Players benefiting massively by quitting a game

[DrewZ]

Under the current rules, when a player is out of time to make a turn (particularly in Turn Based games), any other player can "Quit the game without penalty". Even the players who are in last place/losing badly at the time. That's fine I guess, you shouldn't have to wait forever for someone else.

But right now those players with no chance of winning the game actually can be awarded huge ELO points by ending the game. This just isn't right. Meanwhile, a player moments away from winning gets zero if their ELO is higher.

This does nothing but encourage people to quit a game and not let the player behind on time a chance to get caught up.

Obviously its frustrating for the player who is about to win, but if they're not going to get any points why should the person who decided to quit the game benefit? For example, look at Rallyman table #154160028, Progression at 99%. The red car (dead last) decides to end the game when another car has run out of time. They got a huge ELO payoff for quitting.

My suggestion: you take the situation right at that moment as if that is the final score, calculate ELO, then distribute it at percentage of progression. For example, if the game is exactly half over, and you would get 20 ELO for a win, you'd get 10. If someone would lose 20 ELO, they would lose 10.

If some players won't receive any benefit at all, neither should the person quitting on the game right before the end. Again, just a suggestion. The way it is now just rewards quitting.

[Silene]

My suggestion: you take the situation right at that moment as if that is the final score, calculate ELO, then distribute it at percentage of progression. For example, if the game is exactly half over, and you would get 20 ELO for a win, you'd get 10. If someone would lose 20 ELO, they would lose 10.

This would only suit a few games. On many games, the actual score does not show who's actually on the path to winning. You would need an individual solution for every game which is just not possible. For ELO-culculation i don't see a better solution than something that's similar to what's been done now (a few tweaks possible maybe)

If you think a player has skipped someone out of time too early (they were still active / made an excuse / other reasons) while you think they should've given them some more extra time and just didn't do it for their own benefit when they were losing - you can't do more than not play with them again, because they're well within their rights to kick the out-of-time-player .

[MoiMagnus]

What's the rule when someone is kicked out of a table?
I believe the ELO earned is proportional to the progress of the game, and you cannot lose points. IMO, it should be the same rule when someone quits, if that's not already the case.

[DrewZ]

This would only suit a few games. On many games, the actual score does not show who's actually on the path to winning. You would need an individual solution for every game which is just not possible. For ELO-culculation i don't see a better solution than something that's similar to what's been done now (a few tweaks possible maybe)

I won't speak for the entire site, but every game I play on here (about 12 or so) has a scoreboard of some kind. Its pretty clear who's ahead and who's dead last. My only point is it encourages quitting a game at the first opportunity if you're losing badly because now you're not going to lose ELO, and you can instead gain the same amount as if you'd won the whole game. Its just..... not good. At all.

And by the very structure of turn based games someone isn't going to burn thru their time bank until closer to the end if you're just a little slow each turn in playing. Every time this happens in my games the progression is 95% or higher, and its someone will no hope of escaping last or next-to-last who jumps the chance to quit the game. Then they sometimes get ELO (a LOT of it) from quitting. There's gotta be a better way, or why have ELO in the first place if its that easily manipulated....

[Meeplelowda]

you can instead gain the same amount as if you'd won the whole game.

Is this how it works, or do you get awarded points as if every non-timed out person tied for first?

[Romain672]

every game I play on here (about 12 or so) has a scoreboard of some kind. Its pretty clear who's ahead and who's dead last.

Let me categorize games I know where it's clear who won at 95% progression:
Very Clear: Not Alone (and still, you can find games where you goes full rover early on, or the opposite), Saboteur, Sushi go (for the most part), Color Pop (for end game, with a quick play by a computer), Marrakech
Clear maybe with some coding: Lost Cities, Can't stop, 7 Wonders (scientific), Tobago, Yahtzee, Kingdomino (I put it here, but still, someone has two group of corn separed by one tile, or someone goes full mine, and two mines are left in 8 last cards)
I dunno: Backgammon
Not clear: Dice forge (someone has full dices, others has full victory points), Carcassonne (someone has a big castle not finished), For Sale, Santorini

Which give 5 very clear, 6 clear, 1 dunno, 4 not clear.

Who, those numbers in very clear and clear are way higher than I though.
But don't forget each game on the 'clear' category would require some coding aspect to take into account weird position, which is pretty annoying. And if you remove them, most of the games became not clear.
And some of the very clear can be totally unfair if per example at Saboteur, you loss your two first round, but was going to win the third one. You would lose elo instead of being neutral.

My only point is it encourages quitting a game at the first opportunity if you're losing badly because now you're not going to lose ELO

If you are the one who leave, you normally lose elo.
If someone else leave, you don't lose elo.
I think.

[DrewZ]

If you are the one who leave, you normally lose elo.
If someone else leave, you don't lose elo.
I think.
[/quote]

Here's what my issue is, and apparently not a lot of others care given the voting:

-- Player A has exhausted their allotted time in a game

-- Player B, who is in last place with no chance of even moving up the standings (let alone win) jumps on the option to "Skip players out of time"

-- Player A, who was out of time but did not directly quit, takes a hit to their reputation but doesn't lose ELO

-- Player B, who was clearly or highly-likely to lose ELO in the game now GAINS as much ELO as if they'd WON the whole thing

-- Players C, D, E etc. who were winning or about to win, get nothing if their own ELO is really any higher at all to Player B

Here's a perfect example: https://boardgamearena.com/9/rallymangt?table=154160028

Symon was dead last in this game. lemalkavfou was out of time. Symon (when it wasn't even his turn) chooses to Skip Player out of Time. Now instead of very likely losing he gains +34 ELO. I mean.... really??

Like I've said, it just encourages people to quit games early, which kind of stinks. Nobody in contention to win is going to bail, just those who want to skip out and save themselves. Then they get rewarded for it. Its a bad setup.

[DrewZ]

Is this how it works, or do you get awarded points as if every non-timed out person tied for first?

It only takes one person to be out of time. As soon as anyone chooses to "Skip Players Out of Time" then the game situation is locked (even if other continues to play), everyone but the person out of time is considered the "Winner" and awarded ELO based on how much of the game was completed. But its ELO based on a multi-way tie I guess, because the highest ELO players won't get anything no matter what the game state was

[Romain672]

Oh you are saying it encourages players to expel player out of time, not to encourage them to leave.

That was the part I miss, and yeah I agree.

[Meeplelowda]

Is this how it works, or do you get awarded points as if every non-timed out person tied for first?

It only takes one person to be out of time. As soon as anyone chooses to "Skip Players Out of Time" then the game situation is locked (even if other continues to play), everyone but the person out of time is considered the "Winner" and awarded ELO based on how much of the game was completed. But its ELO based on a multi-way tie I guess, because the highest ELO players won't get anything no matter what the game state was

I took issue with the characterization that the player that was behind massively, but not timed out "gain[s] the same amount as if you'd won the *whole* game." Now that would be wrong. But as you confirmed, you only gain points based on a tie for first, not based on "winning the whole thing."

If the alternative you propose is to award ELO points based on where you would have finished if the timed out person is placed last, then everyone else otherwise given their position based on points, that makes sense for some games, but not for others. Also, the extent to which it makes sense depends on how far along the game is. There are some games where being massively behind is nearly impossible to overcome, and others with high variance where it is fairly common.

Another problem is that not all games play well with a dummy player replacing the person who was expelled or by skipping the expelled player's turn. It seems equally wrong to force anyone to stay in the game with a dummy/skipped player if that completely changes the game, regardless of their rank at the time the timed out player is expelled. Between that and awarding them "fake" ELO that they likely wouldn't have earned, *shrug* the fake ELO doesn't seem like it causes some great imbalance in the universe that can't be abided. Does it encourage them to "call flag"? Sure, but nothing's perfect.