What is the logic of lichess rating giving system ? • page 1/6 • General Chess Discussion • lichess.org

ABID-1HASAN

#1

I saw in my games when I win I got 5 but when I lose I got -8. What is this ?

#2

Autofill

#3

Facing different opponents with different rating

#4

Autofill

#5

@Blundered_the_queen said in #4:

On a similar subject, I've been wondering for a while now - is it generally easier to gain rating against higher-rated or lower-rated players?

I'm basically trying to figure out what is the best strategy from a more meta perspective. Should I play 10 strong players a day and hope one of them dozes off in the middle of the game, or should I go sealclubbing and bully 10 noobs a day, and just hope I don't hang checkmate in 1 to any of them?

My insights tab suggests that I generally gain rating against much higher rated players as well as much lower-rated players, but I'm dealing in a very small sample size in the double-digits here. Can anyone else who's played a lot of games share their insights so we could hash out the best strategy? Thanks.
My insights show me losing rating against weaker players

@Blundered_the_queen said in #4: > On a similar subject, I've been wondering for a while now - is it generally easier to gain rating against higher-rated or lower-rated players? > > I'm basically trying to figure out what is the best strategy from a more meta perspective. Should I play 10 strong players a day and hope one of them dozes off in the middle of the game, or should I go sealclubbing and bully 10 noobs a day, and just hope I don't hang checkmate in 1 to any of them? > > My insights tab suggests that I generally gain rating against much higher rated players as well as much lower-rated players, but I'm dealing in a very small sample size in the double-digits here. Can anyone else who's played a lot of games share their insights so we could hash out the best strategy? Thanks. My insights show me losing rating against weaker players

AlexiHarvey

#6

Strictly speaking as far a rating goes it doesn't really matter whether you target strong or weak players. However the rating gain/loss is capped at +/-32 per game for regular players so for efficiency you should remain within a +/- 400 range.

For example, if you are 1000 and play and win against someone 1800 you gain +32 rating, the exact same if you had won against a 1400 player. You have a 1 in 10 chance of beating someone +400 and a 1 in 100 of beating someone +800. These figures maybe slightly off, but this is the general pattern.

As far as chess improvement goes you're best aiming to play better players as you learn more from losses than wins. So if your ego can take it, nudge towards higher rated players with an expectation of losing more than 50% of your games.

Strictly speaking as far a rating goes it doesn't really matter whether you target strong or weak players. However the rating gain/loss is capped at +/-32 per game for regular players so for efficiency you should remain within a +/- 400 range. For example, if you are 1000 and play and win against someone 1800 you gain +32 rating, the exact same if you had won against a 1400 player. You have a 1 in 10 chance of beating someone +400 and a 1 in 100 of beating someone +800. These figures maybe slightly off, but this is the general pattern. As far as chess improvement goes you're best aiming to play better players as you learn more from losses than wins. So if your ego can take it, nudge towards higher rated players with an expectation of losing more than 50% of your games.

ok lol

@Blundered_the_queen said in #4:

On a similar subject, I've been wondering for a while now - is it generally easier to gain rating against higher-rated or lower-rated players?

makes no difference. If you play stonger you win more per game but lose more. and other way round against weaker.
That is the whole point of rating system.

Probably it is not accurate if there more than 300-400 points difference but those games are quite boring anyway

@Blundered_the_queen said in #4: > On a similar subject, I've been wondering for a while now - is it generally easier to gain rating against higher-rated or lower-rated players? > makes no difference. If you play stonger you win more per game but lose more. and other way round against weaker. That is the whole point of rating system. Probably it is not accurate if there more than 300-400 points difference but those games are quite boring anyway

AlexiHarvey

#10

In a magic world where chess talent was fixed, it would make no difference whatsoever whether play was biased towards strong or weak players. However giving reality, you are best playing strong players getting worst and avoiding weak players getting better. This maybe stating the obvious somewhat but noticed the difficulty in detecting the future direction of a player's rating!

A much bigger factor would be the general fuzziness of players' ratings at a giving moment in time. For example, usually something like 30 games are required to establish a player's classical/OTB rating, yet such games can't be played within an hour. Worst when you consider the Blitz format you could easily be looking at 300+ games to establish a player's rating. Clearly ratings can only be absolutely accurate if and only if playing chess results in no changes to game talent, which is extremely doubtful.

The short answer to your question is that the matter is undeterminable, and biasing your gameplay one way or another will result in a gain with the same frequency as correctly predicting the outcome of a coin toss.

In a magic world where chess talent was fixed, it would make no difference whatsoever whether play was biased towards strong or weak players. However giving reality, you are best playing strong players getting worst and avoiding weak players getting better. This maybe stating the obvious somewhat but noticed the difficulty in detecting the future direction of a player's rating! A much bigger factor would be the general fuzziness of players' ratings at a giving moment in time. For example, usually something like 30 games are required to establish a player's classical/OTB rating, yet such games can't be played within an hour. Worst when you consider the Blitz format you could easily be looking at 300+ games to establish a player's rating. Clearly ratings can only be absolutely accurate if and only if playing chess results in no changes to game talent, which is extremely doubtful. The short answer to your question is that the matter is undeterminable, and biasing your gameplay one way or another will result in a gain with the same frequency as correctly predicting the outcome of a coin toss.