Furthermore it’s a fact your WR improves if you as a player improve. Even if it doesn’t your MMR may increase. Why do you think there are sub-40% platinum and such players around? Because the matches they won meant more than the matches they lost.
1 Like
That’s true, that’s how it works.
The issue with HOTS, however, is that MMR here feels too spiky. It seems to put too much value to your recent games, so just a few wins makes your MMR skyrocket, and you’re matched against people that are much higher skill than you, and you lose.
Then you lose a few time - and MMR plummets, and you’re on a win streak again.
The weight of wins and losses is too high, and your system seems to struggle finding any semblance of a middle ground, so you’re always on that roller-coaster of winstreaks and loss streaks.
This also makes it so that when your MMR is high - sometimes you get given teammates that play EXTREMELY bad. Just because they had a recent win streak and now they’re in your MMR bracket even though realistically their skill level is much lower. So you get that problem of “Losing because idiot teammates” which is the result of the same spiky-ness of the ranking system.
That’s a real problem, and even after hundreds and thousands of games it seems that the only thing HOTS cares about is my recent 5-10 games.
The “50% winrate” is a result of an MMR. THat’s true.
However because the MMR is highly flawed, it feels unfair.
The unfairness feels only stronger because of other two factors:
- A Plague of AFK/Ragequit/Nonparticipation players that Blizzard doesn’t do ANYTHING about. They don’t get banned, they just keep playing and ruining games to others.
- An idiotic matchmaker that follows incorrect ruleset. For example, it matcher 5-man premade vs 5 randoms, because the premade queue got a bit too long - which creates an ABHORRENT frustrating experience to 5 people and unfulfilling mediocre one for the premade.
Secondly, the way Matchmaker matches Counters needs reworking. I often see matchups where one team literally is a 5-man hard counter to the other team. The matchmaker doesn’t actually care how much counters either of teams has, which results in frequent 4-5 counter setups. Worst of all, you can sometimes see how re-shuffling a couple of players could make the match fair and interesting, but the MM doesn’t look that far. It just looks through the table and finds Mirror or Counter pairings and then randomly assignes them, it feels like.
Answer me, AZJackson, does it feel like Blizzard is aiming to create the most fair matchup when they match a premade vs randoms?
Does it?
No.
It feels like shorter queue times is much more important than fairness or fullfilling experience.
The problem is not that Matchmaking is unfair by design.
The problem that it has massive oversights in its design, and it feels (and actually IS) unfair de-facto.
It is underdesigned. That’s the problem.
It has extreme flaws that make it unfair.
2 Likes
Yes. You should never look at your mmr as a constant, fixed value. It’s more like an interval. Always going up and down. But there is a range, which represents your average performance.
3 Likes
This holds truth even if I have an MMR of 2800.
I had a match where my teammates had on average 2100 while the enemy had 2550.
https://www.heroesprofile.com/Match/Single/?blizz_id=4816225&battletag=lucifer®ion=1&replayID=27946219
The disparity between the matchmaking from QM and SL can be so huge, even when no one is above 3k MMR.
At least in QM, it’s nice to find a match so quickly, haha! /s
1 Like
Why are professional players in a sport not simply rated based on W/L, but on a multitude of stats such as (just for example) pass completion? I think the reasoning is pretty straight forward.
“Proven” huh? IDK lol… You have a 5-man team on coms vs 5 randoms, and you are going to argue that the randoms stand a better chance of winning? I think if that is in fact Blizzard’s stance, it explains a lot of issues.
3 Likes
Human sports have managers and scouts evaluating individual player performance though. They draw and recruit the best of the best from a pool of a couple thousand professional players.
Try doing that in a moba with tens or hundreds of thousands of players (in the case of more popular mobas - millions)…
Rating individual impact in HotS would require human resources. A real experienced HotS player (preferably a Grand Master) going through each replay and rating performances.
Because it’s freaking obvious?
Professional players in a sport don’t have to deal with minions, talent selections, map varieties, player synergies with 80 heroes, different player sets, graduated power scaling, and multiple skirmishes on an ever evolving map dependent on player location, objectives, and ancillary localities of interest that can be combined and collapsed upon in real time. You can’t “moneyball” HOTS. This isn’t baseball where you have players going all in the same direction and each person takes turns hitting with a defined role where you do relatively the same thing.
It’s not a stance. It’s data. Just because you don’t like it because it clashes with your self prescribed ideology doesn’t make it less true.
It’s already rated and recorded though. It just does not count… You know that little star that pops up at the end of games to tell you that you did better at X than the average, or that you did better than you usually do? Well, they’re writin that stuff down somewhere.
A good example for sure.
That may be the goal and the intention but that hasn’t been my experience at all at least when it comes to QM. Once i go on a win streak and losses start i am still the best player according to every stat on the stat sheet but i cannot carry the players i was given because they don’t seem on my level. I dont think the system intentionally does this but i also feel that its not working exactly as you intend it to either because i truly don’t feel that i am the worst person on my team when i get on those loss steaks. ( i know everyone thinks this lol )
Isn’t it possible that the population is so low now that the systems average MMR starts to weigh you down with worse team mates if it cant find proper MMR range for you fast enough? Not so much that its purposely putting potatoes on my team its just the only way for the MM to make a quick game.
2 Likes
I mean, that’s a whole other thing. You are speaking about leagues/tournaments/teams.
When we had HGC, players where going into teams according to how they played, stats and everything. This is no different than a real sport league.
The thing is, if everyone was a solo player in sports and there was a system to put them together in a team to play against another team, W/L (mmr) would be the most valuable stat taken into account.
2 Likes
Can you also explain the samuro I had last game? he went afk/inted after doing 8k siege in 15 min, how is he anywhere close my mmr+logans who is a master?
2 Likes
Let me guess - he saw my TLV so he decided to not waste his time to try to win.
1 Like
He replied that to my kind whisper “lost because of vikings, dumb QM”
2 Likes
Of course, that is what is happening. So it is the goal and intention, and also your experience; because any MMR is predicated on having critical mass.
1 Like
When people decide to troll, the system just can’t do anything. How that person got there? Well, probably just another person boosted in TL who got a high rank when SL merged, or just boosted by friends in SL. Same thing for QM/UR.
1 Like
So TLV has a passive to off my teammates brains? How wonderful.
And that guy didn’t have nuts to say me those words by himself. Coward.
2 Likes
Generally speaking though there’s a positive correlation between stats and how good someone is. More often than not, the better the stats, the better the player. However there are many situations that make this untrue as well. There are too many variable (composition, map, and each player). For example Leoric might get value out of dying, but not if there’s a Butcher on the enemy team who will just stack off of it. Same with Murky.
I also said that stats are not everything but they can be a good indicator of who’s good and who’s not. But just that, and indicator, nothing more. You are able to see this first hand in game better than just through a stat sheet.
heroesprofile is not a good metric for trying to appraise mmr.
For one, they actively point out they’re trying to recalculate how they do MMR.
Second, in the hour since you posted that, your team of 2100 is now appraised at 2258; the values it puts on players does change and will continue to change.
Third, the post-match mmr changes are in a range that suggests the system is less confident on the appraisal of that match and is likely to be more volatile than what the change would have been live.
Fourth, profile doesn’t have an equal amount of games for all the players in that match, some of them are very low, so it’s estimation of their mmr is not going to be all that great.
The More You Know!
Stats don’t mean much. Take WoW for example right now in mythic+, the hyper-infalted values from corruptions have gotten out of control that getting a 4k+ score is not even impressive currently and there are many, many trash-bags with “good” scores.