If you have played 1000 games and are still worse than the average player

Competitive Discussion
75

I’d be okay to wait. Blizzard would determine, which they already do for MMR, the wait times versus the SR differential.

As I see it, the major handicap is placing low MMR players on teams to offset high MMR players on the same team.

I’m not sure you mean team closeness as in the SR differential between two teams, which yes, I think the matchmaker should use as it’s overarching factor with some variance allowed to deal with long queue times.

If you mean SR closeness within a team, SR differential within a team is fine if it’s a result of a small player base or off times, etc. What I don’t agree with is the system intentionally offsetting high SR players by actively seeking and adding low SR players to their team.

I’m not so sure of this, and definitely disagree with your last point. SR could decay, sure, if Blizzard wanted to offset players who are out of practice or aren’t up on the meta, but this meddling could be avoided as they’d simply lose SR when they played again (same as someone who starts playing an unfamiliar hero in comp).

And I can say, without a shadow of a doubt, the main reason MMR exists, for better or worse, is to allow Blizzard to engineer player progress behind closed doors.

The system I outlined (no MMR, no 50/50, no PBSR, match on SR only, start everyone at 0 SR) is dead simple, transparent and would work great – for dedicated players that are improving and anyone who wants honest feedback from the system.

Blizzard will never do this because it’s not the most optimal system when it comes to fostering engagement, it’s too efficient and stark. Bad, mediocre and stagnant players would stop playing, and that’s bad for business.

76

i started off at 600 sr, the slowly climbed to my sr over the time of around 7 seasons. It just takes tonnes of time. im at around 2100 rn.

77

so what you are saying here is that, if the computer found a 6v6 match that one team had a 40% chance of winning (the lowest percentage of blizz will allow), that if the computer then waits for a minute or two, and within that minute throws a low SR player onto the team that has the 60% chance, raising the underdog’s chance to 45%… this is what you call handicapping.

If we followed your guidance the vast majority of games would be a 60/40 chance, and those are face stomps, and no one likes face stomps.

So we either “handicap” the games OR every game is a face stomp. I’ll take the former.

78

I have 3 accounts and they are all within 150 SR of one another. I feel like I am placed where I belong.

79

The system doesn’t need to calculate chance-to-win at all, and it shouldn’t (or it shouldn’t take them into account). If it matched on one visible metric, everyone would be (within a queue friendly variance) pretty close in rank, and their skills would be a true representation, and random distribution (at least between the two teams), of the skill levels that exist at that rank.

80

ok, so in the wee hours of the morning, there are only 6 GMs in queue and 6 bronzes in queue. How does the computer decide who’s on who’s team? does it just randomly assign them?

81

It shouldn’t match outside of ranks. Though an argument could be made for getting rid of ranks altogether. The queue you outline doesn’t yield any matches relevant to a competitive ranking system for a team based game.

82

no, but answer the question… Let’s throw ranks out the window, no more ranks.

It’s 6 oclock in the morning and there are 6 people with SRs of 732 and 6 people with SRs of 3954. How does the computer decided who’s on who’s team?

83

As I said previously, Blizzard would have to decide on an SR variance allowed for matches, outside of that variance matches would not be allowed. They already do this. In the proposed system if there’s no players in queue within say 200 SR, then you don’t get a match in comp, go play QP where they could remove the variance restriction.

And in return, I can ask you, do you think a match of 3 Bronze and 3 GM versus 3 Bronze and 3 GM will yield useful results as it relates to a competitive ranking system?

84

Okay, very well. It’s 6 oclock in the morning there are only 12 people on. Six of them have an SR of 1400 and six of them have an SR of 1600. How does the computer decide who’s on who’s team?

85

Rando yo, rando. If you were to intentionally even out the teams the results of the match would not be all that useful for ranking (that’s why they added PBSR).

86

hold on a sec, let me continue with my questions then I’ll answer yours

87

I condemned calculating win chances for the purpose of match making. Not looking at historical win percentages – though I’m not sure what that has to do with the conversation. The win/loss data resultant from that proposed match would be nothing but noise (again why they introduced PBSR, because the system is broken otherwise).

You’ll have to continue on your own for a bit, I appreciate the discussion, I’ll check back.

1 Like
88

I would be fine with Rando. But if the computer wants to organize the teams in such a manner as to get the SRs of each team as close as possible, I would say this would expedite SR movement. As with the random way it would take longer. We’d get to the same place in the end, speed is the only difference.

There must be another metric than just throwing a crappy player on to a high SR 5 stack to lower the team SR. It’s gotta be something more than that, and if it’s not, yes, I agree with you…that just feels wrong

1 Like
89

Your theory is sensible, but it is not consistent with the system as implemented, in a few key ways.

This is probably approximately correct, though Blizzard has never given this level of detail.

There used to be win streaks bonuses/loss streak penalties, but they were removed because they tended to throw people far from their proper rank, through random chance alone (Overwatch Forums).

Indirect evidence suggests that MMR moves slower than SR.

  1. Since SR chases MMR, that implies that MMR moves slower (otherwise SR would have a very hard time catching MMR).
  2. During placements, the average movement per game is +/- 19 SR. I suspect that this is actually the average motion of MMR. See How Competitive Skill Rating Works (Season 11) → Season Transitions, last paragraph.
  3. If SR is well above MMR because a player has won many more games than he’s lost, less SR than MMR will be gained on a win, and more SR than MMR will be lost on a loss. This is an SR debuff. At high win percentage, a player can gain up to 6 SR less on a win than he loses on a loss.(Overwatch Forums).
90

You didnt earn anything.

You started placements and performed badly for 10 games, your SR went down from 2500 extremely fast (but hidden from you due to placements).

Then you placed low gold and kept falling to where you belong.

91

I tryharded qp on this alt and was playing mostly diamonds/masters yet my first placement match was a low plat/high gold average game lol.
Went 10-0 and ended up placing barely 3.5k back in s8.
QP MMR does not affect comp at all.

2 Likes
92

You’re not a very nice person.

93

I am just telling you the truth. The same thing happened to me on my first try on my first account.

Its pretty common. They really should just start everyone at 1k and rank up outlier players quickly during initial placements, rather than creating the salt they do by placing the majority in gold and having 30% of them drop.

1 Like
94

Eventually people will figure this out. They clearly changed it. But people keep using old info.