4 ms·

A rating difference of 100 points means that the stronger player will collect 3 out of 4 points over 4 games. A difference of 200 points is a full category jump

A rating difference of 100 points means that the stronger player will collect 3 out of 4 points over 4 games. A difference of 200 points is a full category jump: Candidate Master to Master, Master to Internation Master and International Master to strong Grandmaster. This is not something that can be achieved in just 16 days of playing/training. As already pointed out by other posts there is a lot of randomness especially when playing blitz and rapid chess. All the improvement suggestions are valid ones ( and mostly common sense for anybody that has read some chess books ) but attributing a 200 Elo point increase to 16 days of training to me tells that this player is probably not worth 2000 yet.

rgossiaux 8y ago
Indeed, the idea of 200 points in 16 days is more or less absurd, at least beyond beginner levels. In fact, looking at the author's Lichess profile I see no evidence of any improvement at all, to be blunt. Improvement takes really a lot of work, far more than I think the author appreciates. Going from 1800 to 2000 in a year would be an excellent outcome and would require hundreds of hours of work on chess. The author dismisses the one method that is arguably most recommended by strong players: studying GM games. Chess is about playing good moves, and studying the games of the world champions helps you learn what good moves look like. Studying GM games is hard work with little tangible reward (no dopamine rush from finishing a puzzle or playing a quick game) but it seems highly effective. Another thing that stuck out to me: the author dismissed tactics puzzles based on what looks like 30 puzzles done on Lichess. That's nothing. I've done 7500 tactics on Chesstempo standard mode alone and it did a lot for my chess. It was basically the only thing I did to go from 1300 to 1600 or 1700 or so (over the board). Tactics aren't a panacea for chess improvement, but neither is anything else. I have some smaller comments (for example: re: "If I want to get difficult puzzles correct, I don’t rely on any pattern recognition. I just spend a lot of time and mental energy explicitly calculating out every combination of moves. I’m really not learning much, most people can get the same puzzle correct if they just thought as hard as I." -- pattern recognition is the basis for calculation skill so it doesn't make sense to say you don't rely on it when calculating) but the main point is that chess is hard and improving takes a lot of hard work, far more than the author makes it seem.
- zodiac 8y ago
  This is unrelated to your main point but I've always been confused why people cited an "over the board" ELO. Is your 1600 rating FIDE, USCF or something else? Aren't they not calibrated against each other?
coryfklein 8y ago
Moving from 1800-2000 must certainly be very different than 2400-2600, is it not? I assume the higher levels of play follow a bell curve with a long tail. It seems within reason that one could move from 800 to 1000 in 16 days with very basic study and coaching. > there is a lot of randomness especially when playing blitz and rapid chess This is likely true, but this randomness appears to be averaged out by the sheer amount of games he played over the ~6 month period in the chart.
tylerhou 8y ago
LiChess ratings are inflated; I'm ~1800 on LiChess but I probably am 1200 in real life.
- JamesBarney 8y ago
  They're inflated cose to 200-250 points not 600. Why do you think you're only a 1200 in real life?
hollerith 8y ago
>A rating difference of 100 points means that the stronger player will collect 3 out of 4 points That would be correct if you'd written 200 points.

rgossiaux 8y ago

Indeed, the idea of 200 points in 16 days is more or less absurd, at least beyond beginner levels. In fact, looking at the author's Lichess profile I see no evidence of any improvement at all, to be blunt. Improvement takes really a lot of work, far more than I think the author appreciates. Going from 1800 to 2000 in a year would be an excellent outcome and would require hundreds of hours of work on chess. The author dismisses the one method that is arguably most recommended by strong players: studying GM games. Chess is about playing good moves, and studying the games of the world champions helps you learn what good moves look like. Studying GM games is hard work with little tangible reward (no dopamine rush from finishing a puzzle or playing a quick game) but it seems highly effective. Another thing that stuck out to me: the author dismissed tactics puzzles based on what looks like 30 puzzles done on Lichess. That's nothing. I've done 7500 tactics on Chesstempo standard mode alone and it did a lot for my chess. It was basically the only thing I did to go from 1300 to 1600 or 1700 or so (over the board). Tactics aren't a panacea for chess improvement, but neither is anything else. I have some smaller comments (for example: re: "If I want to get difficult puzzles correct, I don’t rely on any pattern recognition. I just spend a lot of time and mental energy explicitly calculating out every combination of moves. I’m really not learning much, most people can get the same puzzle correct if they just thought as hard as I." -- pattern recognition is the basis for calculation skill so it doesn't make sense to say you don't rely on it when calculating) but the main point is that chess is hard and improving takes a lot of hard work, far more than the author makes it seem.

zodiac 8y ago
This is unrelated to your main point but I've always been confused why people cited an "over the board" ELO. Is your 1600 rating FIDE, USCF or something else? Aren't they not calibrated against each other?

coryfklein 8y ago

Moving from 1800-2000 must certainly be very different than 2400-2600, is it not? I assume the higher levels of play follow a bell curve with a long tail. It seems within reason that one could move from 800 to 1000 in 16 days with very basic study and coaching. > there is a lot of randomness especially when playing blitz and rapid chess This is likely true, but this randomness appears to be averaged out by the sheer amount of games he played over the ~6 month period in the chart.

tylerhou 8y ago

LiChess ratings are inflated; I'm ~1800 on LiChess but I probably am 1200 in real life.

JamesBarney 8y ago
They're inflated cose to 200-250 points not 600. Why do you think you're only a 1200 in real life?

hollerith 8y ago

>A rating difference of 100 points means that the stronger player will collect 3 out of 4 points That would be correct if you'd written 200 points.