ELO calculation change

[Lucan 2]

I feel ELO calculation should be changed. I just lost more ELO than opponent when he had more than me. How does that work? Should've lost like 4 or 5, wasn't very big gap between us

[Jellby]

Winning/losing against an opponent with the same elo should be +/- 10
Then that's modulated by the elasticity (K) factor (more for your first ~20 games), consecutive wins/losses, number of players...
And of course, you win more by beating a higher ranked player, and lose more by losing against a lower ranked player.

[dschingis27]

I suppose you refer to this game: https://boardgamearena.com/table?table=770676875

It's just a rounding/display error. Your opponent won exactly the amount of Elo points that you lost, you can see it when you hover the cursor over the Elo scores (not sure how it works on mobile). You got -9.62 pts and your opponent got +9.62 pts. Apparently it's then rounding both values down for the quick display, and then it looks like you had different values, but you had not. (In the backend, your Elo is saved with at least two decimal places even though on most pages on BGA, it is only shown as integer.)

In 2p games, the Elo won and lost must be equal (unless you are a beginner of a game).

[BarnardsStar]

hover the cursor over the Elo scores (not sure how it works on mobile).

For the most part, “hover” text/graphics are triggered on mobile with double-tap where the second tap is held for a half-second or so… kind of like an A in Morse Code. The site can be a little finicky about it; it doesn’t always work and it’s nearly impossible to trigger the same one twice in a row (but you can trigger another one and then go back to the first again), but that is usually how it works. Usually.

[FrankJones]

I have no idea how people play on a touch-screen device. Besides the tiny screen, there are also issues with "hover" displays that I sometimes cannot figure out how to do on my touch-screen devices.

[Gfunk7676]

My only issue with ELO scoring has been in 3-4 player games. Due to the natural increase in luck factor, not to mention how players tend to cooperate against the leader in any given game, there should be extra credit for beating 2-3 opponents or a lesser loss in ELO for losing to 2-3 opponents. But multiple opponents doesn’t seem to figure into ELO scoring at all as far as I see.

[Meeplelowda]

But multiple opponents doesn’t seem to figure into ELO scoring at all as far as I see.

It says in the FAQ, linked to at the bottom of nearly every page in the site:

A player's ELO change is determined by comparing their finishing position against each of their opponents'. For example, finishing second means a loss against the player in first place but a win against the player in third.

[Jellby]

My only issue with ELO scoring has been in 3-4 player games. Due to the natural increase in luck factor, not to mention how players tend to cooperate against the leader in any given game, there should be extra credit for beating 2-3 opponents or a lesser loss in ELO for losing to 2-3 opponents.

OK, now imagine a group of 3 players with the same skill who play 3-player games. No matter how difficult it is, there will always be a winner in every game (let's ignore the possibility of ties), and since all players have the same skill, in the long run each of them will have 1/3 victories against 2 opponents and 1/3 defeats against 3 opponents. With your suggestion, they will win more with their victories than lose with their defeats, and therefore the total elo in the group will increase over time. That's not something desirable.

Or do you suggest that also winning against 1 player and losing against 1 player should have the elo adjusted, in order to make the total zero sum? That brings other problems. The first is that it's not true for all games that it's more difficult to win against 2 players together than separately (e.g. Yahtzee).

[FrankJones]

But multiple opponents doesn’t seem to figure into ELO scoring at all as far as I see.

It says in the FAQ, linked to at the bottom of nearly every page in the site:

A player's ELO change is determined by comparing their finishing position against each of their opponents'. For example, finishing second means a loss against the player in first place but a win against the player in third.

I can't recall exactly how the other somewhat recent discussion about this went, but -

Didn't we figure out that the Elo change is modified with some unknown formula in such situations?

For example, when 3 equal Elo players play, the winner gets credited with 2 wins (explaining how some people can have a win% higher than 100%), but is not warded the Elo that would be expected from beating both players individually.

In fact, I recall two instances in which the winner of a 3-player game received exactly as much Elo as though it had been just a win in a 2-player game.

[Meeplelowda]

But multiple opponents doesn’t seem to figure into ELO scoring at all as far as I see.

It says in the FAQ, linked to at the bottom of nearly every page in the site:

A player's ELO change is determined by comparing their finishing position against each of their opponents'. For example, finishing second means a loss against the player in first place but a win against the player in third.

I can't recall exactly how the other somewhat recent discussion about this went, but -

Didn't we figure out that the Elo change is modified with some unknown formula in such situations?

For example, when 3 equal Elo players play, the winner gets credited with 2 wins (explaining how some people can have a win% higher than 100%), but is not warded the Elo that would be expected from beating both players individually.

In fact, I recall two instances in which the winner of a 3-player game received exactly as much Elo as though it had been just a win in a 2-player game.

It is modified (reduced), so that it is less than two wins in two-player games against the same two opponents. I don't know what it means for "the winner of a 3-player game received exactly as much Elo as though it had been just a win in a 2-player game." A two-player game against a player with what rating?