Why are tactics critical for chess improvement at all levels?

@Wodjul said in #8:

Rapid recognition of simple tactical patterns is essential. I don't have this rapid recognition and I have to calculate everything, even one move and two move tactics. It is hard to even call a one move "tactic" a tactic at all. It is hard to believe that a one move "tactic" needs calculation or at least checking. But I can tell you that in my case it does need this. This slows me down terribly when I play chess. I am always getting into time trouble in Rapid 15+10 and it is impossible for me to play time limits faster than that.

I believe I actually need to do great numbers of one move and two move puzzles and that these have to be done by themes for the repetition effect. I have been doing a Chessable course that is actually a little bit too hard for me in some puzzles; meaning quite a few are 3 moves and some are even 4 moves. I can see how lost I get in the harder 3 move puzzles and in almost any 4 move puzzle.

Sight-of-board exercises, one move puzzles and two move puzzles are really the basic building blocks of training I should be doing in bulk, I think. If I can't see every one-mover in at least five seconds to ten seconds and every two-mover in say twenty to 30 seconds then seeing / calculating any longer tactics will be completely beyond me in Rapid, The longer tactics are completely beyond me for the most part.

I just wonder what the sweet spot for learning speed is. Should I do random selections of one-move puzzles at say 1700 rating to test myself? When I can do 50 in 250 seconds (2.5 minutes) would that indicate I am fast enough, at one-movers anyway? Or should I, as a low rated player, still learning very basics, be happy to do 50 of 1700 rating one-move randoms in 5 minutes?

Note: I will still study by themes but I will test my speed by random puzzles. I think that might be the way to go to get a handle on this.

I think that it will be a mix. You should always work on one move and two move tactics. Once you are comfortable enough with those, you could add a bit of a 3 movers, lets say. I think you are thinking about this correctly, and that as long as you work on tactics, there is no bad way of doing it.

@Wodjul said in #8: > Rapid recognition of simple tactical patterns is essential. I don't have this rapid recognition and I have to calculate everything, even one move and two move tactics. It is hard to even call a one move "tactic" a tactic at all. It is hard to believe that a one move "tactic" needs calculation or at least checking. But I can tell you that in my case it does need this. This slows me down terribly when I play chess. I am always getting into time trouble in Rapid 15+10 and it is impossible for me to play time limits faster than that. > > I believe I actually need to do great numbers of one move and two move puzzles and that these have to be done by themes for the repetition effect. I have been doing a Chessable course that is actually a little bit too hard for me in some puzzles; meaning quite a few are 3 moves and some are even 4 moves. I can see how lost I get in the harder 3 move puzzles and in almost any 4 move puzzle. > > Sight-of-board exercises, one move puzzles and two move puzzles are really the basic building blocks of training I should be doing in bulk, I think. If I can't see every one-mover in at least five seconds to ten seconds and every two-mover in say twenty to 30 seconds then seeing / calculating any longer tactics will be completely beyond me in Rapid, The longer tactics are completely beyond me for the most part. > > I just wonder what the sweet spot for learning speed is. Should I do random selections of one-move puzzles at say 1700 rating to test myself? When I can do 50 in 250 seconds (2.5 minutes) would that indicate I am fast enough, at one-movers anyway? Or should I, as a low rated player, still learning very basics, be happy to do 50 of 1700 rating one-move randoms in 5 minutes? > > Note: I will still study by themes but I will test my speed by random puzzles. I think that might be the way to go to get a handle on this. I think that it will be a mix. You should always work on one move and two move tactics. Once you are comfortable enough with those, you could add a bit of a 3 movers, lets say. I think you are thinking about this correctly, and that as long as you work on tactics, there is no bad way of doing it.

Wodjul edited

#12

@HelloItsDmitri said in #11:

I think that it will be a mix. You should always work on one move and two move tactics. Once you are comfortable enough with those, you could add a bit of a 3 movers, lets say. I think you are thinking about this correctly, and that as long as you work on tactics, there is no bad way of doing it.

There is perhaps no bad way but maybe are there less efficient ways and more efficient ways? I have a quick question.

If I was studying 1-move forks, do you think it would be less efficient to do 70 random forks and more efficient to 10 forks for the pawn and then 10 for each of the six pieces in turn, to get a better repetition reinforcement effect? If training was to be even more efficient, then perhaps we would need data on how common each type of fork was in games above say 2000 rating and then proportion study by those frequencies. If knight forks are 30% of all forks in live play above 2000 (just a guess) then the training in knight forks should be 30% of all fork training, perhaps?

I have more questions of this type but is this sort of thinking simply going too far in trying to make training a fully efficient use of time?

@HelloItsDmitri said in #11: > I think that it will be a mix. You should always work on one move and two move tactics. Once you are comfortable enough with those, you could add a bit of a 3 movers, lets say. I think you are thinking about this correctly, and that as long as you work on tactics, there is no bad way of doing it. There is perhaps no bad way but maybe are there less efficient ways and more efficient ways? I have a quick question. If I was studying 1-move forks, do you think it would be less efficient to do 70 random forks and more efficient to 10 forks for the pawn and then 10 for each of the six pieces in turn, to get a better repetition reinforcement effect? If training was to be even more efficient, then perhaps we would need data on how common each type of fork was in games above say 2000 rating and then proportion study by those frequencies. If knight forks are 30% of all forks in live play above 2000 (just a guess) then the training in knight forks should be 30% of all fork training, perhaps? I have more questions of this type but is this sort of thinking simply going too far in trying to make training a fully efficient use of time?

NM XadrezLogico

#13

Very good article! In my own routine, before playing matches or while studying chess, I always try to practice 30 minutes of tactical exercises, including in different modes, such as puzzle storm and puzzle streak, both available on Lichess. In addition, I always try to teach my students the importance of daily tactical training, for pattern recognition, improving calculation, and always accompanying the analysis of positions with a positional assessment. I believe that this is one of the essential keys to progress in chess and to maintaining it when you reach a higher level.

IM HelloItsDmitri

#14

@Wodjul said in #12:

There is perhaps no bad way but maybe are there less efficient ways and more efficient ways? I have a quick question.

If I was studying 1-move forks, do you think it would be less efficient to do 70 random forks and more efficient to 10 forks for the pawn and then 10 for each of the six pieces in turn, to get a better repetition reinforcement effect? If training was to be even more efficient, then perhaps we would need data on how common each type of fork was in games above say 2000 rating and then proportion study by those frequencies. If knight forks are 30% of all forks in live play above 2000 (just a guess) then the training in knight forks should be 30% of all fork training, perhaps?

I have more questions of this type but is this sort of thinking simply going too far in trying to make training a fully efficient use of time?

I will be honest with you, as I do not like people that pretend to know everything - I have no idea. My knowledge on the subject does not reach such depths, and my suspicion is that it would be a fun scientific study-like idea to pursue but, for now, I think it is best to acknowledge that, more often than not, being worried with 100% efficiency just stops us from actually doing the work. As they say, its better to do something than nothing, and sometimes being on an infinite pursue for the optimal setup prevents us from doing that.

@Wodjul said in #12: > There is perhaps no bad way but maybe are there less efficient ways and more efficient ways? I have a quick question. > > If I was studying 1-move forks, do you think it would be less efficient to do 70 random forks and more efficient to 10 forks for the pawn and then 10 for each of the six pieces in turn, to get a better repetition reinforcement effect? If training was to be even more efficient, then perhaps we would need data on how common each type of fork was in games above say 2000 rating and then proportion study by those frequencies. If knight forks are 30% of all forks in live play above 2000 (just a guess) then the training in knight forks should be 30% of all fork training, perhaps? > > I have more questions of this type but is this sort of thinking simply going too far in trying to make training a fully efficient use of time? I will be honest with you, as I do not like people that pretend to know everything - I have no idea. My knowledge on the subject does not reach such depths, and my suspicion is that it would be a fun scientific study-like idea to pursue but, for now, I think it is best to acknowledge that, more often than not, being worried with 100% efficiency just stops us from actually doing the work. As they say, its better to do something than nothing, and sometimes being on an infinite pursue for the optimal setup prevents us from doing that.