It's one thing to say the MM sucks (which it does), but to say that it's deliberately setting people up to lose several matches in a row is something else entirely.
The patterning I'm seeing is beginning to look more like a bug than it being rigged. It may be rigged, I'd still consider that an option for player distributions through lobbies, but it strikes me more as a calculation error, specifically if you end up in a particular pool of playstyle/performance where it's more likely to happen.
Therein likes the problem though. It's very easy for bias to impact how a person feels they've played, so their own self assessment may not be reliable. I can tell you that I feel I did exceptionally well, but someone may watch me play and think I did horribly. I would be more apt to believe the observer than myself because they aren't subject to the same level of bias regarding my performance that I was.
It's not too difficult to assess performance. Your k/d gives you an idea of how you handled combat relative to other matches, and your own performance in pushing the objective and/or defending the objective is obvious to you. If you pushed the objective the farthest of your team, or were the only one able to stop the incoming objective, those are very good indicators. I get much more frustrated when I know I'm playing badly than when I'm losing. I know when I do well or bad. And with support/tactical weapons rather than aggro shooters, it's much easier to have a handle on that. With aggro weapons, the adrenaline frenzy can make it difficult, I agree. I used to think I did so well with splash-o-matic, but my numbers and objectives showed otherwise. The adrenaline made it feel that way though. With eliters, chargers, hydras, blasters where every action is deliberate and methodical you get a more even feel for what you're doing and when/why you're doing it. I know very much when I'm doing well or poorly. it's also harder to get frustrated that way.
. Similarly, putting a cap on win streaks helps even more,
This, seriously, circles the heart of the problem itself. Any game where you have to prevent yourself from WINNING in order to avoid being made to LOSE, is inherently fundamentally broken. To actually have to be stressed and stop playing, because you're doing well and the matchmaker is actually working is awful. having to consider THROWING matches in order to WIN more to game the system is the antithesis of what ranks are supposed to represent.
This statement right here is the crux of why I say the only solution is to come to terms with the nature of the game and play to enjoy it, ignoring rank. If you don't it's going to be awful.
What you describe, where you only get a series of losing streaks with no winning streaks is contradictory to how Elo functions. In statistics, this is called an outlier. There is a possibility of it happening, but it is far from the norm. An overwhelming preponderance of evidence established that, if a person goes into a win streak, they can expect a proportional losing streak to follow, then a likewise proportional win streak after that, and so on. By limiting the extent of these streaks (both winning and losing), you stabilize this trend and make these win and loss streaks less drastic. This is not a new phenomenon nor is it unique to Splatoon.
That's exactly my point, the series of losing streaks with no winning streaks should indeed be an outlier. And yet, this frequently occurs in Splatoon, at least for certain players. The wins come slow and steady with an alternation of losses, exactly as you'd expect to climb ranks. When the losing streak kicks in it's absolute. And although you'd point to the psychology of it, once you've not played it with that sense of frustration because you already accepted it will happen, that's not an influencer. The fact that what statistics says should happen is not happening is precisely what I've been saying. There's something else that happens.
What you describe after that, though, also goes back to the fact that B+ through S is really all one giant rank and most of its players are caught up in one giant win/lose streak somewhere between. But I'd also point out the streaks do not appear to be proportional. The upward climb is always more difficult (and actually more frustrating) than the downward descent.
The advice of limiting the streaks by making sure to NOT win too much is horrible both in that it suggests not playing the game you want to play when you want to play it, as well as alluding to the possibility of intentional losses to game the system. That's horrible on its own. Catastraophic in a team game.
Splatoon is a somewhat unfortunate case of this, as Splatfests pretty much encourage binge playing to maximize snail pay-outs. In other games that use Elo, the incentive to maintain a stable player value that gives more consistently fair matchups outweights the incentives behind marathon play. With Splatfests, people are more or less driven to keep playing in spite of win or loss streaks to hit King level.
Binge playing is how I play all games. :)
Absolutely. I've never dismissed that as a possibility. I don't think it happens regularly enough to explain someone dropping three full ranks, but I do believe it is a factor in short-term losing streaks and general losse
eople often blame teammates or unfair match-ups for their losses, and I get that they are blowing off steam.
SZ yesterday, Carbon roller, my standby main on Mahi. I got a full team wipe next to the zone. Traded on #4. My team did not reclaim the zone. Its inexplicable. That's not "blaming my team" to blow off steam. It's a stated fact that they're simply bad at playing Splatoon.
Teams like that are not at all unusual. It happens a lot more than you think. It's not that they're bad shots, they're just inattentive entirely. That's the dreaded A range players I spoke of. Once I fell to B+ again games started getting a lot more competent. You wouldn't think so....but seems normal.
This is because of the very stupid way Elo handles promotions in games where you can win several times in rapid succession. Elo doesn't assess skill nor does it determine the level you can perform an consistently. It literally calculates probabilities of winning. If you win several times in a short window (like a winning streak), it tricks Elo into thinking you have a higher chance of winning against tougher opponents than you can realistically defeat. What happens is that win streaks slingshot people into a pool of players that are actually well beyond what they're capable of beating. That's when the losing streak begins.
You then keep losing until you fall back to the level you should have been at the entire time. Walking away from the game will not reset this: if you are out of your league, you're out of your league. What it WILL do is prevent you from getting stressed out so much that you aren't even capable of playing at your usual level, thus causing you to fall into a group of players that are much worse players than you are. If that happens, you'll go into another win streak because you'll get matched up with a bunch of players that are no challenge at all. This will slingshot you back up to the level where you're outmatched, thus the whole process starts again. By playing with a cool head, you're more likely to maintain a consistent level of play, thus your losing streak will even things out sooner. So, you can either stop after so many losses or just ride it out. If you ride it out, at least you'll get a feel for when losing streaks are coming and can prepare for them I suppose.
This is all very true, and all very stupid, BUT it ignores a few aspects of Splatoon. It's a team game. and being assigned these slingshotted players, in a game that is likely to create LARGE numbers of them, will carry you with their trajectory. If you happen to have a stat that makes it more likely to be assigned these players, you will often end in that trajectory. Not everyone, in fact most people, experiencing this problem seems to have been slingshotted past their own skill. Additionally, the slingshot effect should naturally not be nearly as wide as it appears to be. Additionally the point system that has different point rewards by rank and the fact that it does NOT generate matches based on rank, but wildly above or below, and then assigns very different penalties/points based on that matchup dump gasoline on the effect. Carrying players in trajectories they didn't earn not by winning or losing more but by winning the "right" matches or losing the "wrong" ones. I know you're aware of all this and agree, but you have to factor in the internal muddling that amplifies whatever elo does to something bigger than just "false win estimates"
Additionally, being in the "pool of players" say, S rank, that outrank you, should mean once you hit them, you start your losing streak. But a player that gets to S 60, then down to S 20 then up to S65, then down to S 22, then up to S 55 then has a losing streak into the A's doesn't sound "in over their heads" unless the matchmaker ISN'T matchmaking based on ranks but only on elo pools. In which case we're backed to the "the rank letter is in fact meaningless, ignore it and play."
This isn't a commercial activity, it's playing a video game. Any conversation that results in "don't win so much" or "don't play as much" points to a deeper problem. It's not blackjack.
Some of what you said made me think of something else though....a potential part of the problem....and maybe something other than elo entirely.
SPLATNET
Splatnet shows you what they call your "rank" both for TW and for ranked. It's a simplistic calculation of win rate * rounds won. It resets weekly. What if THIS is what they're using to seed the matchmaker? That would imply that players who play the most rounds, would inherently have a higher score so long as they're fairly good players that win fairly consistently. Not by actually having a higher win rate, but by having a good win rate and playing many many rounds compared to other players at the same point in the week? That would also mean "playing less" would be beneficial for matchmaking. It would mean more frequent players would be matched against harder players. It would also imply that you'd have a weekly curve of difficulty. And could account for fueling horrible win/loss streak swings.
If the matchmaker then mostly disregards rank or uses it only as a modifier against this score, that would account for horribly, horrendous matchmaking...and would account for different players with different play habits being affected differently by it.
EDIT: Just checked my Splatnet ranked "rank" and my score is quite a bit higher than several people currently in S rank on my list. I assume squads dont contribute to that score at all, as jsilva's score is low enough to suggest only a few rounds were played. Mine is compratively exceptionally high to all players on my list despite losing about 4 rounds for every one win - but I've played a lot of rounds so far. Which if it uses this for matchmaking also suggests this score isn't used and could be why squads get better teams.
