Can't create game with specific side any more

@NotTakenUsername said in #529:

''Risk facing an abuser''. I have created custom games several times, I was never suspicious of any ''abuser'', the stats prove this. Even if I faced one or two of them, it's still regular chess, no risk included. Playing with white doesn't give you a huge advantage.

It doesn't give you a huge advantage, but playing exclusively with white certainly gives you an advantage. I don't know which stats prove that you were never suspicious of any abuser. But again, even if you were not bothered by it, it was still there. I have never suspicious of sandbaggers in the lobby, they were still there though.

The post is irrelevant here, the post was about new things added and the reasons why such additions may not be accepted, even if well implemented. Unless color selection causes such a massive cost/resources upkeep (press X to doubt)

Would be interested to see your cost estimation of cost and resources per year for this feature

Make a poll and ask the players whether they want the color selection wiped. Pretty sure the majority of the players want to revert this, even though there are some ''abusers'' lurking around.

That would be interesting. But only people with say 1000 games, approximately 50/50 white/black and 1 year on the platform are to vote. Open to have my mind changed about who is bothered by it, and what solution they deem acceptable. Going back to the way it was, is a definite big no.

Choosing the color isn't the same as cheating at all. I'm pretty sure I wouldn't be able to beat Magnus with white, but maybe you are built different. I know you brought cheating up as an analogy, but it's a terrible one.

It's a great analogy, they both provide unfair play for the opponent. Okay, so let's say we allow use of chess engine, but only during the opening (first 12 moves). You would still not be able to beat Magnus, but definitely have a better chance of beating your peers overall.

It's either an unfair advantage or it isn't. If it is, using it excessively means you're abusing it to your advantage. Just because at some level you are unable to consistently convert this, doesn't mean it's fair to other players.

It's indeed great that I didnt notice this personally and given that I have a large amount of games, statistically wise, it indicates that there aren't many people abusing the feature.

I'm curious how you determined this statistically. Could you share the numbers/calculations on this?

The benefits of keeping it outweigh the damage that those ''abusers'' cause. If I'm in the lucky minority that wasn't affected by those people, I would be glad to see a poll that disproves what I claim. If the result of the poll is in favor of the removal, I'll agree with this change. But I have a suspicion that the poll would suggest otherwise.

I've played with cheaters in the past. And? Do you remove standard chess because of those cheaters or do you take care of each individual case?

No, this is actually making my point. The cheaters abused a feature (not having cheating detection), so the feature is removed (in this case, cheating detection is added to detect and disallow cheating). They did the same for color picking. In both cases, the abusers are limited in their ability to exploit the site, but the normal people are playing chess like they always were

@NotTakenUsername said in #530:

Burden? Really?
Yes

There was not a problem. Deal with individual cases with 90% games with a specific color instead of punishing everyone. Or just keep the feature, if moderating is costly.

There obviously was a problem, 89k games with white vs 0 with black is a huge problem.

Create a poll (not you personally, whoever took the decision) asking the players whether they want the feature or not. If the majority of the players notices something wrong, remove it.

The poll should be a bit more sophisticated, giving several options, but I would welcome this for sure!

The correct analogy in this situation would be this: There are cheaters playing chess, therefore we remove standard chess from Lichess.

No, because both cases are people using something to their opponents disadvantage, and that something should be taken away to protect fair users

@mstef0 said in #520:

Please revert this at the very least for casual games. When I try learning a new opening, I'll play hundreds of CASUAL blitz games with a given color against humans and bots. I find it very useful.

You can still do so against bots! Also against humans, on invite

This thread is by far the most active one this month – with over 530 responses!

It's important to emphasize: When someone offers a website or software for free, no one has the right to demand changes. If the service were paid, the situation might be different.

As a software developer who has been providing Windows freeware for years, I speak from experience. When I believe a change is necessary, I implement it – regardless of how loudly users complain or threaten to switch to another software. If someone doesn't like my software, they're free to use something else.

That said, I do listen to users. Often, it's their suggestions and ideas that drive development forward. However, I don't run my software purely out of goodwill – I also have business interests. Good software can be great marketing. It might lead to opportunities to display ads, offer paid services, or even develop a franchise.

Lichess is free for users, but the operator may have economic considerations as well. Here, they must weigh whether the current change truly benefits the project or if it merely caters to the fairness concerns of a few players.

I found the idea of a survey very intriguing. Is this change really seen as a positive by many users here? I'd be genuinely interested in knowing that.

I'm deeply impressed by the stamina of @BeDecentForAChange who has stood unfazed in this hailstorm of attacks, garbage and worse for four days now. I bow to you.

@duran_was_the_mvp said in #531:

It doesn't give you a huge advantage, but playing exclusively with white certainly gives you an advantage. I don't know which stats prove that you were never suspicious of any abuser. But again, even if you were not bothered by it, it was still there. I have never suspicious of sandbaggers in the lobby, they were still there though.

Okay and? The core of the analogy is: People abuse feature, therefore feature must be removed. What do sandbaggers abuse? Rating. So the correct analogy in this situation would be: There are people abusing their rating, artificially lowering it, therefore we must get rid of the rating.
Does lichess do that? Ofcourse not. So why don't you do the same thing for the so called abusers?

Would be interested to see your cost estimation of cost and resources per year for this feature
From the tone of your message, I can see you are far more qualified, so by all means, please, shed some light on this.

That would be interesting. But only people with say 1000 games, approximately 50/50 white/black and 1 year on the platform are to vote. Open to have my mind changed about who is bothered by it, and what solution they deem acceptable. Going back to the way it was, is a definite big no.
If reverting a downgrade is out of the question, even in the case where the poll result shows that most people want the colour choice to be reinstated, what is the point?

It's a great analogy, they both provide unfair play for the opponent. Okay, so let's say we allow use of chess engine, but only during the opening (first 12 moves). You would still not be able to beat Magnus, but definitely have a better chance of beating your peers overall.
Not really.

It's either an unfair advantage or it isn't. If it is, using it excessively means you're abusing it to your advantage. Just because at some level you are unable to consistently convert this, doesn't mean it's fair to other players.
Deal with individual cases if that's such a big problem. Ban people with a certain % of games playing with a specific colors, if manual moderating is costly.

I'm curious how you determined this statistically. Could you share the numbers/calculations on this?
I guess I'm in the marginal minority that was treated fairly. Lucky me.
Even if I got ''color abused'' in 5% of my games, the games were still the same and fair. I'm pretty sure most players would agree with this, that it's more important to have the feature than to remove it for the sake of ''fairness''.

As to how much rating I have lost to those abusers, I've no idea how to track them so it's hard to calculate, however I found a relevant post on Reddit, calculating how much more Elo a top 2700 player would have if they played white only.

FIDE uses the Elo rating system. Under this system, a player's expected score against another player can be calculated using the formula:
Expected Score of Player A = 1/(1+10^((Player B Rating - Player A Rating)/400))

White's exact advantage is not conclusive, but in top level play white seems to score about 55%
Plugging that into the formula, we can find the rating differential that yields that expected score.

0.55 = 1/(1+10^(x/400)) simplifies to (9/11) = 10^(x/400) which simplifies further to x = log_10(9/11) * 400
Solving for x gives the result of -34.86.

''A player is always expected to score 50% against another player of the same rating (1/(1+10^0) = 1/2).
This means that if they were to consistently score higher, in this case because they always have the white pieces, they would gain rating. As calculated above, this specific advantage would lead to a rating gain of ~35 points.

Therefore to answer your question a 2700 player would gain a new rating of 2735.''

Yea, ofcourse some players perform better in white than in black, but for the sake of this conversation, lets say, I score 55% with white and 45% with black.
If I was constantly paired with abusers, my elo would be 35 points lower. Of course, Lichess doesn't use Elo, but the results wouldnt be that far off.
Yea, this whole discussion is done for 35 rating points. I would argue thats the same as sandbagging and cheating.

No, this is actually making my point. The cheaters abused a feature (not having cheating detection), so the feature is removed (in this case, cheating detection is added to detect and disallow cheating). They did the same for color picking. In both cases, the abusers are limited in their ability to exploit the site, but the normal people are playing chess like they always were

If cheaters abuse the feature of playing chess, remove chess. I solved it guys.

@BeDecentForAChange said in #532:

Yes
Nope

There obviously was a problem, 89k games with white vs 0 with black is a huge problem.
Deal with them like you deal with cheaters.

The poll should be a bit more sophisticated, giving several options, but I would welcome this for sure!
We are on the same page here.
No, because both cases are people using something to their opponents disadvantage, and that something should be taken away to protect fair users
Deal with color abusers like you deal with cheaters.

@NotTakenUsername said in #536:

Okay and? The core of the analogy is: People abuse feature, therefore feature must be removed. What do sandbaggers abuse? Rating. So the correct analogy in this situation would be: There are people abusing their rating, artificially lowering it, therefore we must get rid of the rating.
Does lichess do that? Ofcourse not. So why don't you do the same thing for the so called abusers?

No, the core analogy: People abuse features, so their ability to abuse it must be stopped. Sandbaggers abuse rating, so their ability to do so is stopped by early resignation detection. Fair players should still be given the option to resign, and fair players should still be given the option to play for rating.
Has Lichess achieved this? Absolutely!

From the tone of your message, I can see you are far more qualified, so by all means, please, shed some light on this.

I could be, but you are making claims about this, so I was interested to hear about the underlying calculations for this

If reverting a downgrade is out of the question, even in the case where the poll result shows that most people want the colour choice to be reinstated, what is the point?

It's a huge upgrade, not a downgrade. Of course we should not roll back abuse-prevention.

Not really.
Yep

Deal with individual cases if that's such a big problem. Ban people with a certain % of games playing with a specific colors, if manual moderating is costly.

Yes, that would be treating it the same as cheating or sandbagging. You could do it a couple of times, but abuse would result in a ban. I could see that working.

I guess I'm in the marginal minority that was treated fairly. Lucky me.
Even if I got ''color abused'' in 5% of my games, the games were still the same and fair. I'm pretty sure most players would agree with this, that it's more important to have the feature than to remove it for the sake of ''fairness''.

I'm sure the majority of players has no direct problematic experience with sandbagging. They get free rating points and very little resistance in their game when he is pushing his rating down. It could be a 2000 level player who's playing among the 1200s. When the games starts, it's totally fair and the same, and it's only one game against that person.
Except it isn't fair, and we should not be allowing sandbagging back, even though I'm sure on a hunch that most people wouldn't mind it.

As to how much rating I have lost to those abusers, I've no idea how to track them so it's hard to calculate, however I found a relevant post on Reddit, calculating how much more Elo a top 2700 player would have if they played white only.

FIDE uses the Elo rating system. Under this system, a player's expected score against another player can be calculated using the formula:
Expected Score of Player A = 1/(1+10^((Player B Rating - Player A Rating)/400))

White's exact advantage is not conclusive, but in top level play white seems to score about 55%
Plugging that into the formula, we can find the rating differential that yields that expected score.

0.55 = 1/(1+10^(x/400)) simplifies to (9/11) = 10^(x/400) which simplifies further to x = log_10(9/11) * 400
Solving for x gives the result of -34.86.

''A player is always expected to score 50% against another player of the same rating (1/(1+10^0) = 1/2).
This means that if they were to consistently score higher, in this case because they always have the white pieces, they would gain rating. As calculated above, this specific advantage would lead to a rating gain of ~35 points.

Therefore to answer your question a 2700 player would gain a new rating of 2735.''

Yea, ofcourse some players perform better in white than in black, but for the sake of this conversation, lets say, I score 55% with white and 45% with black.
If I was constantly paired with abusers, my elo would be 35 points lower. Of course, Lichess doesn't use Elo, but the results wouldnt be that far off.
Yea, this whole discussion is done for 35 rating points. I would argue thats the same as sandbagging and cheating.

Okay, so it is an unfair advantage, but only a small one. That's my whole point, it's abuse of a feature to gain an unfair edge

If cheaters abuse the feature of playing chess, remove chess. I solved it guys.

Wrong analogy again. The player's ability to use an engine should be removed, nothing more or less. Which is what Lichess achieved

@NotTakenUsername said in #537:

Deal with color abusers like you deal with cheaters.

This could be a good solution, mb better than to punish all players. This would give players, they train for openings (both colours one after the next) here, a chance to stay.

I am irritated, because the existing problem not has been discussed with us players, before a wrong decision of changes has been brought to us.

Another voice asking to revert this feature for casual games, I enjoyed selecting my side as black to train defenses, please bring it back, thanks!