Board Game Arena

There might be some improvement to make in the way ELO score evolves through out games.
For one thing, the actual difference between the winner's score and the other does not seem to matter...
And the most unfair of all: when there is a tie, and that the winner has to be determined on a secondary criteria: in that specific scenario, there should only be a gain of ELO points for the winner, and zero change for the "tie looser".

Hello

You can check how ELO is computed here:
https://boardgamearena.com/#!faq?anchor=faq_account_elo

As you can see, the difference between the ELO does matter.

If there is a tie breaker in the game rules to determine the winner, there is no reason not to consider the winner as a full winner tor ELO

I believe the score difference mentioned means the actual in-game scores achieved by players, presumably with the idea that winning by a huge margin should be worth "more" than if you manage to barely eke out a win.

Personally I don't agree with that line of thinking, as it punishes risk-taking and further encourages playing for safer, more average scores to minimize potential loss. Which would not lead to an enjoyable experience on tables involving players who care about their ELO ranks, especially so with the unclear nature of how a match with more than 2 players would distribute the score difference values.

Edit: On another note, for games such as Innovation the proposed idea would actually change how you perceive a victory obtained through the special game-end conditions on certain dogma actions. Since these are encoded as setting the end-game score to 1–0 and "regular" achievement-based wins can range from 6–0 to 6–5 (in rare cases, 7–x and 8–x might be possible) relying on this measure means in a game that's not as close, you might be incentivized to actually wait around needlessly in the hope of clearing out 6–4 or even better, instead of going for your instant win that's "only" worth one score-difference units. Not my cup of tea :-)

I think I see this get brought up everywhere I go. I also never see it implemented.

While I understand that winning by a lot of points seems like a "better win," in the long run it should even itself out. That is, in one game you win by a lot and in another you barely win, or get a tie-breaker win (or loss). (Which is why I "got over" this idea a long time ago.) And for many games, the scores themselves are not that relevant.

If you win by a lot of points, is that because you're that great at the game, or because you're playing someone who is lousy at it, not only because they cannot get very many points, but maybe also not good at blocking you. So it seems to me that the other player's experience/ability level would be more relevant.