So if it is true that these matchmaker are algorithmically handcapped to give the players something that’s close to 50/50 or 40/40/20 win/loss rates. And if we were to assume that it is thanks to this sytem that enables heroes like mercy which arguably has one of the least solo carry potential in the game to climb the ladder. Then wouldn’t that lead to flex players in their respective role having easier/harder time climbing?
Say for example if you’re a mercy one trick, If these matchmaker is giving mercy mains a tailor-made game that’s allows her to have somewhere to 50/50 or 40/40/20 win/loss rates. Then I’m going to assume that mercy one trick ponies are going to enounter less type of matches where the only win condition for them is picking main-healers like ana/bap/moira to pump out sufficient healing.
And I’m guessing this kind of tailored matchmaking only starts taking effect after the support player has played at least 100 or more of competitive games and the matchmaker developed enough confidence in being able to tell which hero the support player picks most.
If this were to be the case, then wouldn’t it mean that if you’re, lets say ana OTP and one of your co-support insta-locks ana. Does that lead to your chance of winning/losing that particular game is no longer accuraely calculated by the matchmaker and possibly higher or lower than it needs to be? This also raises the question of whether the matchmaker tailor-makes matches for flex players as well in their respective role. Thoughts?
No more than it would if a DPS or Tank player swapped to a less familiar hero as well.
That situation is correctly resolved in an increased likelihood of losing more SR points. The matchmaker cannot change based on conditions it knows nothing about, like offroling, hero pool, if you’re playing left handed, if you’re playing without your monitor turned on…and it shouldn’t be expected to.
This leads to the argument that you the player are inconsistent, so you can’t blame the match maker it’s out of scope.
My only counter is that this whole game is inconsistent. There’s more inconsistencies than just the player, to further the point that the match maker is so out of scope perhaps it was obsolete.
“But gms can still climb so can others”. And then this goes to carry potential. And some characters have more carry potential than others so I agree with the original poster using the mercy reference.
So then here comes another question. If these matchmaker are tailor-making 50% winnable games based on your most picked hero. What if I my most picked hero is ana. But at the same time, I have this issue with my co-supports always insta-locking ana in my elo?
Unless the matchmaker is so manipulative to the level where it even finds a co-support player who will less likely be insta-locking ana, this issue is always going to be there. And I doubt the matchmaker will go that deep as that will likely increase the queue time.
So if I want to win more games that’s designed to be won on ana but I don’t have a reliable access to picking ana in my elo. I wonder if it’s better for me to stop using ana all together and just start maining other, less frequently picked support heroes in ranked?
Exactly and you have 11 other players in the game who could be inconsistent. Additionally you have smurfs and boosted accounts that could be on both teams. Results in a very messy matchmaking.
People are trying WAY too hard with this matchmaker.
I come at it from the background of tournament chess. There, they also use SR, and in some bigger tournaments where ranks are involved, they pair people in their rank, where theoretically you should have a fairly equal chance of winning. Instead, you always get some people who are a little bit sharper or more rested or more prepared or less tilted, who play steadily and get positive results, and you get people who play tired or aren’t prepared or having a bad day or distracted and they don’t. Sometimes you drop a game with a dumb mistake, sometimes you swindle a win where you shouldn’t. You don’t always play at your rating; it’s not an absolute predictor.
Now in Overwatch, multiply that uncertainty by 12, and then add in that it’s much easier to throw or smurf in OW, and you see why you’re never going to get a perfect matchmaker.
The matchmaker is also trying way too hard with the people. It doesn’t need to be complicated. Just ship 12 people within a close SR range into random teams. Instead they try too much to make it a 50/50 which obviously fails for all of the reasons you stated in your post.
The whole thing is based on an individual skill score, made poorly to fit into a team game where the entire team is a skill score (team average.) It’s flawed at the very foundation. Comp was probably originally intended for regular full teams only.
As someone who was hard stuck Gold and Platinum for many seasons (eventually made it to Diamond and Masters), my experience is that flexing won me games in solo-queue.
Another thing to consider is, Performance Based Skill Rating PBSR, the additional (or reduced) SR you get for doing well (or poorly) in a game, is also tied to your hero.
So if you’re putting out 7.5-9k heals/10 mins on Ana, you’re probably putting out average heals. If you’re putting out that much healing but also getting huge purples and finishing low HP targets, you’re going to get more SR.
Same goes for Zen, Lucio, Brig, and Mercy. Your level of pro-activity in your game is what allows you to gain those additional SR points. And if you’re not performing very well, the game will take points off (kinda like a math test where you showed your work and messed up your arithmetic).
You need to be always doing the extra credit assignment. That’s how you ensure you get the PBSR that shoves you up the ladder.
Aside from the whole debate about player skill, this is going in the right direction. The game itself is just unpredictable as hell. There are SO MANY things one person or the whole team can unconsciously screw up, that even an advanced AI can’t 100% predict the outcome of a match.
So don’t say it’s the matchmaker holding you back. In Diamond and below, it’s more likely that your decision-making ability (and consistency) just sucks in certain situations.
And one-tricking only works until you run into real counterpicks and people who actively want to take you out.
Its kinda like u agreed but then failed to see the big picture again.
match maker is something that doesnt help, if its not WITH you, its against you. So stacking is urged. Can I blame my entire career on match making… no… but its not helping.
Other players. We all basically agreed one person and or multiple people can unconsciously screw up - so again other folks might hold you back. Whole career? no, but we atleast acknowledging these factors exist.
So 1 issue cant be blamed, but both issues and a mix of yourself to improve. Some nights I can definitely see 1 - 30%, 2 - 50% - 3(self) - 20% . Other nights depending on the “time of day” and “region” - it could be 10/10/80.
At this point we also have to acknowledge that other people who play at different times, and different regions have different mixes.
This is why I never denied self perfection and seeking knowledge.
SBMM matchmaking isn’t supposed to be looking directly at what heroes you play. It looks only at the data/performance and adapts (rigs) around that. In that sense, a OTP might be too restrictive and overspecialized for achieving higher MMR, whilst over-flex would be too budget washed and also fail to max the MMR.
The best min/max (imo) would be to approx 4-pool yourself. Basically 3-4 trick a small pool of heroes 1-2 of them are very universal “carry solve” for bulk of the ladder and meta-agnostic, like raw frag raw dmg (junkrat or reaper, say) and the other 1-2 are more niche and win based on surprisal/mastery. Same idea for tank/support roles you might opt to master 1 mainrole and 1 offrole or one shield and one dive tank or one solo/pocket healer and one group/brawl healer. Best hedging between flex and rigged matchmaking.
Note however there is likely EOMM and “desirability” scores that go into the MM. We don’t know for sure (but the patents are in-use and show Overwatch). That would mean the MM adapts down to what heroes you’re using and tries to rig you into lobbies that put you against mirrors and other otp specialists to “test your diff” against them for faster mmr adjustments. Plus you get all the engagement effects of pushing you in/out of the win/loss cycle for max dopamine and maybe buy an OWL skin for ur main.