Yup, and I already mentioned why these games are many degrees simpler than HOTS and how they play far fewer games, have fixed teams against other fixed teams where there is one ball. There are thousands more metrics in HOTS and trying to determine what is and isn’t useful is incredibly subjective, especially as you determine what is and isn’t useful at various skill levels. If you still can’t see why distilling a few metrics and why HOTS is more complex, when W/L is just more elegant than you’re just not making accurate comparisons.
I’m not ignoring it at all. I’m saying the vaaaaast majority of teams aren’t able to utilize the trivial benefits that MIGHT come from playing together. The biggest determinant of success is skill level of the players. A group of bronzies don’t magically become gold players when grouped. You want to believe there’s a difference, the data just doesn’t prove it. You’d rather blame the matchmaker and feel sorry that there is a perceived advantage. You’re pointing to outlier cases and thinking that’s representative of most games. It isn’t.
In most cases MMR, is really all you need. Now, I think repeat teams can have a possible adjusted MMR, but that requires a great deal more data points to store, but could possibly be beneficial to account for those kinds of teams, but most likely they would have been doing well anyways because surprise, they’re probably already good players to begin with.
That’s what it feels like to me.
It goes without saying that I would rather stand in queue for 10 minutes than be weighted down by potatoes.
Yesterday I had a game with a Butcher, who at 7 minutes had 550 hero damage. That, while having 4 deaths. He finished his meat quest at level 22, only because we hard-carried and won a few teamfights for him.
I liiked at his profile afterwards, and he’s sitting at 46% winrate, and he had 299 games on Butcher.
So that is a player that is clearly much lower MMR than myself or any of my other teammates, somehow appearing in our game with abysmal performance and winrate.
We’ve won basically due to a miracle. A play that was way above my own skill, personally, but just happened to luck out. If not that dumb luck - we’d lose hard. We were losing every fight and every objective until 16, and all map was theirs.
So I ask you - what happened? There really is two possibilities:
Somehow that 46% winrate Butcher’s MMR got high enough to match mine
Or our team got weighted down and the only thing Matchmaker cares about is average MMR
We are wired to try and see patterns so that we can act on them. Sky gets gray, find shelter, it’s going to rain. You streak, eventually you get an opposite difficulty. In this case, unfortunately, this pattern can be demotivating and thus upsetting. I got electrocuted in bright sunshine, damn it!
There are outliers, and the ranking system combined with smurfs isn’t exactly bullet proof. People in Bronze are often more skilled than those in Gold. Discussing picks, understanding meta, more soaking than Gold, and 2-3x as much hero damage compared to GM matches. People do have 30-40% winrate over hundreds of matches in ranked. (I have made it from legit Gold 1 to legit Bronze.) You don’t see it too much because these either smurf, quit, or it stabilizes later. The opposite cannot happen to everyone since you’ll climb.
@Aax above: Sometimes players are super hard countered, be it skill or hero choices. Quite possibly it was a match which didn’t work for him. Hammer is also a super OP hero - but do you think about her that way? Specifically, a Valla one match can have 100k+ hero damage by 20, in the next, 10k by 20. Not a joke. Been there, done that, seen that, even from 60-70% winrate people.
APPARENTLY THE GOAL OF THE MATCHMAKER IS TO MAKE “FUN” MATCHES NOT FAIR MATCHES.
45:20 “The approach where it’s still kinda 50/50 but you don’t always give a 50/50 match, you give a variety of matches, you have some smart way of saying no for this match i’m going to give him actually a hard match
for this next match i’m going to give him an even match for the next match i’m going to give him an easy match.”
45:35 “it gets to the point when matches are super even all the time after a while it actually get kinda tiring…so i come up with the opinion that its ok to allow a little bit of balance here and there and have a mix of even hard and easy matches.”
There is an individual QM MMR and it has nothing to do with your account level. Someone could be level 5000 and still be matched with complete beginners because they are this insanely bad or incompetent. Also QM is and always has been a complete coin toss. Comps are random and most people don’t care about the outcome of the game. Also a lot of people in QM are only (or almost only) playing QM and have no idea of the actual fundamentals of the game.
its like asking why the sum of a triangle angles equal 180
its forced by mathematics, because its how mathematics works
if we toss a coin a lot of times we get an average of 50% (call it the law of large numbers)
but we cant use this as an explanation for the outcome of the individual tosses
if you play a single game and win you have 100% win rate
it does not mean you will lose the next one to get 50%
on the other hand, if you are not in your league, you will keep losing/winning until you get to your league then you will see your wins and loses getting even, its just logic
My friend sent me an example of the same MMR averaging you outlined.
Can you guess who won? Oh yes of course it was the team with the two GM’s, in an 11 minute game. I watched the replay my friend sent, a complete stomp even with the Azmodan and Qhira feeding. I feel very sorry for the single Master Valla player who was expected to carry her team. My friend also said the enemy team without the GM’s was favored to win by MMR averaging, so he picked up extra points for the win!
I don’t believe in forced 50% never have, but MMR averaging, especially for ranked games is just absurd. Note, this was SL, not URD!
So if i win 10 matches in a row, climb from let’s say Platinum to Diamond as a resault and then start to loose every single match back into Platinum it’s because Platinum is my intended rank?
What if i’ve climb leagues just like that but objectively observe/notice that the taxing losing streak wasn’t because of me and me being dragged down to my “intended” rank but that it’s because of the mass hoarding miss-matchings of brain combinations in the team causing 0 chemistry again and again?
IMO There’s your “intended” rank and then there is Luck (the coinflip luck).
And because the coinflip luck in this case ain’t set to garantue 50% but is “probable” to garantue it, you could place a limiter and toss a coin 100 times and get HEADS 90 times and TAILS 10 times making it 90% HEADS chance for that particular toss count.
You could get a higher percentage enough times to get you out of a league to a league where everyone is very skilled and you manage to fit in even thou it’s not your rank
and im saying “manage to fit in” because you might not be nowhere near as skilled as the other 4 in your team but you get such good combinations of them that they in a sense keep on carry you long enough for you to learn by repeated matches to practice high rank matches and improve your skill as a resault. and all that because you got luck on those 100 coin tosses.
The below could be the case for 50% of the player base.
Lucky -> Low Skill = Rank up
Skilled -> Unlucky -> Rank down
This is how i experience it with my second paragraph above as refrence.
It’s not. However, the matchmaker already accounts for that by adjusting the matchup (5-mans face tougher opponents) slightly. And despite their supposed advantages they do not actually perform better, so apparently things balance out.
It’s also not as if every 5-stack is a fully coordinated killing machine on voice comms with perfect synergy. There’s plenty of players that just queue together because they enjoy playing together. Similarly, it’s not like every solo-q is full of morons that don’t know how cooperate.
Not surprising. If it was a single GM with those allies, I’d have bet on the team on the right. Alone, that Master has absolutely zero carry potential with the teammates he was assigned, against two Grand Masters.
This type of match shouldn’t exist in the first place though, a literal Bronze has no business playing in a GM match and vice versa.
I think the MM believes the Bronze player makes it “even”, but not even close. The Valla player was actually very good, but her teammates were completely outclassed.
No one says equally.
The good formula would be a decreasing value proportional to how old the game is. And the period that should have the most weight should be WAY greater. Currently, it feels that I have 5 win streak, then 5 loss streak, give or take, and it is repeating over and over. This means, that if we presume that it’s the workings of MMR rising and falling, that the last 5-10 games hold the most weight. Which is just STUPID.
To calculate MMR well, an average of your performance in the last MONTH should be taken. At the very least.
It is all very well to point out that match making can’t make completely fair matches, but what alternatives are being proposed?
Should MM track your MMR for individual heroes? Should it then lock your heroes in for you to ensure an even match? Or should it swap players between teams as heroes get picked?
Should it track you’re win/loss rate vs specific heroes and do your bans for you too?
What about playing with and against specific players? Should it rearrange the teams and give you hero a with hero b played by player x vs hero c played by player y?
Should you be forced to fill out a questionnaires between matches that tracks your work, sleep and eat habits and adjusts your MMR if you are stressed or hungry or tired?
I’m being ridiculous, of course, I’m not suggesting that any of this is being advocated for.
But all of these factors do have huge impacts on our performance in any given match, and it really is on us as individuals to be aware of these factors, before we start looking for much more complicated systems to match our biases.
All it takes is one good player on a strong hero, that wasn’t countered in draft for a match to feel unwinnable. Throw in a tantrum and the match is definitely lost, none of which the matchmaker can do anything about.
It is Blizzard politic,I am also faced with this problem,for example in every 10 match,5 win and 5 lose decussate.
At all,if you have ever 10 win Consecutively,you should have 10 lose in future,it is Blizzard program,they finaly force you to have 10 lose even they match you with AFK or leaver people!!! It is a clear issue and can not be hidden.
1 - Left team is probably in a 5 stack.
2 - If they aren’t in a 5 stack, both GMs are queueing in a time where almost no players are queueing, which means they were probably in queue for several minutes (like 20-30+).
3 - Low pop server, which has rainbow matches all the time duo to low pop.
Then it the match happened when no one was playing. It is normal to have these rainbow matches in low pop times. It has always happened just because there is now way to find balanced matches. Also since both GMs are in a group, it is even harder for the MM to find a match at that time, so after a really long long queue it just puts them in a match so they can play, otherwise they will be stuck in queue forever.
Like, the same would happen if I queue in my server now. Lots of rainbow matches with really long queue times (15+min), but if I queue at night after work, it gives me balanced matches.
Then “dead game” is not a meme, if skilled players have a very hard time finding balanced matches (or if they can’t be found at all unless playing at the exact peak 2-3 hours of the day).
It is true to some extent.
I wouldn’t call it alive if even Platinum must play with Bronzes due to lack of more suitable allies and opponents. Not to even mention Grandmasters.
Just because a match starts with such rainbow ranks all over the place does not make it alive except technically.
Specifically, a Valla one match can have 100k+ hero damage by 20, in the next, 10k by 20. Not a joke. Been there, done that, seen that, even from 60-70% winrate people.
