It’s not actually “hundreds” of SR if we had real data. OP did the calculation assuming an average of 20 or 25 SR per win. SR gains and losses clearly vary across a much wider range than what was presented. Given OP’s winrate, they are losing games they should be winning and under performing in games they should be winning.
1 Like
Because the average SR is 25 20 is below average. You cannot assume that those 40 games would have been exactly like the PBSR in other games, We don’t even know which games were low PBS are in which were potentially normal. When dealing with unknown variables it’s best to use the average. Either way I clearly address the reasons in my post of why this possibly happens. The post is not about me being confused or not knowing what could have possibly happened to cause this. The post is to prove that this can happen. PBSR are my MMR are unknown factors. I don’t see what’s the point of trying to calculate my PBSR really.
It would not prove anything It’s not already said or assumed.
I’ve already made topics and/or replies concerning the PBSR I get while playing Brig compared to the PBSR I’ve got and while playing Zen. Simply by getting kills with Zen I could get a high PBSR however games where I feel I was equally impactful with Brig are rewarded with terrible SR.
I’m well aware how PBSR works I’m saying I don’t like the system because in games where I can tell I’ve made impact I get low PBSR. I’m not completely clueless to be unaware of what impact looks like either, and PBSR seems to work fine in terms of DPS or Tank or apparently while playing the Zen. I don’t know what it measures for brig but I feel like Brigs PBSR is just at least the style which I play is very underrated. Whatever statistic It is measuring It is obviously not one which I feel important. I play brig as more of an assist type character, and that’s the way I find it I get the most values through assists and displacing enemies and you know crowd control. If the game is not measuring cc value or doesn’t measure it as high as healing value for example this could cause a discrepancy. Which like I said is particularly the reason that I don’t like the system. Because this would force me to play specifically for PBSR not for value I always play in a style which I feel contributes the most value. And what does this say about the game really does that mean that we should specifically be trying to farm stats in order to rank up?
See but you were talking about facing teams that were lower MMR ratios not my PBSR ratio. Of course using the PBSR ratio it makes sense but that’s not what you were talking about.
Sure you’re right the more games you play the less errors there are but that only goes up to the standard by which it can measure value. If I say that I’ll give you my house location accurate up to the mile/kilometer Then sure you could I can guarantee you some pretty good accuracy but the accuracy is irrelevant if the initial system isn’t good enough. But I have experience in video game programming aswell To where I kind of have an idea possible methods they can measure value and no method is perfect because no AI is able to look at a game the same way that a person looks at a game and measures its value. It is hard to describe/understand a situation using only numbers and variables You can get the basic most obvious points He’s done easily enough but not the details more abstract versions of value.
Well I’ll tell you one thing not the same rank that I’ve been in for the last seven seasons because that pretty much just says that I have not improved one iota since I began playing competitive and that is not true because I know at the very least I know know more about ults and how players uss abilities and have gotten better at many things.
As to what I deserve, I think at least gold. Realistically when my other friend used to play online who was a plat / diamond I could pick you back onto his games and we would win most of them on average up until mid-plat where It started getting difficult to get past. Ironically probably as about as difficult as my silver games. People in plat surely are better their aim is better their strategy is better everything is better, technically this should mean harder and it is but Silver is more stressful which makes it seem harder because the amount of carry that you have to do to win a game is nothing compared to the amount of carry you have to do to win a plat game having reliable teammates does wonders even if at the cost of your opponents also being more put together.
So I’ll say yeah definitely not 1700 I think at least gold, but realistically I’ve seen myself play in the lower plat levels and succeed at a positive win ratio.
Whether my replay is reflective of that I don’t know, It’s not a plat game for me to be playing like Plat. Initially when I did do those plat games with my friends I had to learn a few things about plat play, One of the more surprising ones were how great people are at timing how fast your brigs shield goes up and down to where they could snipe you in between something that doesn’t happen at silver but small things we’ll make the gameplay inherently different.
2 Likes
We don’t have the real beginning and ending sr was 300 sr?? Also talking about loosing games he should have won and winning with low stats - isn’t that definition of prediction handicapping?
Just last night I was joking about how the tinfoil hat guys are a little too far sometimes but now we have statistical information and we discrediting the poster and down playing outcomes. Shame
And now we’re defending a system that negates actual wins over stats? We all know stats don’t accurately portray plays. This dude has more wins over losses and he is 300 sr lower than he started.
It’s a sad day for OW.
2 Likes
I could literally post my stats from the last 3-5 seasons, that actually are the opposite of the OP (you could just look at my profile ingame), I had a negative win rate, and gained SR.
1 Like
All due respect I do believe you. I just don’t like the idea of what was suppose to be a minor adjustment to overrule the almighty win / loss stat.
I even think Jeff himself said mmr / sr is all good but it only matters if you win or lose. To me this opens cans of worms like even if you win should you lose SR based on pbsr? Or win but not gain(which happens here) based on PBSR.
More importantly whether you win the match or not.
2 Likes
We’ve actually talked about this, and my stance is, PBSR should outweigh MMR in ranks below diamond were it exists, I don’t always feel that when you lose a game, you should lose SR, I will also say, sometimes when you win, and you’ve legit done nothing, you shouldn’t gain SR, but, this being a team based game, you are reliant on the other 11 players in any given match, but as you said, this opens a can or worms.
In my opinion, and I very well could be wrong, PBSR help us stay at, or climb to the rank we deserve (below Diamond, as those above earned it), my all rights, with my win rate, I should be bottom 500 bronze (we’ve played together, should I be ranked there?), but my stats say I belong where I am +/- 200.
1 Like
I lose between 14-21 sr per loss, and gain 28+, pbsr does help me stay where I belong.
1 Like
You are being measured against other player performance data for the same hero and map. You’re essentially arguing that you get more value than higher ranked players with better stats. That doesn’t make sense.
You’ve seen yourself succeed in Plat because you were getting carried. The logic you’ve presented here is: “If my teammates carry me, I will win.” Your current gameplay is commensurate with your SR. You can certainly improve if you start taking responsibility for your gameplay and reconcile mistakes. It can be hard to know what you’re doing wrong because you almost never get punished for doing ridiculous things at your elo. Check out HolyShiftKid’s Youtube and Twitch for examples/guides of what you should be doing.
Correct. If you play like a Plat, you will be in Plat. If you play like a Silver, you will be in Silver. If you want to be in Gold, play like a Gold.
If you take the player’s peak SR and compare it to the current SR. If you take last season’s SR and compare it to their current one, there’s a 79 SR increase.
No, I don’t think so. If the MMR matchup predicts a 60% win chance for you, but then you lose, you lose more SR. This is just like chess Elo–if you’re the higher rated player and get rekt, you lose more. If you get rekt by a higher rated player, you lose less Elo.
When you’re predicted to win (MMR) and you do win, your SR gain won’t be as high as it could be with consistently better performance (PBSR).
Your win percentage doesn’t matter if you’re only winning against ridiculously easy opponents. It’s the equivalent of a chess super Grand Master destroying someone a high-school chess club and expecting to gain hundreds of elo points. Now, if that Grand Master gets destroyed by that high-school chess club, she will lose a ton of elo.
PBSR helps you climb faster and remove the grind if you’re performing above your current elo. If you’re average for that elo, then you don’t get extra SR when you win.
If you want smurfs and higher ranked players crushing lower elo games, then go ahead, remove PBSR.
Now we’re ignoring the myst post?? Minor influence and most important factor is win/loss??
We can’t control or blame our teammates, that’s a pill. Now at its most fundamental level of having a winning record and ending lower which is against what the devs even communicated.
Controlling someone’s sr by their mmr predictions should even out if your at 50/50. No1 knows the game is predicting ahead of time unlike a chess game and the willing to take the risk. AND UNLIKE a chess game Elo is pure skill mmr accounts for region, ping , and other undisclosed factors that blizzard doesn’t tell us. MMR is not elo.
At the very least argue about pbsr but calling his data fake too?
So this dude wins, but when they lose at the wrong games and wins with the wrong stats. Please see the bigger picture, it’s not just this dude, he represents plenty of others and you hear about elo hell. Folks I don’t know what else to tell you.
ALSO the whole 40-60 is edge cases and again they’ve won more games than loss so at some point it’s probably impossible to have always lost when at 60%. Unless your saying more than average he was suppose to win?? (Over multiple seasons too). It doesn’t make sense. For 300 points too??
I think my soft threshold would have been 100, what is everyone else’s threshold? If they posted 500 sr diff is that when the community bans together? At what point will people come together?
2 Likes
Not sure why did you put that giant pillar of text in there, when it is known that winrate is not equal absolute winrate.
For example, I won 10 matches of 10 playing one same hero, with winrate of 98% 
Ok more detailed:
Soldier76:
Time played: 1 hour
Games Played: 10
Games Won: 10
Games Lost: 0
Win Percentage: 98%
1 Like
No that’s not what I’m saying, which is why it doesn’t make sense.
Oh ok here we go, the guy who knows my situation better than me. 1st You can’t really carry if MMR matchmaking system actually works, unless you specifically make an alt account to boost. 2nd of all he would only climb when we played together. Though lastly you can go ahead and assume everyone you talk to must be inferior to you not incapable as if the game of Overwatch was programmed by the gods themselves, but there is only assumptions Behind what you say.
You don’t understand statistics very good do you. Within of span of 10 games Maybe there’s a chance that I got lucky. But within the span of 884 games if I was consistently the weakest link I would not have a ratio higher than 50%. Because the people who carry will not be exclusively on my team They will even out to a 50/50 ratio in the long run. It would not be possible for me to have above a 50% well underperforming and relying on lucky carries. Either that or you’re arguing that I have some insane amount of luck.
Assuming the system is completely 100% infallible. What I’m offering is simply what I’m offering. And while I would need to have the entire game code and it would require much larger investigation than this. Clearly statistically it shows signs of potential error. The statistics do not make sense If I was playing below average to merit below average pbsr, in my win ratio should be below average. With 884 games played any potential argument that I got carried or this and that happened goes out the window because of averages over time and the sheer amount of games I played, for the very same reason that you could get a bad team every now and then and supposedly due to the averages over time it will not matter in the end because everybody gets bad teams aswell.
The only reason I didn’t get a higher win ratio was because I pretty much switched off a brig. Based on the sheer amount of games played we know that I am above average at least with Brig for the rank that I’m in and I’ve always maintained above average performance in every every season as literally shown by the game that you so adore and statistically I’m not even sure why you even try to contend this. I don’t know up until what points because the matchmaker hasn’t allowed me to rank up. I only know of the plat games because I played the black games with my friend. I can tell you what happened and I have no reason to lie, this post wasn’t even about my skill It was proof of redundancy. My skill is obviously proven within the win ratios unashamedly even for the characters which I really do suck at.
Unless you’re telling me that you can underperform and still get higher than 50% when ratios for close to 1,000 games. That logic doesn’t even make any sense…
Statistically It doesn’t make any sense, with a test pool of 884 games, You cannot be underperforming and maintain a positive win ratio, The more games you play the more your win ratio becomes reflective of your own skill.
Having above a 50 win racial means that I’m at least doing better than averages which means and if I perform at least better than average I should have at least average PBSR If PBSR calculates your value correctly. And vice versa if I had low PBSR that means I am underperforming and if I am underperforming I should not have higher than 50% win ratio after this many games. So for your argument to work you got to choose one which one is the one that’s not adding up Is it math that’s wrong or is it PBSR that’s wrong because either mathematical statistics is wrong and making me seem better than I am or PBSR is wrong and making me see more than I am. Pick
3 Likes
Yea this goes back to obfuscation- these screens on stats stink. Like the medals we get fed bits of info but it ultimately doesn’t matter or mean anything. It’s so unlike the bigger games like wow and Diablo.
Nobody is saying otherwise. Yes you can drop down even with 55% winrate, just like you can climb with 45% winrate. Its really not a secret.
3 Likes
Although I left the weird percentages that were in the game stats for each characters individual winrates, the totals were calculate using familiar math (Basically added up the wins and the losses to get total games and then divided that the number of wins by the total games getting a decimal which I converted into percentage) I just put it the weird win ratios for the characters for the sake of just completion, or in case it mattered.
True, but nobodies saying and much of anything. And then It doesn’t really add up when you think about it at least not during a span of close to a thousand games. Because if someone is consistently getting a 45% win ratio over a thousand games then they are legitimately underperforming and they wouldn’t rank up with PBSR, unless they were just insanely unlucky for some reasons but after so many games you figured the average is over time would take care of that. Same thing with 55% win ratio if you were to maintain a 55% win ratio over a thousand games then you are performing better than average unless again you are just insanely lucky. PBS are can’t stay below average and when Rachel stay above average for a long time, If they do that means that something’s inherently wrong because either the PBSR isn’t measuring value correctly despite performing better than average or a guy is performing below average and somehow still ranking up this would only make sense if he was being boosted intentionally over the span of a thousand games
Which is why I added up all the wins and losses throughout the seven seasons for a grand total. I don’t know why I would play close to 1,000 Just on support alone spend like a year playing a thousand games not to mention the other games like tank and DPS that I played on this account since it’s my main account Just to prove this. It’d be easier to make a new account play as minimal games as I can possible to prove this. I don’t need to go seven seasons. It’s inherent with within the first season actually you could see that I come out with a positive win ratio and go down in SR.
Even If it were true and I did manipulate the stats the point still stands that it can be done, That alone is substantial in itself. Think of it similar to white hackers who just try to break the program on purpose to point out any flaws in the system.
That said I did not do it on purpose. The purpose of is to show that there might be some problems with the PBSR and MMR calculations You cannot maintain such ratio for that long and have low PBSR That is unless I was intentionally being carried by friends for over a thousand games. Which was not the case I solo q most of the time probably over 95% of the time or more. It would not be possible for me to get carried solo Q because of averages over time would make it a redundant factor.
Really it just falls on you to make your own best judgment on the whole situation, because it’s either I’m lying or playing the system or there’s a bit off with Blizzards system. I just put the information so people could see it and because I never knew it was possible to be hampered to this degree so it seemed relevant to share.
You don’t have to believe it but surpass there’s another player who feels the same way like he’s been stuck in the same rank despite doing better He could check his previous games and he might see a similar pattern.
1 Like
Btw you realize that winrate is sometimes weirdly distributed amongst more heroes if you play more of them in match right?
2 Likes
Yes somebody else has mentioned it before That’s the response I gave them. So only the individual character stats will have the weird percentage but the stats that actually mattered I made sure were actual mathematically sound percentages
Too bad you didnt write down SR gain/loss for these games.
2 Likes
Well yeah but I don’t see what difference it would have made. Also this wasn’t my intentions to eventually prove this this is just something that I looked at and saw that there was something weird about it. Then realizing that there’s something very weird about it to a bigger degree than I thought initially. Because I had noticed playing support that I would lose more as are than I would gain most of the time not by a lot but maybe by 3 or 5 SR difference despite on whether I was doing good or not.
However maybe this number was higher than I originally thought or I don’t know but the definite culprit is going to be PBSR. It doesn’t make sense for the culprit to be previous or after so many games but that’s going to be what it is It wouldn’t benefit me much to know detail per match.
However one thing that can benefit me is if from now on I specifically try to play certain ways with support. And seeing which ways get me the most PBSR a because there’s obviously some type of way that I’m playing which isn’t getting its value measured correctly (considering I’m still getting higher wim percentage therefore I can’t be underperforming after close to 1,000 games). Once I know if the system just wants more of a heal bota or what it wants in particular then I could just play & game the system instead of getting played by it and maybe then start making some actual progress lol. Unfortunately that’s going to take some time and it risks my SR even doing that during comp and my SR already faces a lot of obstacles already as it seems. The fact that it might have actually come down to this is pretty crummy though.
Your point is well taken. The absurd grind that you are talking about is mainly a consequence of Match Making Rating, which is what performance-based skill rating adjustment is supposed to compensate for. But as you see, the relationship between win/loss record and rank is totally broken. More information in my thread on the subject.
1 Like