I’m surprised you don’t do puzzles. Gotham says anyone who’s trying to improve should be doing 15 puzzles everyday.
i’m like 2100 in puzzles and 1400 blitz. i probably take too much time with each puzzle tbh. really i love doing them, but they don’t transfer well.
im around 900 points higher in puzzles than in blitz. i often will take 5-6 minutes to solve one, feel really smart when i do but then also realize i never would have found that move in a timed game. i think endgame puzzles are the one place where i have actually been able to translate the puzzle solutions into my timed games.
2 Likes
I’ve been teaching my 6-year old chess using Chesskid.com. It’s pretty good, but quickly goes beyond this age range, imo. Probably a better site for 7-8 year-olds who can read a bit better. Now I’m trying out the book Chess is Child’s Play and it’s a fantastic supplement. It includes lots of mini-games that teach piece movement and other fundamental concepts. Most kids want to play a game rather than learn endless rules, something these authors really understand.
My kid has the ladder mate down :)
1 Like
They’ve got a twitch stream with an FM named Mike that is pretty entertaining. It’s the right balance between being for kids, but not treating them like they’re stupid.
Pretty great mini-match between Carlsen and Nakamura today. Every game decisive and high-quality play.
Carlsen-Nakamura recap starts 17:00
Hikaru played one of the games at 99.7% CAP score on chess.com, which is actually 100% because it’s the highest it will ever give you. I know this because I once played a 16-move game consisting of 3 book moves and 13 best moves and got 99.7.
Edit: Sad news, I dug the game out and ran depth 30 analysis on it and it downgraded one of the moves to Excellent.
Yeah OK computer, the forcing move which forces a win is a little less strong than the non-forcing move where it turns out they also can’t do anything. This is why nobody invites you to parties.
1 Like
I suck at chess, but why exactly is Ng5 better? If black follows that with g6, it seems like he buys enough time to mount a defense.
g6 Nf3 threatens Bh6 and Bg5?
If g6 there’s Nxh7. If you take there’s Qh5+ as g6 is pinned and you’re getting mated in short order, if you don’t take the knight comes back to f6 and you’re getting mated in short order. I saw this myself but I’ve confirmed it with engine just now.
However, it turns out that Bxh7 is actually better, in fact it’s a forced mate in 12. I found this accidentally analyzing it in browser just now. Stockfish didn’t find this on depth 30. It’s a common misconception that engine depth means that the engine has exhaustively searched that many ply ahead, but that would be impossible. Instead it’s the nominal search depth that the engine has reached. Individual branches might be searched deeper than that (for example in the case of something like quiescence search) or shallower than that (if terminated by a pruning algorithm).
When the engine can’t find a forced mate, Ng5 is “better” because White’s position is so overwhelming that at depth 30, what it wants in both lines (meaning Ng5 immediately or Bxh7 Kxh7 Ng5+) is for Black to sacrifice his queen for the knight to delay mate. Ng5 is therefore better in the sense that if Black is going to sac the queen anyway, there’s no point giving up a bishop first. But it turns out that saccing the queen in the Bxh7+ line doesn’t actually delay mate.
If he takes the knight, you can take his pawn with your bishop and mate.
2 Likes
I think? it’s meant to be the equivalent of being up 14.8 pawns.
3 Likes
the typical scale for chess is pawn =1 , bish/knight=3, rook=5, queen=9
so +14 is like being up a queen + a rook
3 Likes
After you have reached the solution (either by solving it or clicking View Solution) the engine will answer these questions for you. I just jumped on for an example problem:
When I play Rf7+, Black responds with Re7, interposing the rook, whereupon Ra7+ wins the rook. Suppose I’m like “wait, why can’t Black just move the king back?”. I can click on the Rf7 move in the analysis on the right and then play Kc8 on the board. The engine then shows me why Black couldn’t do this:
3 Likes
I’m really bad but trying to get better.
Chess log for 5/22 - 5/29
All games on Chess.com at 15+10 time control.
Puzzles: +99 (start: 1468. end: 1567)
Score: 45/70
Games: -44 (start 1097. end: 1044) :(
Score 4/11
Lessons: 6
More detail
Summary
5/22/2021
Puzzles1468: 11/15 – +52
Lessons: 0
Games: 0/1 – -8.
Lose: Great position until blunder in the end game. Didn’t see I was getting mated.
5/23/2021
Lessons: 2
Puzzles 1520: 11/15 — +38
Games: 1070: 2/4 +0
Win: wild game – opponent blunders a piece and resigns despite being +8
Lose: - bad calculation of a trade blundered a rook -6
Lose: bad exchange and low time. -8
Win: good pawn pushes. Caught some blunders. +6
5/24/2021
Lessons: 0
Puzzles: 1558. 7/15 – -10
Games: 1070 +2
Win: punished a wild opening. +6
Lose: into endgame +1 Then blunder rook for no reason -9
Win: Punished a wild opening. Just played boring and solid. +5
5/25/2021
Lessons: 0
Puzzles: 1548: 12/15 +58
Games: Lose with Black. – crushing +8 and then blunder several times in time pressure. -7
5/26/2021
Games: Lose with Black. Two random blunders in time pressure. -9
5/27/2021
Games: Lose with Black. Blunder Queen for no reason. -8
5/28/2021
Puzzles: 4/10
Lesson: 2
1 Like
Do you replay the games in the engine and try to fix mistakes?
Yes, I use the chess.com game analysis to review. Usually I try to sum up the game in a sentence or so. Some of that is in the More detail drop down above. I’m sure I could do a better job analyzing.
Your rating will settle down to 1000 really quickly if you just start playing.
1 Like
Your first ~10 games your rating fluctuates by a few hundred points until it settles in, then it starts the normal ELO scoring where it’s ± ~8 points/game.
I guess killing it means “barely made it out of the qualification stage.”
1 Like