This just popped up on my Youtube feed and it is a very interesting watch. It was posted yesterday so it is the most recent one. A lot of questions have been answered and a good insight into OW matchmaking has been given.
h ttps://www.youtube.com/watch?v=d_Ut8pCH9QM
(if someone can post link please post it)
Itâs an interesting look behind the scenes but wonât change the minds of people who already distrust them (they wonât watch it at all)
if your matchmaking is good you dont need to keep going two times a week and yaaap for hours with some random sweat tuber about how its not your fault but players,. .ubb ubb u muggles they dont understand.
these are backend stuff which if functioning properly need zero explanation and thesis papers.
you have a good matchmaker when nobody knows who your matchmaker dev is.
https://www.youtube.com/watch?v=6s-6q0n7izM
These are basically what my games look like. It really is bad that I keep running into players like these that donât have any clue of how to handle this.
Thank you for sharing cause I donât have X and I donât watch OW content creators so I would have never seen this one.
Its an interesting one as we learn that now, Displayed sr IS mmr (except for demotion protection).
They also gave a better explanation of the matchmaking âpredictionâ.
And I kinda feel that some problems actually donât come from the matchmaking but rather on whatâs going on during a match which has always been a problem in Overwatch : thereâs no clue to know what youâre missing out while playing. You might have to look back on the game and try everything, putting more randomness inyour upcoming games. Making MMR varying, making it longer for a player to improveâŚ
The main problem with the matchmaking isnt actually the matchmaker, its how the game ranks people and pretty much the entire comp system that is the issue here. But you cant see any of that from the raw rank data they use. You need to look at performance metrics which clearly blizzard do not care about when its arguably the most important metric when assessing matchmaking
They keep looking at useless data like how closely ranked people are of a certain rank in a given match, what % of games that are matchmade features player of a wide/close range etc.
They see things like â80% of matches made have players within 1-2 divisions of each otherâ and thus believe the system is working accurately cus its putting a large amount of people of equal rank into the same match so it must be working right?
RIGHT?
What happens when you separate a players rank and a players skill thoughâŚ. The system immediately falls apart.
Just because you are putting 10 people of equal rank into the same match DOES NOT MEAN you are accurately putting 10 people of equal skill into the same match.
This is most often felt when you load into 1 game and everyone is playing normally, then you load into the 2nd game and it literally feels like you are playing with people who have literally 0 clue what they are doing.
The game is saying those 2 matches had players of equal ranks in them but it clearly highlights players in that rank are of such widly varying skill levels. Which means the ranked system isnt working!!!
If you can have players of such drastic skill level, where 1 teams dps is 30/2 and the other teams dps cant even get 1 kill, the game is not ranking people properly. A dps who cant get 1 kill LITERALLY cannot be the same skill level as a dps who gets 30 kills. But they are the same rank according to the game and thus get matched into the same game
How can a tank who runs in by themself every fight be the same rank as the tank who is playing out of their mind?
Make it make senseâŚ
They said 6x/3x % games are fair competitively.
There is no loserâs q
The matchmaking is perfect and the data says so.
Basically is what I gathered from it. Some matchâs are 60/40 but âthatâs is still competitiveâ âŚâŚ
Basically âŚâŚ. They canât do wrong
And howâs itâs to hard to track stats for players
Yes because I should totally believe the devs of a company known to be constant liars and promise breakers. âItâs all in your head, the matchmaker is perfect guys! Trust us!â I donât think I will. Trust is earned, and they have broken trust too many times.
One of the main âknownsâ that Overwatch is famous for is the atrocious match making.
It is probably the most recent update weâve had on matchmaking, it is from a week or two ago though. Regardless, itâs always good to post for those that havenât seen it already.
This I agree with. I donât understand how you ignore individual performance in a ranking system. I mean, how do you justify deranking a player when he goes 50-0 in a game? I can understand losing points, but he should lose as many as his teammate that went 0-50 against the same opponents?
They argue that they suck at figuring out what metrics help a team win, but you can at least get a ROUGH idea, and make SOME GUESSES. Like, the guy with 20k damage and 40 elims is PROBABLY playing better than the guy with 2k damage and 5 elims, you know what I mean? I guess you can argue having an off day, or just a bad matchup. Or like maybe the mercy is pocketing one DPS or something, but still.
Thatâs not what they look at though. Their matchmaker looks at the game and says, âI think this team will win 51% of the time,â and that team WINS 51% OF THE TIME. If it is actually that good at predicting that, then the matches are actually pretty closely balanced.
There is a direct correlation between having a short attention span and blaming the matchmaker. We need to figure out how to explain matchmaking through a series of 15-second TikToks with lots of flashing lights and one-word-at-a-time text captions.
Uh⌠when the matches are lopsided stomps⌠not so much.
Well, they said 50% of their matches fall within the 51-49% threshold, in competitive narrow matchmaking gameplay.
My guess is that there is a reason that they didnât speak about QP at all.
I would really like to know how wild QP matches typically are
The Matchmaking is so good, that at Masters, I gotta spectate a bunch of morons charging 1v5 like bots non stop into the enemy team; dying 2 seconds after jumping from highground, when using ults
While playing Duos, and doing placements on main accounts, my GM/Masters buddy in all roles canât get out from Plat 4? While Iâm at Diamond 2? We both have 70% winrate
But anyways, having multiple accounts at Masters, I decide play on an oldie; just to see what is happening.
With 2 games, two nightmares, where I gotta call everything that is happening on VC, Iâm already in Plat. Which was faster than playing in group? Ok, letâs keep going.
That was it.
30% winrate! And for the next day, as for a Letâs see what happens after placements, I just Alt F4 a game in Gibraltar with a 1st round ending in 2min and 34s.
⌠more roflstomps I guess xD
Its time to wak up
OW1 GMs canât boost Silver accounts to Plat
and then I see GMs and top500s playing like Diamonds; at best.
And no one argues the ability from Real top tier players; but having the opinion from that 0.1% ranked players, on what is happening bellow, its hillarious
Overwatch competitive is designed to keep people playing. But because everyone knows its trash, the algorithm needs to inflate and manipulate matches, so actual Top500s donât have to wait 78mins in Queue, for a game that ends in 3
I find better match quality in QuickPlay
Just in Duos, Masters start filling the lobby.
3 Stacks, and top500s start popping from under a rock
but if you ask me⌠they are not top500 xD
Anyways, have fun with the next MMR reset. The following UI change displaying those who play comp; and the patch after that, saying, once again, that they are fixing and fine tuning the matchmaking. Like with every patch for the last 8 years
this formatting is worse than WW1 chlorine gas attacks
They do mention QP is more loose but they donât tell details.
lol
No, is not; but not because QP is better
It is just the same experience
But I donât have to wait 9 mins for a game that ends in 2
Got a tl;dr? That is a seriously long video.