Proposal: Reduce All FFA Standard Ratings by 700

Sort:
HSCCCB

Ideally, chess.com two player ratings, FFA, and Teams ratings should be similar, where a 2000 blitz player, I would say, has gained the same amount of skill as a 2000 FFA player or a 2000 Teams player. 

Functionally, this is impossible to line up completely, but we can still try to move ffa ratings to a point where FFA ratings seem to at least vaguely parallel two player ratings. 

Right now, FFA ratings start at 1500, while normal ratings (typically) start at 800 or 600 or something like that. From a logical perspective, I don't think this discrepancy is needed, as it inflates all ffa ratings far above where they should be. I propose reducing all ratings by 700 points and starting at 800, which I think is much closer to similar skill in normal chess.

For example, a 1500 would be an 800, 2000 a 1300, 2400, a 1700, and a 3000 a 2300. While not perfect, I think this makes a lot more sense. 

Similarly, I believe FFA ratings are above Teams ratings by about 500 points, give or take. I think it should be considered to make their ratings line up more as well.

The argument against this is that the current rating system looks better. It is cooler to be a 2000 than a 1300, yes. But I think under the current rating system, players either ignore the current rating tiers versus 2 player chess, understanding they are completely incomparable, or are misled by their rating, neither of which are desirable. While reducing ratings by 700 may make ratings less attractive, I think having a rating system that reflects something that is actually somewhat comparable with regular chess ratings is more desirable in the long run.

Thanks for considering,

Caleb

JkCheeseChess

then we can never see radon getting 3300 or rojitto getting 3400!!!

JkCheeseChess

on a serious note, we've actually considered this before. I don't remember what the consensus was but our plan was to use a certain formula that doesn't necessarily change all ratings by 700, rather changes by smaller and smaller amounts the smaller your rating is and by larger amounts the larger your rating is

for example if you were exactly 1500 you wouldn't have your rating changed, if you were 1600 it might become 1575, if you were 2000 it might become 1900, if you were 2500 it might become 2250, etc

thenomalnoob

Hmm... elo exists to compare w/ other players, so no matters to change

HSCCCB

@CheeseDuck

Just a straight line reduction wouldn't affect rating records nearly as much, which would still make rating goals like the ones you mention attainable. Instead of 3300, radon would be going for 2600. Again, just a small loss.

The rating alteration you mention may or may not be better; that would, however, mess up rating things like that. Similarly, I still don't like starting at 1500.

HSCCCB
thenomalnoob wrote:

Hmm... elo exists to compare w/ other players, so no matters to change

Yea, this post is dealing less with "in-game" elo, and more with "inter-game" elo

ChessMasterGS

I’d wager that unless we also reduce the Glicko RD default, rapid might return to “normal” in ~1-2 years and Hyperbullet maybe months instead of years

HSCCCB

Oh probably, it is slightly a different issue, though I guess it will invalidate this one.

I assume that another rating system change is off the table?

Shrinking ratings (cheese) and reducing the RD (gs) would set it back, but even, then, eventually, wouldn't it go back? eta: of course, maybe that is what it will have to be

JkCheeseChess

It wouldn't necessarily go back. I mean, mayyyyyybe it might happen in another 5 years. But even then you still have 2pc players who are 3300 in blitz, so it shouldn't be that big of a deal because the people that do achieve 3100+ will be the select few that deserve it and not every single person on the leaderboard.

Otherwise we should have a good distribution of players and their ratings.

Arjun1516

This is a great change. I feel like I am super good at 4 player chess but I just ain't. Really breaks my heart when I realize that.

HSCCCB
TheCheeseDuck wrote:

It wouldn't necessarily go back. I mean, mayyyyyybe it might happen in another 5 years. But even then you still have 2pc players who are 3300 in blitz, so it shouldn't be that big of a deal because the people that do achieve 3100+ will be the select few that deserve it and not every single person on the leaderboard.

Otherwise we should have a good distribution of players and their ratings.

Not sold exactly, but I will say I am not sure what the alternatives will be. We can't change the rating system, but I am positive that ratings will keep going up. So what does that leave besides this? Anyway

jujocu

Yo apoyo también un ELO más realista, aunque deba tener 1500.wink.png

Arseny_Vasily

BeautifulGoose

I don't want to lose 700 pts

jeremyhu

same

flyfishsullivan

i see what your trying to do, i think dropping everyone by 700 is a good idea but to prevent rating inflation you should only be aloud to join/make rated games 200 below your rating. You see alot of leaderboard people (3000+) play a 2250+ game thats whats causing this

ChessMasterGS

Wouldn’t they otherwise never be able to get into a game?

MuppetRobin

4 player chess has a really small player base. Those 20 players are not that active, it's a ghost town anyway. So, players being able to farm up to 3000, have a huge gap (big rating difference) between 100-60 and top 60. If we would somehow bring players' ratings closer where the leaderboard would vary more often and we would have more competition. Also, where are 2800+ games? Everyone is focused on farming and their rating is far above the other players. In regular chess top 20 play against each other all the time and they boost their rating. 2600+ or 2650+ should be the highest rating range queue.

JkCheeseChess

The solution is obviously to reset inactive radons rating to 2200

Arjun1516

Perfect. JK