Insightful commentary on ranked by Jeff

Competitive Discussion
18

I’m saying forcing even teams does not keep people from ranking up or down.

and this is preceicley the problem. At 3am on a Tuesday. the 12 people it will select will be 500 or more SR apart. the computer needs to balance the teams.

check this out, you know a the computer won’t create a match that has a greater than 60% win rate, correct. so there’s a way to get 12 people when shuffled someway it’s greater than 60 but shuffled another way less than 60.

listen, if you want to find out any more of what I think about matchmaker (I think it’s pretty good btw) let’s talk on comms. typing at length is cumbersome.

19

It doesn’t prevent them from ranking up or down, but it makes it far more difficult to rank up than it should. It makes for a confusing, frustrating, bizarre matchmaker that most people feel “something” is wrong with but they can’t put their finger on it. People are playing in matches where they are obviously significantly better than the people on both teams, except for one or two of the players on the opposing team which are there to even the teams out. Let’s put it this way, I can halfway try and stay at the same rank as I do when I try as hard as I possibly can. The reason is because when I halfway try (not throw) for a few matches, the matchmaker realizes this (by analyzing my in game statistics) and places someone on my team to make up the difference. So what’s my incentive to be the person who tries hard? None.

That should be the exception, not the rule. If after two minutes of trying to find people of similar SR to place on each team they are unable to do so, then it should change over to team balancing.

1 Like
20

no, the matchmaker doesn’t know your past SR

at least there has been no official statement on this.

edit, well maybe it kinda does because there is an unsureadness variable in play, but regarless

you know all the facts but come to the wrong conclusions, you’re a deciple of the one who shall not be named :slight_smile:

21

I value contributions like this. They remind of how little some people understand of both what they responded to, and what they are talking about. Especially when they give a contradiction within the same post, and proceed to a veiled insult.

22

let’s get to the point, kabaji can climb, you can’t, nothing is wrong with the matchmaker

the only flaw it might have is that it is not accurate enough, but to think that it actively keeps people down?? how does it decide who it keeps down, the examples given happen on both ends, but only one end is used for the argument.

1 Like
23

You now subscribe to the “performing well/poorly causes the matchmaker to balance you out” theory? You previously have subscribed to the “performing well causes the matchmaker to rig games in your favor” theory.

Have you changed your mind since last week? Or do you not realize that these two statements are directly contradictory?

3 Likes
24

You really need to work on your reasoning skills. There are a vast number of ways that a matchmaker can be flawed, but still allow people to climb.

2 Likes
25

The way I say the matchmaker works is the same way I have said it has worked for a very long time, and it’s the actual way it works:

You are placed in matchmaker created games where each side has as close to a 50% chance of winning as possible. If you play well then you will be matched with scrubs in order to make the teams fair (handicapping), you must carry them just in order for the teams to be fair. If you play poorly then you will be placed with better people in order to make the teams fair. The matchmaker continues like this unless you:

  1. get really bad stats individually in which case you will be placed with other people who under performed and will rank down on a loss streak. In these games you could try your hardest and you would probably not win.

  2. get really good stats individually in which case you will be placed with other people who over performed and will rank up on a win streak. In these games you don’t even have to try in order to win.

that’s exactly the way the matchmaker works, you know it and I know it. Our time would be better spent talking about why this is so bad for matchmaking and why people hate it.

2 Likes
27

Why though? There is still QP, Aracde etc which wouldn’t force 6 stacks. There is also nothing wrong with being a scrub: http://www.sirlin.net/ptw-book/introducingthe-scrub

Let’s consider two groups of players: a group of good players and a group of scrubs. The scrubs will play “for fun” and not explore the extremities of the game. They won’t find the most effective tactics and abuse them mercilessly. The good players will. The good players will find incredibly overpowering tactics and patterns. As they play the game more, they’ll be forced to find counters to those tactics. The vast majority of tactics that at first appear unbeatable end up having counters, though they are often quite subtle and difficult to discover. Knowing the counter tactic prevents the other player from using his tactic, but he can then use a counter to your counter. You are now afraid to use your counter and the opponent can go back to sneaking in the original overpowering tactic. This concept will be covered in much more detail later.

And there is nothing wrong with that, they can still play comp, they just won’t be successful.:

Let’s return to the group of scrubs. They don’t know the first thing about all the depth I’ve been talking about. Their argument is basically that ignorantly mashing buttons with little regard to actual strategy is more “fun.” Superficially, their argument does at least look valid, since often their games will be more “wet and wild” than games between the experts, which are usually more controlled and refined. But any close examination will reveal that the experts are having a great deal of this “fun” on a higher level than the scrub can even imagine. Throwing together some circus act of a win isn’t nearly as satisfying as reading your opponent’s mind to such a degree that you can counter his every move, even his every counter.

Can you imagine what will happen when the two groups of players meet? The experts will absolutely destroy the scrubs with any number of tactics they’ve either never seen or never been truly forced to counter. This is because the scrubs have not been playing the same game. The experts were playing the actual game while the scrubs were playing their own homemade variant with restricting, unwritten rules.

It would only be enforced for competitive, just like at your local beach volleyball competition, you can’t enter unless you have a team of 4 people. Sure you can just go play at the beach with less people to have fun (QP) but of course there has to be a minimum rule for comp.

Similarly there is a general concept in software design called the ‘pit of success’. The idea being that you design your software in a way it is easy for people to fall into the right patterns.

At the moment that is not the case with Overwatch and we can see where full freedom in competitive has brought us. People turn on each other and blame the system, symptoms of a failed system.

I just fail to see why just like almost any other team based competition, why the minimum requirement wouldn’t be to have a valid team. Sure I guess you could just turn up and sub for a team with a man down, but the current system in Overwatch just doesn’t work.

(*) Please note the italicized text is taken from the link in this post

1 Like
28

Sounds like they’re moving towards breaking down SR by role and providing more after-game context (hopefully a scoreboard, graphs, etc) as to what contributed to your SR gain vs what negatively affected it.

If that’s true, I’m hyped for these changes and OW comp will be moving into the right direction!! Still not convinced it will mitigate the smurf/thrower and toxicity problems that suffocate the ranked experience tho…

P.S. LFG isn’t bad at all bad… in theory. In practice, I get laggy games (90+ ping occasionally) and filling the group takes well over 5 min. Queue times another 5… unfortunately, not worth the wait either since often times teams are widely unbalanced (more so than soloq…). Comms aren’t everything Jeff.. LFG search does worse job finding equal MMR teammates than soloq…

1 Like
29

Not really ofc. But if there is simple/easy solution and blizzard are just stupid enough to see it, then they should prolly share it. Its easy right? Why cant you explain if its soooooo easy?

1 Like
30

So are you saying something like this?

Imagine a 2500 SR player, Joe. His average MMR is also 2500, but MMR moves very quickly.

2500 SR, playes well for a few games, his MMR goes up to 2750. The game matches him with 2500 SR / 2250 MMR player to keep things fair.

He then plays poorly for a few games, his MMR drops to 2250. The game matches him with a 2500 SR / 2750 MMR player to keep things fair.

He then plays a game or two of his life. His MMR shoots up to 3000, and he is put in games with 2500 SR / 3000 MMR players, against 2500 SR / 2000 MMR players, leading to total stomps in his favor.

31

For many players, solo-queue competitive is their favorite mode. If you remove someone’s favorite mode, they will likely quit. They don’t want to be shoved into the quick play ghetto. This is either self-evident to you (as it should be), or it is not.

I’m familiar with play-to-win, but I’m not sure it really applies to this discussion, as he is focusing on 1v1 games, and you are complaining about your teammates. I doubt you mind much if the enemy has scrubs.

Nothing is keeping you from making a valid team. Why don’t you LFG or form a friends lists / guild on your own?

32

You don’t think enough people are leaving/have left already?

1v1 can be extrapolated to teams. This is either self-evident to you (as it should be), or it is not?

I’m not, please find the reference in this thread. This is either self-evident to you (as it should be), or it is not?

Correct, easy wins.

I think you have misconstrued the proposal and are trying to suggest I personally have a problem in that regard? In fact your response is what I tell many other forums goers when they complain about issues with team-mates. Myself I’m happy enough as is.

I am simply trying to offer a general solution that:

  • Removes a large percentage of the complaints about any kind of toxicity in game and on the forums
  • Encourages people to take responsibility for their performance because they would have no-one to blame but themselves
  • Trying to present an analogy like many other competitive team events. I thought the beach volleyball one was ok and thought that This is either self-evident to you (as it should be), or it is not?

This proposal, which I have made in many other threads doesn’t seem to take so it would seem to indicate it isn’t the ‘pit of success’ as far as game design goes.

This is either self-evident to you (as it should be), or it is not? (*)

(*) I thought I would use this from your above quote as it seems like you like using it to emphasize your points and perhaps it can aid you in understanding my response?

33

Hmm, that’s a decent way of describing it, but I doubt the MMR variation is that wide between the players he is playing with. I would imagine it’s closer to something like :

2500SR/2500MMR plays well for a few games, he is now 2600SR/2700MMR The game matches him with 2500MMR/2600SR players to keep things fair.

You would think that the skill difference between a 2500MMR/2600SR player and a 2700MMR/2600SR player wouldn’t be that much (200MR difference), but it is because a 2500MMR player with 2600SR indicates that the player has been playing below his current SR threshold since his MMR is lower than his SR (possibly far below), where as our 2600SR player with 2700MMR indicates he is playing well and playing beyond his SR (who knows how far beyond) since his MMR is higher than his SR.

So, the key takeaway is this: a 2600MMR player with 2700SR indicates that the player has been playing below his current SR threshold (possibly far below) and so the matchmaker should not pair him with someone who is say, 2700MMR with 2600SR because the max possible skill variance is far too high and will lead to frustration.

34

The team must believe. If that was done. Than team only mode would fall apart. The Q times would be so high. People would abandon it and just go solo.

35

You will always have different metas on various ranks but the important thing is to balance it from the top, mainly for the pro scene because if they would have balanced it for your average joe player (which would be 2.2-2.4k I believe) you would have the most one-sided and broken gameplay in the higher ranks.

And yes, 2-2-2 is the most common meta, usually with main tank like Rein and off-tank like Zarya/D.Va, a short-ranged dps with a flank/backline harass purpose such as Genji/Tracer or a more situational hero like Junk, Mei, Reaper etc. paired up with a more medium/long range DPS such as McCree/Ashe/Widow/Hanzo and then you have two healers usually a main heal and a flex healer (or commonly refered to as off-heal) which means you combine a healer with lots of general heal output aka Ana (or Moira on some points) with one of the other healers that offers less heal in general but have something that make them worth picking, Mercy for Res, Lucio for boops, Zen as ulti-counter & focus fire, Brigitte because she’s broken etc.

Now GOAT do see play, a lot of it, but it’s nowhere near as popular as your average 2-2-2 set up and just like Sym/Bastion plays which increased in popularity lately it’s still extremely rare if you look at the playtime it sees compared to “normal comp games”, most deviations from 2-2-2 come as a response of having to go all in. Like you should start a game by keeping in mind that the opponents might run something like GOAT but most of the time they wont and those games tend to either be 1) very short due to GOAT comp steamrolling through the points 2) People adapt and force GOAT to switch 3) GOAT picked as an all-in effort when things looking grim so the overall minutes on GOAT is low, however it doesn’t mean it’s not having a bad effect on the ladder and pro scene because few people actually seems to enjoy playing GOAT, playing against GOAT is even more frown upon and watch a game with GOAT is rather lacklusting. Now as long as such oppressive meta is in the mix everything else have to adapt to it in one way or another in order not to get extremely punished by it so something can be meta yet not played that much.

36

Spamming this line makes you sound angry, or tilted, or … something. You certainly seem beyond the point where we are having a useful conversation.

The line is not used to emphasize a point. To say a statement is self-evident, is to say that it is axiomatic, is to say that it is a first principle, is to say that it is difficult to argue about, because you either accept it or you don’t. It is not an argument or a conclusion, but a premise.

If you say that everything you are writing is self-evident, and I disagree, then we really have nothing else to talk about. At this point, you can either rewrite your last comment, or we are done here.

1 Like
37

I could narrow down the numbers.

But anyways, this system as described is completely ridiculous and absurd. Sometimes it rigs matches to 50/50, sometimes it rigs matches to 80/20 (or whatever) depending on some sort of detailed statistical analysis of play.

It is massively easier and more correct to group people by MMR, increase MMR when they win, and lower MMR when they lose. Fortunately, that is what Blizzard does.

1 Like
38

And here in lies the inherent problem:

  1. He’d like to fix things
  2. Nothing major to announce, maybe sometime in the distant future
1 Like