Incorrect Rating for Tactics Problem?

Sort:
Avatar of Ziryab
philidor_position wrote:
Ziryab wrote:
(...)

White had sixty seconds or less left. Here and later, he played the moves differently than I would have. He created a passed pawn on the g-file instantly, while I would have maneuvered my king to the queenside first. We disagreed on which constituted the "simplest win". Engines calculate thousands or even millions of positions and favor the win requiring the fewest moves. "Simple" to the human relies upon recognition of "patterns," but may differ between players of approximate equal strength.


I would do the same thing (create the passed pawn). It's like, things are more concrete and visible when you have a lone pawn instead of just a majority.

And maybe move the king to the black pawns with the e4 route.


A majority is a passed pawn, and the threat is often more powerful.

The position seemed a bit more complex last night. I've reversed colors, and may have other things wrong too. One of the players is a member here, I'll see if I can get him to weigh in.

 

I repect your effort to get a high score percentage, but it will be far more meaningful when you have completed over 4000 TT problems, as I have, than when you've completed barely over 100.

Avatar of philidorposition
Ziryab wrote:
philidor_position wrote:
Ziryab wrote:
(...)

White had sixty seconds or less left. Here and later, he played the moves differently than I would have. He created a passed pawn on the g-file instantly, while I would have maneuvered my king to the queenside first. We disagreed on which constituted the "simplest win". Engines calculate thousands or even millions of positions and favor the win requiring the fewest moves. "Simple" to the human relies upon recognition of "patterns," but may differ between players of approximate equal strength.


I would do the same thing (create the passed pawn). It's like, things are more concrete and visible when you have a lone pawn instead of just a majority.

And maybe move the king to the black pawns with the e4 route.


A majority is a passed pawn, and the threat is often more powerful.

The position seemed a bit more complex last night. I've reversed colors, and may have other things wrong too. One of the players is a member here, I'll see if I can get him to weigh in.

 

I repect your effort to get a high score percentage, but it will be far more meaningful when you have completed over 4000 TT problems, as I have, than when you've completed barely over 100.


I have around 17K problems in my pocket in other online servers Smile. I tried TT here mainly to check it out, it took a little time adjusting to it, and it's very fast for me so I realized I won't be very good at it anyway. Still I don't think I did too bad, I was expecting worse.

In CTS, which I have abandoned a long time ago, it's literally 90 with a very low rating with around 14000 problems. In the current one (chesstempo), I'm currently solving at 90% in blitz mode but I've started low so it's still 86, that rating still pretty low with 3500+ problems. In standard mode, I just can't get it up to 80% with a pretty high rating as the problems are too damn difficult for that. Smile

Avatar of Ziryab
philidor_position wrote:
Ziryab wrote:
philidor_position wrote:
Ziryab wrote:
(...)

White had sixty seconds or less left. Here and later, he played the moves differently than I would have. He created a passed pawn on the g-file instantly, while I would have maneuvered my king to the queenside first. We disagreed on which constituted the "simplest win". Engines calculate thousands or even millions of positions and favor the win requiring the fewest moves. "Simple" to the human relies upon recognition of "patterns," but may differ between players of approximate equal strength.


I would do the same thing (create the passed pawn). It's like, things are more concrete and visible when you have a lone pawn instead of just a majority.

And maybe move the king to the black pawns with the e4 route.


A majority is a passed pawn, and the threat is often more powerful.

The position seemed a bit more complex last night. I've reversed colors, and may have other things wrong too. One of the players is a member here, I'll see if I can get him to weigh in.

 

I repect your effort to get a high score percentage, but it will be far more meaningful when you have completed over 4000 TT problems, as I have, than when you've completed barely over 100.


I have around 17K problems in my pocket in other online servers . I tried TT here mainly to check it out, it took a little time adjusting to it, and it's very fast for me so I realized I won't be very good at it anyway. Still I don't think I did too bad, I was expecting worse.

In CTS, which I have abandoned a long time ago, it's literally 90 with a very low rating with around 14000 problems. In the current one (chesstempo), I'm currently solving at 90% in blitz mode but I've started low so it's still 86, that rating still pretty low with 3500+ problems. In standard mode, I just can't get it up to 80% with a pretty high rating as the problems are too damn difficult for that.


CTS is much faster than TT here, and in my limited experience (a few hundred there) rely much more on instant recognition of patterns. Far more here require long sequences of calculations.

I prefer to use databases on my computer and not play for rating, score, or percentage. Presently I'm focused more on the Anthology of Chess Combinations and the CI Solver's kit. I could use testing mode, which tracks my score, but I prefer quick access to the correct solution when I get one wrong.

Avatar of philidorposition
Ziryab wrote:

CTS is much faster than TT here, and in my limited experience (a few hundred there) rely much more on instant recognition of patterns. Far more here require long sequences of calculations.

I prefer to use databases on my computer and not play for rating, score, or percentage. Presently I'm focused more on the Anthology of Chess Combinations and the CI Solver's kit. I could use testing mode, which tracks my score, but I prefer quick access to the correct solution when I get one wrong.


That's one of the reasons I stopped training on CTS.

Speaking for myself, having ratings, an interactive community, comments etc make the training experience much more enjoyable. I believe ratings in various platforms have a very important role in motivating me for improvement. I find it OK to care a lot about ratings, but many (you too, I guess) have different opinions about this.

I sometimes look at Dvoretsky's tactics book while trying to sleep, but find it difficult to show the discipline to actually solve the problems there like I solve rated problems online. Usually I reach quick conclusions and just have a peek at the solutions.

Avatar of Ziryab
philidor_position wrote:
Ziryab wrote:

CTS is much faster than TT here, and in my limited experience (a few hundred there) rely much more on instant recognition of patterns. Far more here require long sequences of calculations.

I prefer to use databases on my computer and not play for rating, score, or percentage. Presently I'm focused more on the Anthology of Chess Combinations and the CI Solver's kit. I could use testing mode, which tracks my score, but I prefer quick access to the correct solution when I get one wrong.


That's one of the reasons I stopped training on CTS.

Speaking for myself, having ratings, an interactive community, comments etc make the training experience much more enjoyable. I believe ratings in various platforms have a very important role in motivating me for improvement. I find it OK to care a lot about ratings, but many (you too, I guess) have different opinions about this.

I sometimes look at Dvoretsky's tactics book while trying to sleep, but find it difficult to show the discipline to actually solve the problems there like I solve rated problems online. Usually I reach quick conclusions and just have a peek at the solutions.


I use ratings to set goals, and measure progress. This use is as true for TT here as for my USCF rating, and my many internet ratings. But, my tactics training regimen does not require constant rating feedback; rather, the correct solution to incorrectly solved problems take precedence. I prefer long combinations that lead to a clear positional edge (such as those in Chess Informant). But, I think it is important to vary my training, including training for quick checkmates, and even elementary exercises that I learned more than thirty years ago, but practice for speed and precision. For example, checkmate with a queen and king against a lone queen requires a maximum of ten moves. I practice doing so consistently in under twelve moves (optimally ten), and in less than eight seconds. I've done it in five seconds with a physical board and Chronos touch clock. A mouse is faster, of course.

Avatar of philidorposition

That's a very logical approach I think, I hadn't taken into account various types of training, tactics training isn't limited to the problem sets and the format of online servers of course.