Quitters at very convenient moments (Kingmaking when the game is already over).

[aZesulfaMe K]

I have seen many different cases of kingmaking, but the latest one feels especially frustrating and unfair. The system should address the fact that, with 100% of the game completed in ranked matches, the lowest-scoring player can let the time run out to be kicked for exceeding the time limit, allowing another player to take their turn instead. As a result, the other two players end up with the same score if they have the same ELO and division, regardless of their final in-game points.

[Suoivax]

How do you know that the losing player did that specifically to boost the second place player?

[Jellby]

Also, the quitting player got a penalty.

[aZesulfaMe K]

Let's suppose he didn't do it intentionally. Why does the second-place player receive the same league and Elo points as I do? The game was already decided; only the forests from the last production needed to be placed. Why does the system punish me if I played my game without any incidents? This is a very important issue that needs to be corrected to maintain the integrity of the game. Despite managing the game better than the second-place player, the reward is exactly the same for both of us.

I would understand if the second player had a much lower Elo and was in a lower league than mine, but that wasn’t the case—we were evenly matched. So what’s the point of playing ranked games?

[Jellby]

Because BGA doesn't know when the game was already "decided". A game that did not complete fully is a game in which the players cannot be ranked. Everyone who did not leave is tied as a winner, and the quitter/kicked player is the loser.

[Suoivax]

Let's suppose he didn't do it intentionally. Why does the second-place player receive the same league and Elo points as I do? The game was already decided; only the forests from the last production needed to be placed. Why does the system punish me if I played my game without any incidents? This is a very important issue that needs to be corrected to maintain the integrity of the game. Despite managing the game better than the second-place player, the reward is exactly the same for both of us.

I would understand if the second player had a much lower Elo and was in a lower league than mine, but that wasn’t the case—we were evenly matched. So what’s the point of playing ranked games?

From a programming perspective, that would take a serious amount of work, teaching some sort of AI what a "decided game" actually means and looks like.

[FrankJones]

1) Now I'm confused. When my opponent quits in Terraforming Mars, my interpretation of the rules is that the ELO points I am given are based on the "percentage of the game completed", which would indicate that the programming does use some sort of formula to determine such things.

2) If A game of TFM is in the plant conversion stage, why does the game need to end when one of the three players quits? Just allow it to continue as a 2-player game, skipping the turns of the player who quit.

3) For that matter, why not just do that in any game that can still be functionally played with one or more absent players (such as Terraforming Mars).

[Acerbic]

% of game completed has nothing to do with who's winning or losing, it's a simple measure of how far some global parameters have progressed. In TM, I assume it's just a measure of the 3 terraforming levels.

It's not even remotely feasible to algorithmically determine who was on track to win the game. It's a gargantuan problem that depends on many factors including how the player who quit would have behaved if they hadn't left. Nor can you continue the game (except, of course, in the sense that you can already continue the game without further influencing the ELO outcome). It may be unsatisfying, but I don't see any better solution than what BGA already does: call it a win against the quitter and a tie against everyone still playing.

[Jellby]

It may be unsatisfying, but I don't see any better solution than what BGA already does: call it a win against the quitter and a tie against everyone still playing.

Well, I would suggest calling it a win against the quitter and nothing (no win, no loss, no tie, 0 elo transfer) against all the other players. But that should be a BGA-wide adjustment, not TM specific.

[Acerbic]

Sure, that's what I meant - the ties between the other players should not trigger any elo changes. The fact it does now is, I believe, a bug. It's been fixed several times yet somehow keeps returning.