Percentage of wins and loss should account for value added to the team

Competitive Discussion
1

Hi,

Supposedly the match making system works well. Supposedly (despite master level streamers proving otherwise) the system quickly moves you to where you need to be.

Supposedly.

How can this be though if in game value, as in, actual value added to the team isn’t accounted for in every game?

Let assume something for sake of argument.
Team A = 2500 SR
Team B = 2500 SR

as in, all players are EXACTLY 2500 SR on both teams. Now, the match plays out and let’s assume one team wins. What happens to the losers?

They all receive let’s say -20 SR. This is more or less how the system works yes?

My question is why aren’t in game actions, value added to the team, accounted for? If 5 people play their role with 5% of the tier average for the role, and one player is 20% below average for the role/tier … why isn’t the potato charged more for the loss?

two off tanks? that’s fine, majority of the lost SR will go to the tanks and not penalize the poor DPS who have no ability to DPS from protected space and the supports are running for their lives 24/7.

do you actually have bad healers? if they’re below average for the role/tier then they’ll assume a larger amount of the lost SR.

got an idiot hard locking widow refusing to change despite double barrier and 2 beefy DPS? That’s fine, she’ll be taxed much harder for the loss.

did you just crush it, getting all the golds, AND adding value to the team in your role? great, that loss cost you very very little due to the other 5 potatos on your team.

I’m not 100% how to do this, however, i’m 100% positive that unless you account for each game, and what happened within each game, you’re not actually moving people quickly to where they need to be. further, you’re round about encouraging selfish play.

if you change it to at least a partially shared risk pool, i think you’ll see players playing more of a team game by default.

i patiently await the assaults on my rank which somehow invalidate my argument’s premise …

5 Likes
2

performance baised sr exists below diamond so what your asking for is already a thing.

11 Likes
3

prove it.

all the evidence suggests that once you’re where the system thinks you should be, every win or loss (the initial one not the 5th in a row) is roughly 20-22 points yes?

so if i get knocked unconcious IRL, maintaining just enough of my faculties to hit w, a, s, or d every 15 seconds, and my team wins despite being effectively 5v6 … how much do I gain in that win?

how much do my teamates win? again, let’s assume we’re all the exact same SR.

now, this is where you prove it.

2 Likes
4

Like what Alayia said, your stats in-game attribute to how much SR you either gain (on a win) or lose (on a defeat)- if you’re consistently outperforming those you are playing with, or against, you should see gradual SR gains over time. That being said, in games that are below diamond rank you should be getting more SR for performing “above average” regardless of what your MMR is.
All the MMR system tries to do is place you in games where you have as close to a 50% chance of winning as possible- you should still see gradual SR gains if you consistenly outperform, and substantial gains further into win streaks as the systems attempts to correct your MMR placement.
In your scenario- if you’re not contributing to a game then, according to performance based SR, you should receive relatively less SR than a teammate who clearly performed above average for instance. The same can be said for the inverse of that scenario. Does that make sense?

5

the word ‘should’ was used a lot in your post. you and alayia are presuming. you don’t actually have proof and again, countless streamers showing the bronze to masters climb does NOT in fact quickly boost them to where they need to be despite GROSSLY outperforming the tier.

you can have your presumptions and your ‘shoulds’ but reality seems to indicate the system isn’t that efficient.

further, are tank’s and support’s value properly accounted for? this would be a first btw as again, countless threads since release have detailed how supports are always placing lower than DPS as an example. perhaps with so few tanks, the system works better with the limited pool idk.

further, this assumption you’re both mentioning. as i recall this has to do only with the lobby as a whole. if you as role dps you outperform the lobby, you’ll gain more or less.

what i’m talking about is in addition to the lobby, wtf happened in the game itself.

  1. did i outperform the lobby?
  2. within this game, did i add value to the team with my role?

let’s assume yes for both and let’s assume my team lost.

tank1 = performed below lobby
tank2 = average for lobby
dps1 = performed below lobby
dps2 = performed below lobby
sup1 = performed above lobby
sup2 = average

performing above lobby, providing value to the team = 50% less SR lost.
average performance, average value = same as now
performing below lobby, gave no value to the team in the game = 50% more SR loss.

now within a game you can perform below the lobby in some stats but still add huge value. let’s say as rein youre well below damage blocked, but hit every shatter and lots of fire strike kills. well, the enemy was running dive so weren’t offering up much damage to block but since you nailed all your ults and lots of firestrikes you should still gain more SR for a job well done.

i think my biggest issue is that losses hurt more than wins. if i crush it in my role, and the other 5 did not, i shouldn’t lose more points for that loss than the next game’s win potential.

as a player, i should be able to control how much i lose by getting as much value for my role as i can even in obvious loss scenarios. as it stands now, within the first few minutes its pretty obvious if the game is a stomp or a close game. losing 24 points on the wrong end of a stomp where 5 other humans make choices I can’t control is a really bad feeling when the next game i only gain back 22 of those points.

doesn’t seem to matter how i play in each game, especially in losses. in fact, in loss situations it is better time management to just 100% throw. lose as quickly as you can so you can get another game that has hopefully fewer potato players.

4 Likes
6

Because you can’t simply measure individual contribution to a match. If it was possible they’d do it…
And will you be matched in the same team as that widow player again? No, if he continues to play like that and causes his team to lose he’ll end up deranking, now he might be the best widow withing your rank nevertheless and in that case he’d stick around longer since he receives bonus SR

7

All the Grandmaster and T500 streamers eventually place in their respective tiers on alt accounts- it just takes the system a buffer to determine where they should end up. I have seen people on these forums reference that it takes roughly 150 games for your rank to reflect your actual skill level. In this essence, you do eventually end up in a rank that you deserve (no shoulds this time).

In critique of your solution, if people were to gain +/- 50% SR based on performance than people would be experiencing highly dramatic drops and improvements in SR accross a small range of games which would not be beneficial to the community in the long-term. The way the system works now is- you DO climb SR if you are better than your team mates. I think what causes a lot of people grief is that these gains aren’t instantaneous. The belief that I should be able to climb 1,2, or even 3 ranks without having previously earned them is a misconception shared by many in the community and prominent in forum posts. Ranking up and improving takes time, it will never be something that happens in the span of a day of grinding.

Ultimately, what I’m trying to say is that if you experience constants in terms of your SR it might be (is) attributed to your skill level reaching a peak. This simply means you need to invest more energy into your play to see further improvements.

1 Like
8

i don’t agree. you’d see smaller gains and losses as eventually your mechanical skill will put you in a place where game to game, you’re performing average for the tier/role.

the way the system is now gives rise to multi hundred point plummets as you can’t control the amount you’re losing regardless of how you perform.

further, i disagree with ‘reaching peak’. I’ve been as high as 2800 for long periods of time. then, inexplicably, i plummet and then climb to mid gold only to plummet again. part of this was switching from 144hz to 60hz w/ freesync. since disabling all that and running 180+fps nonvsync my aim numbers are climbing steadily, while my win percentage continues to dive.

again, game to game, according to stats, after upping FPS i’m up as much as 10% with some weapons.

my most recent game is a great example. these are my stats from that game versus the lobby according to overbuff.

Lucio:
Elims - 30 top 1%
obj kills - 16 top 1%
obj time - 2:25 top 21%
acc - 27% <-- needs work
crits - 9% top 20%
env kills - 2 top 2%
dmg - 13.7K - top 1%
healing - 10.7K top 28%
deaths 10 - this is higher than the 5ish i usually have
off assists - 15 top 8%
def assists - 20 top 1%
sound barriers - 26 top 1%
e:d ratio - 3 top 9%
medals - 3

so tell me. why did i lose 24 points for that performance? a completely win-able game where one of the two dps locked sym and ran it even after the other team hard countered her with reaper and mei. we got zero value from a dps spot, i provided top 1% value in almost all categories … why do i get equal share of that loss?

it doesn’t make sense. i do NOT believe you can’t score individual performance per game, total BS. this is a completely solvable problem.

2 Likes
9

Because you are the only constant in every game, and quite frankly, you’re a potato, too. If you weren’t, you’d pull off whatever was needed for the win. GIT GUD

10

Is what you just described not an effective MMR system? Surely you belong in a tier where you have an equal chance of both winning and losing the match. Also, if you’re plummeting SR (consistently losing games) it’s a reflection of your skill being unequal to those you match up against; similar to if you were to go on a 10 game winning streak, the system would increase the SR you gain per win and match you up against higher skilled enemy players because you’re demonstrating that you do not belong in your current rank.

What I was trying to say is that if (through performance based SR) in a span of ten games you both lose and win five games (5-5), but you consistently outperformed all other players in the matches- you would see an aggregate gain of 5-8 SR (for example). That’s climbing. That’s you being rewarded for being better than those around you.

1 Like
11

I would also argue the reliability of Overbuff because it does not account for private profiles meaning that the results and data shown on the site are likely inaccurate. And the way I see it, you lost 24 instead of 25 or 26 for performance.

12

if i were dps and could influence the game by way of raw mechanical skill, sure. support CAN hard carry and I do but support’s ability to hard carry is significantly limited in comparison.

further, the assumption you’re making requires supports, tanks, and dps to all be evaluated efficiently within the system. when has this game ever properly scored support? are you suggesting that right now support gains the same amount as dps when they crush it?

this has never, ever, been the case. if you have proof they’ve resolved this, feel free to link me said proof and i’ll read it and update my understanding.

further, you’re suggesting “maybe you have hit your peak”. ok, which one? the high platinum i used to be at and was easily hanging around at? the gold i climb to over and over, easily? or are you suggesting i’m actually a silver player because this system sends me on losing streaks I can’t carry my way out of?

further, how is it so many people have such varying experiences? we have these supposed experts falling on the ‘git gud’ sword claiming the system ‘just works’ then we’ve got the other camp that thinks the system is close, but far from perfect and needs some updates to account for cases where you just get stuck on 10 terrible teams in a row.

the individual needs to have some ability to control slides. currently we have none. play out of your mind, and 5 other people will regularly ruin games for you.

every loss isn’t my fault. if it were, how would you prove that when the system isn’t accounting for performance within the context of that game, with those players, and who did or didn’t help that team of potatos attempt to win.

i have no doubt that if you’re one of the top 10% of the best players in the world, the system ‘eventually’ puts you where you need to be. it isn’t quick though. what about the much larger population that is 1% better than the average? you’re telling me that population is accurately placed in the SR system?

i don’t buy it and you can’t prove it.

1 Like
13

Well, first of all, you’ve got the burden of proof absolutely backwards since you’re the one pushing the conspiracy theory.

Also, the last time I climbed an old, underranked account, I was gaining 80+ SR for every win up until ~mid plat. I know that’s still a thing because a friend was getting the same thing while climbing his Smurf account this past weekend.

14

every loss is 24 points. well, that isn’t true. if im 20+ points higher average SR for the team the loss would be even higher, lower for the inverse.

the system will be flawed as long as it isn’t accounting for your performance with those terrible teamates you were placed with. the SR system itself needs to highlight value. how many ults did you get off, how many kills from said ults? how many ults blocked or eaten? there are plenty of ways to figure out value per game, implement something.

comparing your performance only to the lobby isn’t accounting for the game itself. i didn’t pick these players, you (the system) stuck me with them. if they’re all horrible, and i’m doing my best, you need to penalize them more which encourages intelligent valuable play.

whats that? you play an off meta hero? do you still crush? great, you should be rewarded. etc etc

the way it is now is lacking. it isn’t broken, but it doesn’t offer much player agency in controlling slides. the system can and does stick you with horrible players that you just won’t win with as much as it will put you with players you can win with. you said it yourself, the system tries to make you lose 50% of the time. i can tell you, it does that … really well at times. in fact, two days ago i went like 12-3-1 climbing like 150 some. the day before, 3-9-2 falling 150 some. both of those are entirely my fault right?

so which player am i? the one that easily won games at a 4:1 rate or the one that the day prior, couldn’t buy a win? or the mid platinum player i used to be before falling into a hole (partly HW related with the FPS)?

you tell me.

1 Like
15

There is also another factor in determining the SR change after completing a match, and that’s a measurement of how well you personally performed during the match. If you perform well than you gain more SR when you win, and lose less SR when defeated. The reverse is also true, so if you perform very poorly you gain less SR for a win and lose additional SR when defeated. The personal performance adjustments have been controversial amongst the community for quite some time, especially since the calculations for these adjustments are not at all transparent.

[…]

So after we get back from the holidays on January 2nd we’re going to turn off the personal performance SR adjustments for players in the Diamond skill tier and above.

2 Likes
16

you’re seeing the recent soft reset in action. accounts that havent been played in awhile have all noticed this.

also, i’m not pushing any conspiracy theory. i’m saying the system could be better if it accounted for performance per role per game whilst also overlaying performance versus the lobby.

all we have now is performance versus the lobby. we have no way to account for mistakes or plays made in each game, apart from stupid medals. if you screw up, and your screw up leads to 5 other people losing … sorry you need to be penalized more than the other 5 people. if you crush, you need to be rewarded more than the other 5, etc.

17

Recent soft reset in seasons 3 & 4? That’s news to me.

I’m confused. Now your complaint is that we don’t have 23rd century AI to judge player performance?

I’m not sure what you want anymore, since your original post was asking for PBSR (which is already a thing)

18

Well I would argue that you won at a rate of 4:1 because you were better than the players you played with, ranked up, and then lost at a rate of 3:1 because you were getting outmatched by the better, higher ranked players. Also, I think you misunderstood me when I said that the matchmaking system places in a game where you have a 50% of winning. This doesn’t imply a “coinflip” scenario, rather, statistically speaking, you have a 50% chance of winning the game based on the average skill of both teams excluding exceptional circumstances that the game cannot account for.

You mentioned getting to 2800 SR, I would suggest that (as it currently stands) is your peak. This would explain why you go on losing streaks after reaching a similar rank to about 2200 where you are then able to hard carry your games because you are vastly better than those around you. Your average SR probably falls in the range of 2500.

As for why some people have mixed feelings about the current matchmaking system, although I have no empirical evidence to support a theory, I would speculate the people who are unhappy are the people who feel personally offended by the rank administered to them by the system.

1 Like
19

Blizzard removed streak bonus, both positive and negative boost, because people complained about boosters and derankers. People taking longer to adjust to the “right” rank is a direct outcome of the community’s request.

Therefore, we’re changing the tuning of the streak multiplier to be quite a bit less aggressive. You now need to win or lose more games in a row before any multiplier is used, and it scales up at a slower pace.

20

ancient post, from the old forums, that includes this …

There is also another factor in determining the SR change after completing a match, and that’s a measurement of how well you personally performed during the match. If you perform well than you gain more SR when you win, and lose less SR when defeated. The reverse is also true, so if you perform very poorly you gain less SR for a win and lose additional SR when defeated. The personal performance adjustments have been controversial amongst the community for quite some time, especially since the calculations for these adjustments are not at all transparent.

so, how exactly are these gains and losses accounted? you don’t know as per the devs, “these adjustments are not at all transparent”. cool, so we have no idea if the system is or isn’t working. we dont know if dps, tanks, and supports are all being properly accounted for. even if we assume its flawless, we the players can’t use the info to inform our own play decisions.

so, explain exactly how this scoring works.