The legitimate comp issues from a Blizz white-knight

Competitive Discussion
I spend a lot of my time here arguing against people mad at Blizzard, because there is a lot of BS complaints. Usually stemming from ignorance. It frustrates me because productive discussion is often pushed behind meaningless whining.

But there are valid complaints about the SR system. In this post, I want to highlight what I believe the real issues are. And specifically, issues that can be solved. (You can't stop leavers from happening, sorry)

1. Trolls and Throwers

While Blizzard takes harsh action against leavers and hackers, players can seemingly throw games at will with no punishment. Some type of community driven system needs to be implemented to punish players for bad behavior (or reward for good behavior). Low priority queue in Dota is a great example of one possible solution.

A unique issue is people reporting those who take off-meta picks. Some legitimately feel this is "throwing", even if the player is trying their best. I'm confident Blizz can solve it. In my personal opinion as a flex player in GM, one-tricks are not throwing and its a valid play style.

2. Performance based SR

The cats out of the bag with the SR system, so to speak. A large group of players understand how the performance system works, and abuses it. Yesterday I literally played with a mercy main who was top 500 with a sub 50% win rate. Not only can it be abused to climb on any unpopular hero, the system itself is broken for low-mechanic heroes like Mercy/Symmetra. It cannot measure their impact properly, and is often rewarding wrong decisions that boost stats.

If they don't want to simply remove it, Blizzard needs to at least turn the weight of the factor down. Better yet, it could be turned off for high ranks (where most of the issues come up) or on low-mechanic heroes like Mercy and Symmetra.

3. Pause feature

After the first few weeks in a season, regular leavers are banned, and a large % of the DCs I see in my own games have players return. The impact of their leaving would be greatly reduced if the competitive system allowed for pauses.

Again, we can look to games like CS:GO or Dota 2 for examples of the system that let a team pause for a reasonable amount of time, but cannot be abused. Tournaments can pause Overwatch, so it's really only a matter of putting the feature into the competitive system. In fact, OW could auto-pause when it senses someone has disconnected, and give them about 1 minute. Possibly with control from the DC team to resume immediately by vote, if they know the player isn't coming back.

------------------------------------------------------------------------------------------

These are clear problems with reasonable solutions. If all three were tackled by Blizzard, I think the competitive system would feel significantly better for a lot of players.

Feel free to add other problems, and solutions, you think I missed. Though knowing this forum, I'd bet most won't meet the criteria of being an actual problem (e.g. 'Elo hell'), or a problem that's solvable (e.g. people leaving).
Point 1, I agree for certain. Throwers, leavers, de-rankers they're all a pain and don't help make it any easier for someone trying to genuinely climb.

Point 2, I sort of disagree and agree with. Unless point 1 shows more being done then I'd rather individual performance not be turned off. With individual performance not being considered at all then a leaver, thrower or de-ranker (or other such vermin) would have full control over their teammates skill rating and that's not really desirable.

I like your idea regarding having individual performance not considered for those in the higher end of the skill rating though. Perhaps another suggestion might be to have individual performance not considered or weighted much less if there's 'not enough' data to base and compare 'individual performance' on for a said hero and map combination, a good example perhaps being Torbjorn.

Point 3, I think is definitely a must.

I'd also like to add a point, from Dota 2, add a requirement for a phone number to be provided (block VoIP numbers). An automated voice call can be made to the phone number provided which could offer a PIN code that you have to enter on the website to confirm the automated voice call was received. Why? This would help reduce smurfing. It won't completely eliminate it of course, but it would help cut it down. Smurfing is a problem when smurfs deliberately play in a lower tier than their main account is in. Of course I don't expect this to be taken seriously on an official level as smurfs make them profit too.
I've said this multiple times before. There are legitimate problems with the Matchmaker, but the MAJORITY of these problems are at the highest elos, where population is the biggest problems.

The issue is that low ranked potatoes use the argument from authority fallacy to justify their own inability to climb from the lower ranks.
07/12/2017 08:58 AMPosted by IXEL

Point 2, I sort of disagree and agree with. Unless point 1 shows more being done then I'd rather individual performance not be turned off. With individual performance not being considered at all then a leaver, thrower or de-ranker (or other such vermin) would have full control over their teammates skill rating and that's not really desirable.


This is why I like the "turn off at high-skill rating" idea. Particularly, they could reduce it from it's current value at rank 3000, to zero at 4000. At high ranks, players all put in enough games to average out random stuff like throwers. The benefits are really for a casual competitive player, but the problems mostly hit the high ranks.

I'd also like to add a point, from Dota 2, add a requirement for a phone number to be provided (block VoIP numbers). An automated voice call can be made to the phone number provided which could offer a PIN code that you have to enter on the website to confirm the automated voice call was received. Why? This would help reduce smurfing. It won't completely eliminate it of course, but it would help cut it down. Smurfing is a problem when smurfs deliberately play in a lower tier than their main account is in. Of course I don't expect this to be taken seriously on an official level as smurfs make them profit too.


I do like the phone idea, called prime match making in CS:GO. A complication with smurfing is there are legitimate reasons for it. Some people do second accounts for solo play only, or second accounts to play different heroes/roles than on their main. I might be GM on my best heroes, but high diamond on others, and want to be able to fully practice these heroes without throwing matches.

I think the only issue with smurfing comes from players who throw games to intentionally be ranked lower than they should. Theoretically, fixing throwing fixes that.
Completely changing this post. I originally wanted to make a point that Im glad this is happening so the higher level players are exposed to some of the troll awfulness we in Gold/Plat and below deal with constantly, and have been complaining about this whole time while the high level players tell us to just "Git Gud". . .but I'm going to go a different route.

The problem isnt the system, it's the community, population, and monitoring. The system would be fine if there was a better system in place for dealing with people who are genuinely throwing the match.

Part of the problem with that, is according to many grandmasters and the devs themselves. . .picking Torb isn't throwing the match. If someone on your team picks Torb and sticks with it your team needs to build around it.

It's no different than when the higher level players have been saying "If you are good enough you can carry the match". . .Well. . .If you -are- a Master, shouldn't you be able to carry a Torb who is mechanically proficient enough to have been placed in Master? If not, maybe you deserve that particular loss.

Seems to me that the new SR system has just more evenly distributed the trolls.
07/12/2017 09:02 AMPosted by BBW

The issue is that low ranked potatoes use the argument from authority fallacy to justify their own inability to climb from the lower ranks.


As the high elo players complain about an inability to stay in their higher ranks now that our smurfs are moving up?

I deff see the point you all are making, and I'm sure it does need to change. . .But I'll be honest, I'm glad those One trick Torbjorn, Mei's, Junkrat's, and so on are being shuffled up so you all can bring the attention of the devs to a problem we've been shouting about for a long time now.
I believe performance should be removed, wins and losses only. Also solo and group SR ratings. Those that play smart will naturally get more wins than losses. There are players with 60% wins with over 200matches played still stuck in gold which doesn't make sense.
The title of this post is hilarious. His/her opinion of their self and issues doesn't equate to legitimate. Lol there are many issues with comp and your opinion cant white wash it nor does the devs give 2 cents what you say...
07/12/2017 09:02 AMPosted by BBW
I've said this multiple times before. There are legitimate problems with the Matchmaker, but the MAJORITY of these problems are at the highest elos, where population is the biggest problems.

The issue is that low ranked potatoes use the argument from authority fallacy to justify their own inability to climb from the lower ranks.


Inability to climb is based on a multitude of factors and it's not purely from an individual statistical performance standpoint. You have no justification for invalidating lower ranks because the circumstances are entirely different.

The game is "different" at higher ELOs. I've played with Master + GMs during odd holiday matchmaking and the baseline for basic teamwork is entirely different. They weren't Genji gods. They just did basic sh#t. It's a night and day difference. The other team also had GMs and they were killed. Repeatedly.

This whole perspective of the matchmaking only being broken at one level is entirely stupid because it's applied across the board. It's those subsets of SR rankings that dictate the type of replies.
Oh hey, this is something I literally said months ago:

This includes how MMR/SR are calculated. One suggestion I would make to a revamped system is to have individual performance (determined by certain stats-per-second that would to be too lengthy to get into here but is partially covered by my On Fire critique) matter more at the absolute bottom tier (Bronze or so) and slowly matter less the closer you get to the top. So, if you want to escape Bronze or Silver, or whatever Hell you're in at that moment, you need to outperform your peers in that area and the game lets you know this by making it easier to completely offset negative SR in a lost match. This makes the message loud and clear: "Losses matter less than improvement, but wins greatly help." This way, even if you go on a two or three game losing streak, you don't suddenly have to win six or seven times in a row to offset it.

But as I said, the closer you are to the top, the less this would matter. At the top, the system should be close (or even identical if top players really see no problem with it) to what it is now, because that's the point where you start playing the "real" Overwatch. That's the point where high level strategies and comps and teamwork actually matter and fundamentals should already be mastered. So yes, at that point, there's no more learning from base system mechanics.
07/12/2017 09:29 AMPosted by NeonBard
Part of the problem with that, is according to many grandmasters and the devs themselves. . .picking Torb isn't throwing the match. If someone on your team picks Torb and sticks with it your team needs to build around it.


Are you saying picking torb is in fact throwing? I just lost to a team with torb last night. You yourself pick Junkrat a lot in comp, are you throwing those matches?

Even if you really believe competitive players should be obligated to play meta and/or swap heroes, Blizzard is never going to tell players "picking one of heroes is a punishable offense".

We should focus on fixing people trying to lose, not people trying to win in a way you disagree with.
07/12/2017 08:46 AMPosted by FriendlyFire
I spend a lot of my time here arguing against people mad at Blizzard, because there is a lot of BS complaints. Usually stemming from ignorance. It frustrates me because productive discussion is often pushed behind meaningless whining.

But there are valid complaints about the SR system. In this post, I want to highlight what I believe the real issues are. And specifically, issues that can be solved. (You can't stop leavers from happening, sorry)

1. Trolls and Throwers

While Blizzard takes harsh action against leavers and hackers, players can seemingly throw games at will with no punishment. Some type of community driven system needs to be implemented to punish players for bad behavior (or reward for good behavior). Low priority queue in Dota is a great example of one possible solution.

A unique issue is people reporting those who take off-meta picks. Some legitimately feel this is "throwing", even if the player is trying their best. I'm confident Blizz can solve it. In my personal opinion as a flex player in GM, one-tricks are not throwing and its a valid play style.

2. Performance based SR

The cats out of the bag with the SR system, so to speak. A large group of players understand how the performance system works, and abuses it. Yesterday I literally played with a mercy main who was top 500 with a sub 50% win rate. Not only can it be abused to climb on any unpopular hero, the system itself is broken for low-mechanic heroes like Mercy/Symmetra. It cannot measure their impact properly, and is often rewarding wrong decisions that boost stats.

If they don't want to simply remove it, Blizzard needs to at least turn the weight of the factor down. Better yet, it could be turned off for high ranks (where most of the issues come up) or on low-mechanic heroes like Mercy and Symmetra.

3. Pause feature

After the first few weeks in a season, regular leavers are banned, and a large % of the DCs I see in my own games have players return. The impact of their leaving would be greatly reduced if the competitive system allowed for pauses.

Again, we can look to games like CS:GO or Dota 2 for examples of the system that let a team pause for a reasonable amount of time, but cannot be abused. Tournaments can pause Overwatch, so it's really only a matter of putting the feature into the competitive system. In fact, OW could auto-pause when it senses someone has disconnected, and give them about 1 minute. Possibly with control from the DC team to resume immediately by vote, if they know the player isn't coming back.

------------------------------------------------------------------------------------------

These are clear problems with reasonable solutions. If all three were tackled by Blizzard, I think the competitive system would feel significantly better for a lot of players.

Feel free to add other problems, and solutions, you think I missed. Though knowing this forum, I'd bet most won't meet the criteria of being an actual problem (e.g. 'Elo hell'), or a problem that's solvable (e.g. people leaving).

You'll recieve a like because you're the first masters to actually acknowledge the problem and address it without telling everyone to get over it or call us noobs.
Literally NOBODY knows how performance is implemented into SR calculations. The fact of the matter is that the impact of performance is quite low. the system is almost entirely based on win/loss ratio. I know this because you win/lose around the same SR every game plus or minus 5-7. So to say people can take advantage of the performance based system is blatantly untrue, because blizz has not released info to confirm how it works. And in reality, it is way too complex to figure out just by trial and error.

The other point I want to make is that it is entirely possible to have a below 50 percent win rate in top 500. if you compete at that level for long enough, you will settle around a 50 percent win rate. Unless you are truly an outlier even in the top 500, ie. the Lebron James's of Overwatch.
07/12/2017 11:02 AMPosted by Valkyrie
07/12/2017 08:46 AMPosted by FriendlyFire
I spend a lot of my time here arguing against people mad at Blizzard, because there is a lot of BS complaints. Usually stemming from ignorance. It frustrates me because productive discussion is often pushed behind meaningless whining.

But there are valid complaints about the SR system. In this post, I want to highlight what I believe the real issues are. And specifically, issues that can be solved. (You can't stop leavers from happening, sorry)

1. Trolls and Throwers

While Blizzard takes harsh action against leavers and hackers, players can seemingly throw games at will with no punishment. Some type of community driven system needs to be implemented to punish players for bad behavior (or reward for good behavior). Low priority queue in Dota is a great example of one possible solution.

A unique issue is people reporting those who take off-meta picks. Some legitimately feel this is "throwing", even if the player is trying their best. I'm confident Blizz can solve it. In my personal opinion as a flex player in GM, one-tricks are not throwing and its a valid play style.

2. Performance based SR

The cats out of the bag with the SR system, so to speak. A large group of players understand how the performance system works, and abuses it. Yesterday I literally played with a mercy main who was top 500 with a sub 50% win rate. Not only can it be abused to climb on any unpopular hero, the system itself is broken for low-mechanic heroes like Mercy/Symmetra. It cannot measure their impact properly, and is often rewarding wrong decisions that boost stats.

If they don't want to simply remove it, Blizzard needs to at least turn the weight of the factor down. Better yet, it could be turned off for high ranks (where most of the issues come up) or on low-mechanic heroes like Mercy and Symmetra.

3. Pause feature

After the first few weeks in a season, regular leavers are banned, and a large % of the DCs I see in my own games have players return. The impact of their leaving would be greatly reduced if the competitive system allowed for pauses.

Again, we can look to games like CS:GO or Dota 2 for examples of the system that let a team pause for a reasonable amount of time, but cannot be abused. Tournaments can pause Overwatch, so it's really only a matter of putting the feature into the competitive system. In fact, OW could auto-pause when it senses someone has disconnected, and give them about 1 minute. Possibly with control from the DC team to resume immediately by vote, if they know the player isn't coming back.

------------------------------------------------------------------------------------------

These are clear problems with reasonable solutions. If all three were tackled by Blizzard, I think the competitive system would feel significantly better for a lot of players.

Feel free to add other problems, and solutions, you think I missed. Though knowing this forum, I'd bet most won't meet the criteria of being an actual problem (e.g. 'Elo hell'), or a problem that's solvable (e.g. people leaving).

You'll recieve a like because you're the first masters to actually acknowledge the problem and address it without telling everyone to get over it or call us noobs.

Those are legitimate issues, unlike everything else that gets brought up on here.
07/12/2017 11:11 AMPosted by alexfly
Literally NOBODY knows how performance is implemented into SR calculations. The fact of the matter is that the impact of performance is quite low. the system is almost entirely based on win/loss ratio. I know this because you win/lose around the same SR every game plus or minus 5-7. So to say people can take advantage of the performance based system is blatantly untrue, because blizz has not released info to confirm how it works. And in reality, it is way too complex to figure out just by trial and error.

The other point I want to make is that it is entirely possible to have a below 50 percent win rate in top 500. if you compete at that level for long enough, you will settle around a 50 percent win rate. Unless you are truly an outlier even in the top 500, ie. the Lebron James's of Overwatch.

Wrong, you clearly have no clue of what you're talking about. You're likely not very high in SR and don't experience getting 15 points for win and losing 35 for a loss, but even with you 5-7 SR example, using 20 as a base you could be getting 13 for wins and losing 27 for losses, good luck climbing with that. You also clearly have not seen the data sheets where certain heroes were compared on SR gains vs winrate which shows that the stats blizzard tracks aren't relevant to winning games, neither have you seen sombras start at 4000 SR and climb to 4600 while winning only 40-45% of games by simply farming stats, same for some mercies.
07/12/2017 08:46 AMPosted by FriendlyFire
there is a lot of BS complaints. Usually stemming from ignorance. It frustrates me because productive discussion is often pushed behind meaningless whining.

But there are valid complaints about the SR system. In this post, I want to highlight what I believe the real issues are. And specifically, issues that can be solved. (You can't stop leavers from happening, sorry)


Bump! lets work with what we can and understand there is no perfect system! Only systems that encourage specific outcomes. That mercy with a sub 50% win rate is not what we should encourage!!!!
07/12/2017 11:11 AMPosted by alexfly
Literally NOBODY knows how performance is implemented into SR calculations. The fact of the matter is that the impact of performance is quite low. the system is almost entirely based on win/loss ratio. I know this because you win/lose around the same SR every game plus or minus 5-7. So to say people can take advantage of the performance based system is blatantly untrue, because blizz has not released info to confirm how it works. And in reality, it is way too complex to figure out just by trial and error.


They told us enough. Your performance gets compared to other players on the same hero, same map, with the same SR, on a per-minute basis.

It's not that players have backed out the exact equations, but they've learned what to do. On Mercy, big rezzes get you SR. Doesn't matter that a 4 man rezz can just feed enemy ultimate and a single man rez can win a game.

On Symm, you need teleporters. Shield gen is super strong and win games, but low teleporter stats means garbage SR.

Or on any niche character in general, just perform well. Run around as Bastion:76, do decently, and get rewarded far more than playing a good game as Soldier.

The other point I want to make is that it is entirely possible to have a below 50 percent win rate in top 500. if you compete at that level for long enough, you will settle around a 50 percent win rate. Unless you are truly an outlier even in the top 500, ie. the Lebron James's of Overwatch.


This is where theory breaks down in practice. With infinite players, this would be true. However there are simply not enough people at the top 500 level for this to work. The truth is every top 500 player has a >50% win-rate. Almost every GM does as well. Except for some Mercy mains.

Note: I'd like to specify there are some good/deserving Mercy players at high tiers, but they are currently out-numbered by (intentionally or accidentally) boosted Mercys. As fun as it is spamming their win-rate in all chat when you win, it's a clear issue that needs to be fixed. As I've said, if Blizzard really thinks personal performance is good for the casual competitive player, just remove it for high ranks.
07/12/2017 11:31 AMPosted by FriendlyFire
It's not that players have backed out the exact equations, but they've learned what to do. On Mercy, big rezzes get you SR. Doesn't matter that a 4 man rezz can just feed enemy ultimate and a single man rez can win a game.


Look, this is STILL not proven.

Not once has anyone actually shown BOTH of the following things:

1. Mercy, in general, having a lower winrate than other characters. (Not specific, isolated players/incidents.)

2. Mercys that climb having above-average Rez stats but are below-average in every other respect, including winrate. (Ching, the player that started all of this, did not climb with Mercy.)

No one has ever proven both of these two things, but have jumped to a lot of conclusions based on random screenshots.
07/12/2017 11:59 AMPosted by Succubus
07/12/2017 11:31 AMPosted by FriendlyFire
It's not that players have backed out the exact equations, but they've learned what to do. On Mercy, big rezzes get you SR. Doesn't matter that a 4 man rezz can just feed enemy ultimate and a single man rez can win a game.


Look, this is STILL not proven.

Not once has anyone actually shown BOTH of the following things:

1. Mercy, in general, having a lower winrate than other characters. (Not specific, isolated players/incidents.)

2. Mercys that climb having above-average Rez stats but are below-average in every other respect, including winrate. (Ching, the player that started all of this, did not climb with Mercy.)

No one has ever proven both of these two things, but have jumped to a lot of conclusions based on random screenshots.

Actually, it has been proven quite conclusively over a large sample size on reddit. I can't find the thread, im sure someone here remember the name/has a link and can give it to you.
07/12/2017 12:08 PMPosted by Dicksmith
07/12/2017 11:59 AMPosted by Succubus
...

Look, this is STILL not proven.

Not once has anyone actually shown BOTH of the following things:

1. Mercy, in general, having a lower winrate than other characters. (Not specific, isolated players/incidents.)

2. Mercys that climb having above-average Rez stats but are below-average in every other respect, including winrate. (Ching, the player that started all of this, did not climb with Mercy.)

No one has ever proven both of these two things, but have jumped to a lot of conclusions based on random screenshots.

Actually, it has been proven quite conclusively over a large sample size on reddit. I can't find the thread, im sure someone here remember the name/has a link and can give it to you.


I'm skeptical, because it's relatively easy to cherry-pick accounts with negative winates. (Especially at the start of the season, when the complaints started. Soldiers, Tracers and Lucios with winrates as low as 30% were everywhere around that time.) On several threads, me and other players found more non-Mercy players with such winrates than Mercys.

But, I would be curious to see this evidence and examine it for myself.

Join the Conversation

Return to Forum