Drives season is so freakin’ bad, my entire team got completely held on Havana after the enemy team pushed all the way through without being stopped, the Zenyatta didn’t even die once, the Mercy only died once.
Their soldier was in no way plat (I got demoted to plat lobbies) because he shot a rocket UNDERNEATH the payload to kill me with splash damage, there’s no way a plat player could think like that.
Whoever made the matchmaker needs to be fired, it’s genuinely criminal just how crap it is.
I don’t complain about the matchmaker ever, but this game, that I cannot describe on the forums what it felt like, was the worst thing I’ve ever played and two people need to be die for it, one of them being myself.
Its broken, so yes. 10 years with this crap and still got even worse matchmaking than ever before. Body Cam game that is early access on Steam has better matchmaking than this game.
Ngl i find it absolutely hilarious people think matchmaking is “broken”
Especially when were talking about blizzard, the developer who has a literal patent for a “pay to win” matchmaking system on file with the u.s. patent office, public record btw you csn find it on google docs.
The system isnt broken, its just a marketing tool masquerading as a matchmaking system..
Blizzard isnt going to do crap, because, “fixing the system” also eliminates their ability to use it as a marketing tool.
Furthermore at this point “fixing” it would do more harm than good for the population of the playerbase..
People coming to forums to be like “I don’t win 80% of my matches in a video game and so 50+ people should lose their livelihood” is just such an insane take though
Well when you consider some of the takes on here it’s not surprising that theres many people on here with just enough brain cells to take those stances as well
The matchmaker isn’t bad – it’s just based off of a 60+ old matchmaking system used for games such as chess, checkers and the like. There are better match prediction systems now (that’s why fantasy sports exist for the NFL, NBA, MLB, Premiere league, etc.etc., essentially all popular team sports), but apparently Blizzard is too stingy to spend for it. They hired some guy who did OW heuristics who tried to develop fantasy OWL (Winstons Lab dude), but they probably didn’t use him for OW competitive matchmaking because it still sucks.
Most of their articles are long on vague math, and short on what the rating means. What is increasingly becoming clear is that it’s just the usual statistical match analysis based on wins and losses.
Most prediction systems have moved beyond that and address the more interesting issue of why there are wins and losses in team sports – it’s essentially moving beyond the visible surface (wins and losses), looking underneath to find root causes (called underlying stats/numbers, shots taken/eAPM, accuracy, first picks, etc.etc.).
I find it weird they don’t do this because the OW HUD tracks these data, or can be extrapolated from OW HUD data. They can get any quality control or financial analyst to look into this and do some match quality control analysis and conclusions – they hired the guy doing that in Winstons Lab, but I’m assuming it’s only that one guy based on lack of apparent progress. Most importantly, it’s being done in lots established sports already.
TLDR – OW has the data to do this, the area of expertise for advanced match prediction is well established, why Blizzard doesn’t want to do this is probably because it’s way below their priority list (the ones on top probably being more skins for the popular ladies in the OW hero lineup).
The MMR used to be much more complex, since Role Queue and OW2, it’s been simplified.
Blizzard apparently spotted an issue with the Support role where the stat tracking, wasn’t working correctly and so a lot of Supports were being placed wrong.
The roles were never balanced properly between each other. So the stat tracking they had was removed for a straight win/loss system of SR and ELO standard. It doesn’t seem like they know how they can address this issue, but it sounds like they might have something next year to add.
Yeap, much of the what they have published are steps backwards to simplify what is at heart an ELO (i.e. wins-losses) system that’s really old.
I don’t think it’s as simple as roles being balanced properly – it’s just that metrics on a hero-per-hero basis differs because Blizzard made the creative decision to make the kits of each hero unique (a feature that elevates OW way above its competitors).
The biggest difference is in the support class because they have the most varied utility abilities, and effective use of certain utilities (speed boost and nade is the most prominent I can think of) affect how well these heroes are used. Speed boost is especially troublesome because only two support heroes (correct me if I’m wrong) have them. It results in possibly apples-to-oranges comparisons which distorts the effectiveness of some metrics.
These are not intractable problems, because these problems exist in a lot of sports – sports analysis has become good at identifying outliers, the data characteristics of players whose value exceeds that which is perceived by conventional valuations.
Lastly, there are heuristics we apply to OW which can be converted from the subjective to the objective – the space and tempo controlled by tanks, the kill pressure from dps. It’s harder to do that for supports, though, because their utility usage and effective decision-making is tough to quantify.
I’ve had several games during drives that were like this (whether I was on the winning or losing team). I’ve never seen it like this before. Most games seem dictated by whichever team gets the better damage heroes–one side’s dps will be 30:2 and the other side will have a negative K:D. I had a Genji today with 23 deaths by the end and 11 elims LOL
At the time OW2 was released, the Support-role consisted of a very generic template of two healing abilities; a primary and secondary heal, a utility, and defensive ability that was either an ultimate or a cooldown. You can count on the number of Supports that increased damage through an ability, cooldown, or ultimate.
Heck, most of the post-OW2 Supports released still adhere heavily to that Support template. If you wanted to see variety in playstyles, abilities and weapons, you’d look to the Tank-role. The Tank-role was the only one that had melee primaries, beam primaries, ranged primaries, shotgun primaries, and explosive primaries. They were the only ones that had area-effect crowd control, clearing debuffs off other heroes, and better resource management than the Support-role. The whole point of OW2’s hero release order was the bulk up the Support-roster choices; and they did that really well! But it still adheres strongly to the Support template made after OW1’s beta.
Even with Doc’s release this Tuesday, doesn’t deviate from that template. Two healing abilities, a primary fire, and a secondary via his E ability is still there, He has additional healing (and damage) through his Infusion mechanic, and his ultimate offers some defensive utility granting Overhealth to his team caught under the effect of his ultimate. And like many Supports, he also boosts damage via his E ability.
That said, Supports have become more creative in recent releases and a bit more varied. And revisions of existing Supports (notably Mercy) have also made Supports more… present… when they’re played and utilized well beyond healing others. But most still lack the variety of the Tank-role, but I’d say they’re catching up to the Damage-role in play-style variety, though not really in specialization (and I doubt they ever will).
Basically, I think this is why they dumped the more stat-based MMR ratings for a win/loss one, because Supports were pretty limited at the time and quantifying that was pretty difficult to measure. Moreover, a lot of that Support-level utility could be bound to Tanks as the biggest benefactors, overall, compared to the Damage-role.
Agree to disagree. While it’s true that supports heal and damage in varying degrees, it’s utility that makes the biggest difference in terms of impact in a game.
Note how JPC is a menace because of that utility “lifeline” is impactful. Even during OW1, the main support was Lucio because of speed boost, while the flex spot varied depending upon whether certain buffs/debuffs/ults were important.
It feels generic in low rank games, but not in high rank games where balance is most important.
Again, agree to disagree. I’m not going to bother with counter-examples because I sorta explained in earlier.
I hope so. The card system hints that they’re looking at metrics specific to heroes – for instance enemies ignited for Anran, negative effects cleansed by Kiriko, etc. That’s why for quite some time I always thought these became factors in determining SR/MMR and matchmaking (until @Baja showed me otherwise).
