lichess.org
Donate

The problem that never dies

Instead of reducing it in all ultrabullet games, we can make an option to create tournaments with different limits for lag compensation. Like 0, 50, 100, 150, 200, 300, 500, 1000. Players with high ping will still be able to participate in tournaments and play games, and their time will be compensated as usually (even with high ping). It just will be more difficult for them to find a game, but still possible.
The problems are these:
1. When playing against a lagging player you have something more like 1/4+0,5, which makes it a longer game, and therefore you have to change your style of play from flagging style to more accurate. Ultra attracts more players who like to play messy moves, that's why these players like ultra. And when you are paired against a lagger, you can't play messily, you have to be more careful. Losing a piece against a lagger might be critical, while losing a piece (even a queen) in a game of two players with low ping won't matter that much, because speed is more important. In a game against a lagger speed is less important and classical chess skills become a key for a victory. Therefore many players (and many players, if not most, play ultrabullet because it allows to be messy and fast, without too much caring about moves) have a huge disadvantage against a lagger, who is used to their slower pace and is prepared to be more careful in moves (because laggers are more of a common type of players, who come in ultrabullet not just for messy moves and speed, but for something which is similar to them as 0+1, so amongst laggers we have more slow, classical type of players, which will outrun usual ultrabullet player in slower time control, even with low ping, because they are better at slower time controls, such as 0+1, 0+2).
2. The second reason is that a lagger will always feel less difference when playing against a lagger or non lagger, because the maximum difference for them is only two. For example, a player with 300ms ping has a range of 310-750. The difference is 2.4. But a player with 70ms ping has much bigger range. From 80 to 420. The difference is 5.2. So a player with lower ping will always have a disadvanage, because they have to change their pace more than a lagger.
"it also interferes with the the meaning of your rating, if the different lag comp tourneys were rated." The interference exists now, because ultrabullet sometimes is interfered by 1/4+0.4 time control, with lagging players we already have messed up rating. I often lose to laggers, even if they are 200 points lower rated than me, but I am on par with 2300 players, who have very very low ping. Because with 2300 rated non-laggers I play actual ultrabullet, and with laggers I play something like 1/4+0.3, which gives more time on every move to think, which is an advantage for my opponents who are used to slower time controls. And in such cases I cannot play messily anymore (which is the gist of ultrabullet), and I lose frequently.
Random idea: cap compensation based on sqrt(x * y) of players' average ping times x and y. Or avg(x, y), or some other math function...

Another random idea: add compensation to both players' clocks, not just the lagging player. Probably unpopular but maybe fair.
"Random idea: cap compensation based on sqrt(x * y) of players' average ping times x and y. Or avg(x, y), or some other math function..."
Seems very interesting. But not sure if it will work, because often a player with the highest ping will quit, so lag cap will be reduced, and then another player with the highest ping will quit, reducing it even more, etc.
> The interference exists now, because ultrabullet sometimes is interfered by 1/4+0.4 time control, with lagging players we already have messed up rating.

You're asking for different time control rules in rated tournaments compared to lobby games, for the same ultra bullet rating category. I'm not going to make further changes until the new lag comp caps goes live and I can look at whether they're reasonable to keep.
What about something that could just warn the player that the oppositon will get lag comp ? Ofc you can just check their lag before the game, but it's a reflex that only comes after you got surprised by some of these games in a tournament .
And if you're aware, then changing the pace isn't really hard (we often slow down against known laggers anyway right)
I agree with Chesstroll and the majority of players wanting a solution to the lagging issue. Issacly has the best of intentions to provide a fair playground and include the majority of players.
That said, it is sad that this variant although it started attracting a lot of players and big names it is now reduces to an endless debate about lag and most of the times a lag festival.
Lag comp does not work well in xtremely fast games. Yesterday i watched a game where one player was lagging heavily and they played 71 moves.
The game was going on for ever. This is the exact opposite of Ultra. Only way is either to follow a lower cap in lag comp or allow during the creation of a tournament the option to restrict heavily lag users from playing. They can always create their own lag extravaganza and have a blast ;-)
> fair playground

Indeed, if we got rid of lag comp for ultra, it would *strongly* favor european players who have lag in the 20ms range over players in North America or Asia.

> @Tr0n Yesterday i watched a game where one player was lagging heavily and they played 71 moves. The game was going on for ever. This is the exact opposite of Ultra

I see a couple solutions:

1) (easiest) Accept that long premove sequences can occur in ultra
2) Reduce lag comp for ultra to 200-250ms. ultra is its own category and it's OK if it is a bit more restrictive in terms of internet quality required to be competitive
3) Improve visibility of how much your opponent is lagging and/or how long they took on each move.
4) Allow tourneys with custom lag comp. Essentially a non-starter, as it's a feature that's only useful for ultra/hyper, it will limit our flexibility to change lag comp in the future, and it adds complexity to our code base for variant with a small playerbase
5) Break the 0s premove sequence. One option: if there's been at least 2 premoves in a row (i.e. both you and your opponent, subsequent premoves incur a small time penalty like 0.05s. This is possible although would need to wait for an upcoming code refactor to allow the clock logic visibility into clock/premove history.
At first thank you for taking time and steps around this issue,
Although 4) is a much sought after feature your points about small player base and complexity are correct and well understood.
For me options 2 and 3 are the best and if done i believe the whole lag issue can be put to rest afterwards.For3) specifically will it be possible to showcase the lag ie 270ms and if so would that be the actual lag or the compensated one? In any case it will be great to see this feature.
Once more thank you !!!!

This topic has been archived and can no longer be replied to.