please give us flat sr gain\loss

Competitive Discussion
They don't have the stats, or the technology to sustain a performance based SR system for competitive players. It's contradictory to the entire concept of what competitive Overwatch is. Why would someone be concerned about comparisons that are being made between players on the same heroes that are playing entirely different matches? They're playing those heroes in a different context. There is too much variation and it produces results that aren't reliable in a game like OW. How you're comparing to the same heroes in entirely different situations should have no impact on SR.

Time to wake up, it's not possible right now. It has the same adverse effect on the entire ladder whether you believe it or not. Players that generally don't belong in the same SR range that you're trying to exceed, regardless of the range. It's an inaccurate indication of skill in Overwatch and it's creating a massive variation in skill level between players in the same SR range. Only the effect has been far more drastic at higher SR.

Trolls, one-tricks, and players that refuse to coordinate/communicate shouldn't be eligible for free agency in the OW league. With the current system these players are being matched with professionals consistently, at all times. The results have been disastrous, and honestly just a huge joke to the competitive scene.

Static SR/MMR seems viable with separate queue systems for solo/duo and stacks, either way it needs to change.

Just wanted to add:

Even as a temporary solution, this system can be tested in a certain SR range so that it doesn't immediately impact or target players that are in a lower SR range. For example, Diamond ( 3000 ) and beyond can be a static system exclusively.

For the following season, you could potentially just roll back all players that were Diamond and beyond to 3000 as a starting point.

Of course the queue system would still have to be separated between solo/duo and premade. Ideally you would still want the Top 500 rank to be exclusive to solo queue. The most accurate indication would be a full SR reset long term however.
NO! You would never have a chance to climb in this slot machine matchmaking. Don't event compare to dota2 - it works there because you can solo carry even if you get one or two imbeciles (naturally if you are more skilled player than enemy). OW is purely based on how well you work as a team - giving it flat SR would make it even bigger slot machine than it is now - got two imbeciles, and no matter how well you play you lose max SR (aka said flat 25SR).
yes
07/29/2017 10:07 PMPosted by basil
i can see where you are coming from but i dont think this is fair because one guy may be getting carried by the team but he'll still be rewarded the same amount of sr as the other people who are doing all the work. this system now is quite alright the amount of sr that you gain or lose depends on how much you contributed to the team and how your play is compared to other people who played that same character on that same map. dota and overwatch are 2 very different games my man


Currently we have the inverse problem where as long as you're playing a DPS and you're semi accurate, you'll get far more SR on average than the healers or tanks will on average. I have a 58% win rate on Mercy this season, and I've gone from 888 to 1490. I am currently 1375, yet still have a 58% win rate (it's actually up from the 56% it was) yet I'm stagnating in this 1350-1490 range, because wins are often steamrolls which cause my stats to be garbage.

This doesn't even address the issue of people from Plat and above deranking to bronze and skewing the system's data by artificially inflating the ceiling of "what players using these heroes are capable of" when comparing you to other people in your same SR range to figure out how much SR you deserve.

Flat gains and losses would be the best situation, because people who end up getting carried will fall back to where they belong even quicker than with this current system.

Edit: I mention my Mercy win rate because I'm trying to play her exclusively this season just to test it. I obviously will swap when necessary or when she is taken and the person doesn't want to swap off her, but I have stagnated at this SR for well over 50 games now with her, so it's an accurate depiction of her positive win rate yet SR not reflecting that.
07/29/2017 08:32 PMPosted by Palsy
Support mains doesn't get as much SR as DPS mains so it seems like a good idea to balance it all out. i fully support this


They already changed it.

I climed from 1700 to 3100 after the change.

It is interesting though, last season I got like 47% win rate but climed from 1700 to 3100,
This season I have 53% but I dropped to 2973 SR
The reason why I dont think blizzard wants to do this is because then comp would be entirely based around win percentage. Who ever won more would be the highest rated. I think this is a good idea though and here is why. Currently the system is all about how you perform, in your team and against others playing the same character, seems ok right? Well sadly no. The reason why this is not ok is because the system tries to give you a 50% win rate no matter what, therefore it is in the players favor to main an off meta hero. Because of this they will likely recieve a 50% win rate because the system tries to give them that for free, all while gaining more sr for winning and losing less sr for losing because they are out performing other players with that character. So if you are stuck at a rank and seem to gain less then you lose, try experimenting playing with other characters. Then document how much you gain and lose. You might find a way for a dimond player like me to become master like my friend who tried maining torb.

what about the phenomena of mercies playing in a specific way to boost thier stats but being detrimental to thier team? While mercies who play for thier team gain less.


This is so true, and it saddens me greatly.
07/29/2017 09:54 PMPosted by Poli
07/29/2017 07:38 PMPosted by Norse
there is not enough time in season for this........ it takes 1000 games to average out to 50/50

maybe they should not reset SR and have mini leagues that last for a month or 3

then we can have flat gains


They don't reset your MMR every season so your placements always put you close to your finishing rating. There is really no difference between seasons.
No difference, other than wasting your time doing the placement, and wasting your effort in previous season
07/29/2017 08:22 PMPosted by Squidbarrel
07/29/2017 08:17 PMPosted by Cakewalk
No,

people get carried will get same amount of SR as the ones doing the hard works?

It's not perfect, but it's far better than how it works now where players can climb ladder with a negative winrate.

This has my full support.


So <removed expletive> the people who are actually doing well and only losing because of teammates.

YOU realize these are legitimate circumstances???

This SR system is infact not performance based ENOUGH

it needs to be 100% performance based.

IF you do good and still lose because 1 person cant carry 5 other people.
(<removed expletive> your logic if you say otherwise)

then you should be rewarded with losing less sr.

[Edited for language by Moderator - please review the Forums Code of Conduct]
I'm not sure that this is what you want.

'ELO HELL' is the concept that your team does poorly and you lose.

Performance based SR allows players who do well, even in bad games, to lose less SR and to gain more on wins. This allows any one to climb out of their ELO. With a flat SR rate, it would cause players to actually be stuck due to the majority of players only having a 50% win rate, or climb extremely slowly. Regardless of skill.
07/30/2017 10:45 AMPosted by Noob
I'm not sure that this is what you want.

'ELO HELL' is the concept that your team does poorly and you lose.

Performance based SR allows players who do well, even in bad games, to lose less SR and to gain more on wins. This allows any one to climb out of their ELO. With a flat SR rate, it would cause players to actually be stuck due to the majority of players only having a 50% win rate, or climb extremely slowly. Regardless of skill.


It's amazing how OP hasn't thought of this.

It's almost as if they're retarded.

I feel bad for them.
Its a team based game, why not just give 300 pts to winning team, lose 150 for losing team, then balance based upon team contribution?
No.

Stats matter.

The game needs a way to know who is getting carried and who is an anchor.
If the point of the game is winning then the only stat that matters is whether you win or lose. I'm pretty confident that the only reason they take other things into account is that you have the same SR for being in a group, so someone could just be carried to a high SR by having a really good teammate for every game. Of course, they still can under the current system, but there's the performance thing there to lessen the effect of that.

But the real problem is that Blizzard doesn't take its own game seriously and thinks it's more important to let people play with their friends in competitive than to have a good matchmaking system. And they go so far as to give you extra xp for being in groups, so that just encourages people to group up even if they don't work well together, which means when I see a group in my game there's a good chance that the game has already been decided by whether the group on my team is a real group that works well together or just some random people who hit the group button after a match.
07/30/2017 05:05 AMPosted by JRoy
They don't have the stats, or the technology to sustain a performance based SR system for competitive players. It's contradictory to the entire concept of what competitive Overwatch is. Why would someone be concerned about comparisons that are being made between players on the same heroes that are playing entirely different matches? They're playing those heroes in a different context. There is too much variation and it produces results that aren't reliable in a game like OW. How you're comparing to the same heroes in entirely different situations should have no impact on SR.

Time to wake up, it's not possible right now. It has the same adverse effect on the entire ladder whether you believe it or not. Players that generally don't belong in the same SR range that you're trying to exceed, regardless of the range. It's an inaccurate indication of skill in Overwatch and it's creating a massive variation in skill level between players in the same SR range. Only the effect has been far more drastic at higher SR.

Trolls, one-tricks, and players that refuse to coordinate/communicate shouldn't be eligible for free agency in the OW league. With the current system these players are being matched with professionals consistently, at all times. The results have been disastrous, and honestly just a huge joke to the competitive scene.

Static SR/MMR seems viable with separate queue systems for solo/duo and stacks, either way it needs to change.

Just wanted to add:

Even as a temporary solution, this system can be tested in a certain SR range so that it doesn't immediately impact or target players that are in a lower SR range. For example, Diamond ( 3000 ) and beyond can be a static system exclusively.

For the following season, you could potentially just roll back all players that were Diamond and beyond to 3000 as a starting point.

Of course the queue system would still have to be separated between solo/duo and premade. Ideally you would still want the Top 500 rank to be exclusive to solo queue. The most accurate indication would be a full SR reset long term however.


This.
07/30/2017 12:58 AMPosted by gungay
"flying around healing your team whilst not being killed yourself" funny you mention io, its actually very difficult to climb with it in soloq and thats basically its playstyle except you don't have a rez, guardian angel has a 12 second cooldown and your damage boost lowers your health.

The fact that you mention it's difficult to climb in solo-q as io proves my point about how playing support with a flat sr gain/loss system is essentially a coin toss.

07/30/2017 12:58 AMPosted by gungay
i don't think mercy mains are climbing out of bronze with a 40% winrate, please provide proof.

I can't provide proof because the system doesn't allow it to happen, except in case of smurfs. When a smurf player starts actually trying to rank up the rank up very quickly because they have a much higher mmr than the sr rank they are in. Essentially the matchmaking system won't put an amazing mercy in bronze in the first place. That's why there are placement matches. And if a good mercy player was put into bronze the would rapidly rank up even on a cointoss for who wins. This is because the matchmaking system realizes "This mercy player is outplaying all the other mercy players at the same sr level. Clearly they don't belong here. I help them climb even if they get poor teamates by making the losses not as bad and the wins much better".

07/30/2017 12:58 AMPosted by gungay
can you define "limited experience"

<20 hours I played unranked and got fed up with how complex the game is. Even to just play one hero. I watched some pro matches to try to get better. I realized I don't have the situational awareness to even be average at the game. I still struggle with situational awareness in overwatch.

07/30/2017 12:58 AMPosted by gungay
there are lots of skillshots in the game and mechanically intensive heros, its not as simple as you describe even with heros who lack any skillshots.

I'm not saying there aren't skillshots. I'm saying isn't nearly as much aiming required. Dota 2 requires more situational awareness ie. knowing when to fall back, push, jungle, and buy an item. Flat sr works for Dota 2 because if you have good situational awareness you can share that with your team and essentially carry by using your voice comms well. Yes, actually mechniacal skill is important too, just not as much when compared to Overwatch.

Your whole argument that "if it works for Dota 2 then it'll work for Overwatch" just doesn't work because the two aren't comparable. If you were arguing that League of Legends should run their ranked system like Dota 2 does I'd completely agree with you. That's because the two games are comparable. What you learn in one is generally transferable to the other. Whereas with Dota 2 and Overwatch the skills you get from one are not nearly as transferable.

07/30/2017 12:58 AMPosted by gungay
https://mindgames.blog/2017/06/22/in-defense-of-purist-skill-rating/

I disagree with the pro player. *audible gasps are heard* Flat sr might work for higher levels because there is less random chance. There aren't many trolls leavers or just plain bad players in the higher levels. That leaves the match up entirely to skill. I can tell you from experience being in gold/silver there are plenty of matches I had no hope in hell of winning because of trolls, throwers, or a crappy team.
Their SR system is so obviously stupid that I don't ever expect it to be fundamentally changed. Something like this involved the whole chain of command and cost a huge amount of money, the reputations of the management chain are probably riding on it. Ergo, the plan will continue to be what it has been: to make it appear that all the past decisions were fundamentally sound.

The fact is though that anyone with a passing familiarity with AI would have told you that accurately assessing skill with fancy algorithms is a pipe dream. All they are doing is demotivating players who have clever game-changing moves, or who are good at managing the team via audio, which can't be detected by their stupid AI. They should have known this would happen.

They also should have known that you have to pick whether you want to incentivize WINNING or whether you want to incentivize GAMING THE AI. The best way to motivate healthy team behavior, where everyone is 100% dedicated to WINNING and without any shortsighted selfish agenda to get the retarded AI to think they are the best player, is to divvy up the winnings equally and then let the best players rise according to their overall wins.
07/29/2017 08:17 PMPosted by Cakewalk
No,

people get carried will get same amount of SR as the ones doing the hard works?


That will not happen on average (people of equal skill would, if the algorithms were competent, be playing with people of equal skill) and in any case is far better than systematically instituting perverse incentives.
hey a bunch of people responded! hi! ill put my answers to a bunch of your posts here.

07/30/2017 02:35 AMPosted by Poptart
07/29/2017 07:35 PMPosted by gungay
what i mean is 25 lost if you loose, 25 points gained if you win. simple. dota 2 already does this.

please, it would help a ton and reduce the number of people complaining about elo hell

ideally give us slightly less sr lost for a game with a leaver, say if you normally gain or loose 25 sr a game with a leaver on your team makes you loose 20, unlike the current performance based system where you are going to perform worse with leavers and loose more sr.

if you dont want to listen to a bronze player, maybe listen to a pro overwatch player, LG evil jake

https://mindgames.blog/2017/06/22/in-defense-of-purist-skill-rating/
why? it would just let people doing absolute !@#$ lose nothing and the people who carried win nothing. wtf? are you autistic or just bronze?


well i do play a lot of symettra : )

did you miss the link where a pro ow player explains this idea in much more detail?

07/30/2017 02:12 AMPosted by Hauven
I'd like to agree that it's worth trying but until the toxic elements in competitive play begin to noticeably disappear (e.g. throwers, deliberate de-rankers, such like) it's the only thing keeping some genuine players afloat. Maybe in a month or two the actual matchmaking will start to feel more balanced now that there's apparent actions taking place to punish the toxic players who have been reported. In the meantime it's still pretty much a stomp or be stomped experience, e.g. 0-3, 3-0, 0-2, 2-0, rarely anything like 2-1, 1-2, 2-3. That's not to say there will be stomp matches still, but hopefully less than there currently is as some players won't be unfairly carried up or pushed down so often by the toxic problems out of their control.


its true! toxic people and leavers are a problem, i think it would help a little bit with them because they cant make you loose more by making your performing worse? What do you think?

07/30/2017 04:05 AMPosted by Sentoki
It's true, the winning and losing streak should go.(btw even if you are bronze player, You may still have an opinion) Getting a winning streak is nice, you can be a part of that. But losing streak not so...if someone leaves or dc, 99% of the time you will lose the game, because the game 'balances' the teams, so the team with 5 players are not a lot better that they can compensate for the 1 leaver. So you can't control the losses. I'd would rather have a system where the other team has 1 random player be disabled, and only can view the game until the game is 6vs6 again.
The player that is forced to become spectator, should get sr for winning, but not lose any sr when his team loses after all. At least that way teams will always be balanced player wise.


hey thats actually a really good idea!!!

07/30/2017 04:25 AMPosted by BlueFountain
Prime example why I hate the current performance based SR.

Liajong Tower: Ana/Lucio 45 elims, 17 deaths, over 15,000 health healed. Game lasted 5 rounds. 20 SR loss

Anubis: Ana ? Elims, 2 deaths, 3200 healing. We held first point and took it on attack on our first push. 21 SR gained

And that was two games in a row. I mean if this system could legit measure us based on personal performance you'd expect the game where me and my team destroyed the enemy to have given me a lot more SR while the game I barely lost to have been much lower SR loss compared to the one I did win.

And both games yes I had gold healing. My accuracy for both were low 70%. I by no means was carried in that win nor was a burden to my team in that loss.

Oh btw the next game half my team left shortly into round 1 and I lost 30 sr. Felt great.


hm well with the slightly more for winning you can climb by going win\loss win\loss but OUCH yea you get a leaver who !@#$s over your performance and BANG loose 30 sr, not cool! its not a fair game and you shouldn't loose sr that requires you to win two games to make it up.

hey id also be happy if they made is so you could never loose more than you gain, just make it go from 25\25 and increase based on performance but not decrease that would be ok too.

07/30/2017 04:31 AMPosted by Ashwynn
07/29/2017 07:35 PMPosted by gungay

please, it would help a ton and reduce the number of people complaining about elo hell


Well, if you look up DoTA2 forums or League of Legends... this particular point is not true. People will allways complain about ELO hell. It's a mindset, not a game system issue.


i think league is still performance based? it still sounds better than overwatch, but i think its actually slightly easier to carry in league even compared to dota, its much more structured than overwatch or dota, it has a role q and you can get banned for qing for support then playing a hero that isn't a support as support or even attempting to go for one of the other roles.

dota does have complaining but less of it in my experience, people are probably always going to complain especially with a game that sometimes puts you into unwinnable games, but that doesn't mean my proposition is a bad idea.

07/30/2017 05:29 AMPosted by walentaz
NO! You would never have a chance to climb in this slot machine matchmaking. Don't event compare to dota2 - it works there because you can solo carry even if you get one or two imbeciles (naturally if you are more skilled player than enemy). OW is purely based on how well you work as a team - giving it flat SR would make it even bigger slot machine than it is now - got two imbeciles, and no matter how well you play you lose max SR (aka said flat 25SR).


well clearly theres a solution to the slot machine problem!

role q! yup! thats something that needs to be added, that would help a lot with the slot machine effects.

Also i think you are overestimating how easy it is to carry in dota.

07/30/2017 10:45 AMPosted by Noob
I'm not sure that this is what you want.

'ELO HELL' is the concept that your team does poorly and you lose.

Performance based SR allows players who do well, even in bad games, to lose less SR and to gain more on wins. This allows any one to climb out of their ELO. With a flat SR rate, it would cause players to actually be stuck due to the majority of players only having a 50% win rate, or climb extremely slowly. Regardless of skill.


elo hell doesn't exist just to be clear.

i want people to be able to climb with a positive winrate, so 51% ?

anyway a solution to your issue would to put a performance based sr bonus system on top of the flat sr system but make it impossible to loose more than you gain, simply gain just as much as you loose.
07/30/2017 05:05 AMPosted by JRoy
They don't have the stats, or the technology to sustain a performance based SR system for competitive players. It's contradictory to the entire concept of what competitive Overwatch is. Why would someone be concerned about comparisons that are being made between players on the same heroes that are playing entirely different matches? They're playing those heroes in a different context. There is too much variation and it produces results that aren't reliable in a game like OW. How you're comparing to the same heroes in entirely different situations should have no impact on SR.

Time to wake up, it's not possible right now. It has the same adverse effect on the entire ladder whether you believe it or not. Players that generally don't belong in the same SR range that you're trying to exceed, regardless of the range. It's an inaccurate indication of skill in Overwatch and it's creating a massive variation in skill level between players in the same SR range. Only the effect has been far more drastic at higher SR.

Trolls, one-tricks, and players that refuse to coordinate/communicate shouldn't be eligible for free agency in the OW league. With the current system these players are being matched with professionals consistently, at all times. The results have been disastrous, and honestly just a huge joke to the competitive scene.

Static SR/MMR seems viable with separate queue systems for solo/duo and stacks, either way it needs to change.

Just wanted to add:

Even as a temporary solution, this system can be tested in a certain SR range so that it doesn't immediately impact or target players that are in a lower SR range. For example, Diamond ( 3000 ) and beyond can be a static system exclusively.

For the following season, you could potentially just roll back all players that were Diamond and beyond to 3000 as a starting point.

Of course the queue system would still have to be separated between solo/duo and premade. Ideally you would still want the Top 500 rank to be exclusive to solo queue. The most accurate indication would be a full SR reset long term however.


Hey thanks! its nice to see high ranking players making reasonable and well thought out posts, your idea about testing it in higher ranks is a good one! They could also make it an opt in thing possibly for people like me who want to use it as soon as possible in bronze

Join the Conversation

Return to Forum