In the current system, rating 1v1 games is borderline rating manipulation

Like really isn’t the obvious perspective here to look at where rating problems come from?

No one is going “ok time to manipulate my rating by crushing Swkoll’s invitational EZClap” they are saying “ok time to go host my all welcome game and beat everyone with a min 10 strat rush because they aren’t even tech 2 yet while I also abuse high uncertainty players and general low levels of mu to gain 15 points a game.” Or “time to lose 200 points a game by accidentally dying in theta or winter duel against players with 1k rating.”

Why would you fix the inherently problematic part of the rating implementation by removing the most properly working area in an attempt to legitimize the former?

No one is going “ok time to manipulate my rating by crushing Swkoll’s invitational EZClap”

Playing these tournaments is the only way I gain rating, I usually gain like 100 rating in these tournies and then slowly lose that rating in teamgames until the next tourney. Unfortunately, this time I accidentally unrated all the games I hosted, so I didn't get that boost 😩

@ftxcommando

This is the fulcrum of it. You have mistake “the point of global rating” with “what I use global rating for.”

At no point in time was global defined as teamgame rating. At no point in time was global defined as map gen rating. At no point in time was global defined as 5v5 rating. That’s the way you use it.

This is really dishonest. You know as well as everyone else that everyone uses global for teamgames (not just mapgen like me of course, but astro seton gap and whatever of course...). It's not just the way I use it, it's the way you use it too, as well as just about everyone else. So you start from this premise that "hey theoretically you could think of global as a representation of your 1v1 performance", but literally nobody does and you know it as well as anyone else. Ladder and tournys are used for 1v1 instead. And no matter how many tournys are rated, or how much u shout at gappers to play 1v1, that isn't going to change.

Your next point is something I've already responded to. 1v1 performance simply does not reflect teamgame performance. Playing lots of 5v5 mapgens will get you the best representation of your 5v5 mapgen skill. Playing lots of Setons will get you the best representation of your Setons skill. That's just how it is. This 1v1 argument is starting to get reminiscent of League of Legends players complaining about their teams being bad and being stuck in "Elo hell". "Ahh I'm such a good 1v1 player, but my teams suck they are just bringing me down every time!". Or the reverse "That guy is such a bad 1v1 player, but his team always carries him!". If you're playing opti, chances are the teams are fair when aggregated over a large number of games.

Btw you are clearly targeting me with these "hide a 2600 with 1600s" example etc. but it's pretty silly since I go out of my way to play the highest "integrity" games possible. Like 90% of my games are 1. random teams (opti) 2. random map (mapgen) 3. 1800+ (with very few exceptions), which is the highest rating range that is generally available. How could my rating possibly be more representative of my mapgen teamgame performance? Certainly would not be improved if I ranked the unrusting 1v1 games that I've played recently.

And while we are on the personal anecdotes, need I remind you that I've been on both sides of this? There have been times where my ladder rating has been higher than my global, and there have been times where I've gained outrageous amounts of points from 1v1 tournaments (and times where I've lost about 300 points in an FFA tournament for that matter). Yes I'm talking about this issue because it's relevant to me, but no, I'm not "just" some high rated teamgamer crying because I want to keep my rating high. I just want to keep my rating tied to my 300 mapgen games or w.e instead of completely randomizing it ocassionally with 1v1 tournys.

Your rationale justifies a position of giving players the freedom to unrank games if both players desire it. It does not meet the bar of enforcing it regardless of the whims of the other player.

I don’t use global for 1v1s because I dislike 1v1 gameplay. I use global for 2v2s. I still gain essentially double the points per game in 2v2 that I do in big teamgames. If I had to deal with a dude forcing me to unrank my games because he wants to keep his points, I’d just intentionally screw around and make RAS SACUs instead of playing the game. That includes a tourney. If you’re scared of losing points to me then your rating is inaccurate and mine is inaccurate; I consider it disrespectful to unrank in that context so I will disrespect you in kind with my gameplay.

And I know you also wanted to unrank 2v2s in the tourney I hosted, so evidently this problem of yours does extend to that game mode where I do in fact use global in a different way than you. And there is zero rational argument for why your way is better than my way. In fact the preponderance of evidence on how TrueSkill was intended to operate would favor my interpretation since I rarely play 2v2 games with extreme rating discrepancy. Just as tournies rarely have extreme rating discrepancy.

I’m not targeting you because I think your rating is fake. I’m using you as an example because it’s the easiest way to convey my point. Even people that try to play legit and within “the rules” of TrueSkill will still face problems at the highest tier. If you want to play 5v5s and you’re 2600, you simply must have games with like 600-800 rating discrepancy. It’s a problem of a low population game at the end of its bell curve. But my point is that this issue is not solved by eliminating a more robust game environment.

@lord_asmodeus said in In the current system, rating 1v1 games is borderline rating manipulation:

@katharsas
When hosting 1v1 Tournaments we do use ladder rating as the entrance barrier.
The issue is we have to host all the matches for these tournies as custom matches which means they affect global rating not ladder rating.

Maybe there would be a way to give 1v1 tournament organizers the ability to count those custom games as if they were ladder games. But i don't know how hard that would be to implement.

How are you supposed to do that without forcing TDs to host every single game themselves, which drastically slows down an already slow tournament process? The only other option is letting people host and use whatever rating they want to use in custom games, which then regresses to making every single rating as terrible and inaccurate as current global rating.

Brutus has a half-finished tournament software that was supposed to automate exactly that iirc