just happens because of bad luck - so I’m excited about that.
it’s really sad to be that one player in 6 vs 5+1. You do bad (or team think you are) and you get 5 reports instantly
why the hell you calculate my possible win chance tho. It can’t be accurate, people are not always going to play their best. They might even throw. And if your basic win chance is high, and someone throw and you lose - you lose massive amount of SR
I am just going to tag these two parts but after reading your post I think it is important to remind you that your skill rating is merely a number between 1-5000 and cannot and should not account for your playstyle or the external factors that you mentioned. Honestly, what do you expect if not a 50% overall win rate after you reach a certain skill rating. Again I go back to my Competitive Deathmatch example, I peaked at 2700ish, but leveled out at the same rank range of gold/plat border that I typically float in Competitive today and I had no one to blame but myself.
In most matches, it is my observation when you are losing what can go wrong will go wrong, and matches that you win, there are no problems.
I have never had this happen to me when I solo queue and get a 5-stack.
Remember that win percentage is based off your skill rating, and without we would have very inconsistent matches… like Season 1 matches when everyone was still equalizing their elos…
The mystery, revealed!
So, let me get this straight, If you win less than 50% of your matches but are matching your win rate, your SR doesn’t go down? WTF, kind of logic is that?
What if you were placed in a rank lower than what you should’ve been at and am dominating and routinely winning like 60% of your matches? Do you have to win more than that to climb in rank? That is incredibly stupid if that is the case and it would explain why it’s hell to climb in ranks.
Other than performance based SR, this would also explain why climbing in ranks was such a pain in the butt. I was winning around 66% of my matches so I was never really climbing in rank even if I was getting gold medals.
It should be that if you win less than 50% of your matches, then you should be dropping in SR. If I have a 40% winrate, then I should be dropping in rank until I start winning again and the win rate goes up to 50%.
I don’t know if someone replied to you, so sorry if I’m just saying the same thing as someone did. I guess the reason why currently it takes a lot of time to find a game when grouping is because most of the players don’t group up, but after the next patch a lot of people are going to be grouping up, so queue times are probably going to be better. I guess solo queue after the next patch is going to have longer queue times.
Yeah, the voicelines for it are in the game, I’ve heard them from time to time, but I haven’t been able to really pinpoint a consistent trigger for them. Probably on YouTube somewhere. I’ve personally heard Mercy and Ana say it on quite a few occasions.
I’d post a topic about it in General, but I’d probably get a lot of people saying “just use mic lul” and I’d rather not have to go through the effort to explain that some people just don’t use microphones. 
Levels doesn’t have to do with skill. As they have said before, it just show how much you have played the game. There’s people with high levels in low ranks and people with low level in high ranks.
1 Like