Tuesday, October 31, 2006

CT-Art, focus on the king

I'm pretty sure I'm learning from the CT-Art exercises, but they seem very artificial to me. I don't recall exactly, but it seems every problem is based on the king in some way, either checkmating or using checks to win material. In contrast, CTS problems are often just winning some material based on a tactic like a pin or fork without anything to do with the king.

The repeated theme of working against the king narrows the search enough that sometimes the problems are too easy. In some cases, the opponent has a mate threat so you know the answer has to be a check and there's only one check! Even if the check means sacrificing the queen, you can be sure it's right and then calculating the line is not so difficult.

Another problem I'm having is that it never marks problem 461 has having been solved. Despite the fact that I solved it the first time it came up and I've tried going back and solving it again. Anyone else ever have this difficulty?

CT-Art stats:
  • level 10: 97% (110 problems)
  • level 20: 92% (286 problems)
  • level 30: 88% (144 done, 77 to go)
  • elo: 2167

Labels: ,

Thursday, October 26, 2006

Tactics while tired

Today I had an interesting experience doing CTS tactics. I was doing problems there even though I was a bit tired. At the time I thought "Is this really affecting me?" and I honestly wasn't sure. After resting a bit, taking a nap, and refocusing, I did some more problems later in the day. Thanks to the reappearance of the ratings graph for the 200 most recent problems solved, I can show you the result:

The bumpy region from 0 to about 135 that hovers between roughly 1642 and 1646 was my tired solving session. My post rest, refreshed, ready to go session starts between 130 and 140 and the rating takes off, getting to 1660 in about 65 problems.

I felt a clear difference when I started my second session. It felt like my eyes were far more active, moving more quickly and finding/seeing the pieces better. I don't know what this tells me about chess improvement, other than I probably wasn't learning anything while tired. Additionally, this indicates that tired is not a good condition to be playing in.

I'm heading out of town for the weekend, so there will be no Friday update, here are the current CT-Art stats:

  • level 10: 97% (110 problems)
  • level 20: 92% (286 problems)
  • level 30: 87% (110 done, 111 to go)
  • elo: 2149

Labels: ,

Tuesday, October 24, 2006

100 Done (level 30, CT-Art)

I just finished 100 problems in level 30. So I'm taking a break to read from my chess books. Here are the CT-Art stats so far:
  • level 10: 97% (110 problems)
  • level 20: 92% (286 problems)
  • level 30: 86% (100 done, 121 remain)
  • elo: 2138

As I predicted, the level 30 percentage has come down a bit. Some level 30 problems are simple, but most require a lot of calculation. There are some really clever tactics on this level. At least, it seems that way having never seen them. Hopefully once I learn them cold they will become simple as well.

I'm happy enough with my rate of solving so far. I'm hoping to finish the 1209 problems by the end of the year. I've done 496 problems so that leaves me 713 to do. I have a bout 9.5 weeks 'til the end of the year, but I'll be traveling this weekend as well as Thanksgiving (late November) and Christmas (end of December), so I probably need to stick with about 100 problems per week. In fact, it's probably best to do a little more than that while I'm still on the lower levels.

I've noticed quite a growing number of people, including even Temposchlucker now, who have decided to do CTS problems slowly. I think it's important to do tactics at all speeds. Lots of problems in rapid succession, tough problems requiring 1-5 minutes or sometimes more, and deep analysis of a single position (Stoyko exercise) or game annotation. For now I'll just say that I think CTS problems are best suited for the rapid succession category because of the way they are rated. I may say more about this later if I can muster enough energy to make my thoughts coherent.

My last tournament is now rated, just a minor rating change for me: 1702 -> 1707

Labels: ,

Saturday, October 21, 2006

Friday Update

As usual, the weekly Friday update. Since I played in the 2-day tourney last weekend, I spent quite a bit of time analyzing and annotating my games. I did get back to the tactics training by the end of the week. At CTS my rating is now at 1641 after 16515 tries. That's about where it was last week. I don't mind the CTS rating holding steady while I delve into other chess activities. I do expect that I need to work on more ingrained pattern recognition in this department though.

In CT-Art I finished level 20 and have just begun level 30. Level 30 threw me for a loop at first, but once I started to focus better I began getting the problems with a healthy struggle. CT-Art stats:
  • level 10: 97% (110 problems)
  • level 20: 92% (286 problems)
  • level 30: 93% (13 problems done, 208 to go)
  • elo: 2076
I expect the numbers to decrease later in level 30 as the problems tend to get harder as the levels go on.

I have also done the first 10 problems from The Reassess Your Chess Workbook. Silman gives a lot of text for the solution of each diagram along with a list of the important Imbalances and Ideas. While his biggest point is to make sure we understand the important characteristics, the solutions must necessarily involve playing the tactics correctly, so it's fairly well rounded.

I've read the introduction to Stean's Simple Chess and I'm almost finished the section on Outposts. I'm not sure what impact this book will have on my game, so I'll reserve any judgements until much later. I do have high hopes for it.

Labels: , ,

Wednesday, October 18, 2006

CT-Art level 20 finished

Tonight I polished off the last of the level 20 problems on CTS. These things got tough at the end! Well, perhaps they just employed ideas I'd not seen before. I'll have a keener eye for these tactics in the future.

The stats:
  • level 10: 97% (110 problems)
  • level 20: 92% (286 problems)
  • elo: 2056

I had hoped to finish level 20 by the end of last week, but fell just a little short. Then I spent a lot of time going analyzing and annotating my games from the weekend tournament. I'm not worried too much about my schedule, but I hope to complete all 1209 problems by the end of the year.


Tournament Report: Moderate Success

In 4 games I had 1 win, 1 draw, and 2 losses. Everyone I played was 100 - 150 points higher rated than me, so this is not a terrible result. For my own ego, I'll show a position from the only game I won:
White to play
Opponent, 1805Loomis
White is to play and win. The answer can be found along with the full game at: www.chesslog.de/users/loomis/?game=538
The other games from the tournament are also posted at chesslog.de. They are games 537 through 540.
Overall I feel good about the tournament. Playing higher rated players and analyzing the games has shown me what some of my flaws are. Now I'll have to focus on correcting them. And of course, back to the tactics exercises!

Labels: , , ,

Friday, October 13, 2006

Friday Update

My rating at CTS took another down-turn this week. I am now rated 1640 after 16008 tries. The fact that my CTS rating has dropped each of the last two weeks coincides with starting the Ct-Art exercises. Of course, this may or may not be related. My first thought is that this indicates that CTS is a skill of its own that must be kept sharp. I would still like to see the CTS rating going up, as that indicates that I am solving more difficult problems, but it may go up more slowly if I am also focused on other chess things. That will have to be fine.

CT-Art statistics:
  • level 10: 97% (110 problems)
  • level 20: 92% (267 done, 19 to go)
  • elo: 2061

I didn't finish level 20 by the end of this week as I thought I would. Many of the problems later in the level took quite a bit longer for me to solve. This is good news, I am happy to see that this problem set will be quite challenging for me. I did get close to finishing and should wrap up level 20 shortly.

The varied difficulty in level 20 makes me question the general intelligence of the folks at Convekta. I always thought it was silly that the levels were numbered 10, 20, 30, etc. What was wrong with 1, 2, 3? Now that I see there are 286 problems of varied difficulty in level 20, I think, couldn't they have broken this into levels 20, 21, 22, etc.? Well, I don't think it affects my chess improvement so I'm not worried, but it is a head scratcher.

I've started concurrently on Stean's Simple Chess and Silman's Reassess Your Chess Workbook. Basically fitting them in where I have time between tactics study. I'll try to write something more organized about this next week.

Tournament: 2-day tournament starts tomorrow!

Labels: ,

Wednesday, October 11, 2006

Catching up with comments

First, I wanted to thank Likesforests and Samurai Pawn for wishing me good luck in my tournament. I appreciate the blogger community for their support -- it's nice to know the long road of chess improvement is travelled in good company. I also appreciate the blogger community for all the ideas shared amongst the various blogs.

A second thanks to Samurai for suggesting Nunn's puzzle book as a final exam after CT-Art. I'll definitely look into that one. It would be nice to get the computer to do it just so that keeping score is easier.

mini update:
CT-Art statistics
  • level 10: 97% (110 problems)
  • level 20: 92% (227 done, 59 to go)
  • elo: 2049
I hope to have the last 59 problems of level 20 done by my end of the week update.


Monday, October 09, 2006

Tournament Report: Doing better

I scored 1.5/3 at this weekend's tournament. But it was quite a confidence boosting 1.5.

In my first game I played a master rated 2290. He dubiously sacked a pawn in the opening and hung a pawn trying to get active play against my king in the middle game. Then he missed a tactic while trying to win back one of the pawns and I was able to win a queen for rook and bishop. So I was up a queen and pawn for rook and bishop. I was slowly making progress towards the win, but when my time got down to about a minute I started to play inaccurately and eventually hung enough material to just be losing. Here is the position where I get a real winning advantage:
White to Play
Black has just played Rf8-f6. This move tactically defends the e-pawn. Can you find the continuation after Qxe4? I have posted the whole game at chesslog.de (http://www.chesslog.de/users/loomis/?game=522) so you can see the solution there. That site is still a little buggy, but I believe the author is working on it -- It's a personal volunteer project, so we should just say "thanks" for what we can get.

In my second game I played a class B player rated about 1680. I felt like he didn't really understand the opening he played and allowed me to dominate the center without really trying to undermine it. He then helped me get a rapid kingside expansion and his position was soon overrun. I was surprised to win a game without feeling any real counter threats from someone of this rating, but I think it came more from his misunderstanding of the position than anything special I did. Eventually I'll put the whole game on chesslog.de, but for now, here are some key positions:

Black to Play


I have just played 14. c4 and my opponent replied 14. ... Bxf3. This move is also Fritz's first choice at a depth of 14, but I believe my opponent and Fritz are both misevaluating the position (Fritz evaluates it as = with a tiny edge to white, .25). There is the obvious tactic that the bishop is now guarding the c-pawn so I have recapture on f3 with the g-pawn. This results in doubled f-pawns and the open g-file leading to my king. Usually these positional factors would count against me, but here they shouldn't count at all because my control of the center means these weaknesses will never be attacked.

The game continued: 15. gxf3 b6 16. f4 Nc6 17. e5. All of these moves follow Fritz's top choice, but they are straightforward so this is not too surprising. Here is the current position:

White clearly has a superior occupation of the center. The black knight and bishop still have their eyes on key squares and if he can shake up the pawns could break through, but I think white has a serious advantage here. On this move, though, I think black makes a real error: 17. ... g5. This attempt to undermine the white center backfires quickly in the game: 18. fxg5 Bxe5 19. f4 Bg7 20. f5 Be5 21. Bd3. This is one of those moves that I'm lucky to have available. The bishop simultaneously i) clears the queens path to defense of the h2 pawn, ii) unpins the d5 pawn creating a threat on the c6 knight, and iii) attacks the weak h7 square. Even at this time I am envisioning Bxh7 sacrifices.

The position after the multipurpose Be2-d3. The continuation was 21. ... Nd4 22. f6 e6 23. Qg2! and now it's all over because the queen is going to the h-file. 23. ... exd5 24. Bxh7+ It turns out Qh3 was simpler, but I had this sacrifice on my brain for a while and calculated it all out. If Kxh7 then Qh3+ Kg8 Qh6 puts the king in a box (Ne6 is the only defense to immediate mate) and the g6 push is on the way. Black trusted me and without too much thought played 24. ... Kf8 and after 25. g6 he resigned. Final position:

I don't want this description of this game to sound like too much braging, but this is as well as I've played in a long time, or possibly ever.

Rating change from tournament: 1683 -> 1702 (back over 1700!)

addition: I forgot to mention my third game which was a draw against an 1870. The only thing of note about this game is that at this point in the day I wasn't focused properly on the game. This is something I need to overcome. I played poorly out of the opening and was lucky to get a better position a few moves later. Not feeling able to focus on playing up to my ability I offered a draw which was acceptable.

Labels: , , ,

Friday, October 06, 2006

Friday update

Time for the weekly update. First, the CT-Art stats:
  • level 10: 97% (110 problems)
  • level 20: 92% (183 problems)
  • elo: 2029

The level 20 percentage is already coming down. Most of level 20 is still relatively easy (a fair bit of mate in 2), but many are complicated enough to make me calculate. Given that the higher levels should be even tougher, it looks like this will be good for making me work hard. With another 103 problems left on level 20, I should finish in the middle of next week.

This week was a down week at Chess Tactics Server (CTS). I finished with a rating of 1650 (high 1672) after 15388 tries. At the end of last week and the beginning of this week when I made my high of 1672 I was focused on seeing all my pieces and their lines. This works very well for me, but unlike other things I have worked on, I can't seem to get it into a subconcious (auto) process.

Another thing I worked on this week at CTS is finding undefended or underdefended pieces. As the saying goes "Loose pieces drop off," so I figure if you can find the loose pieces, then it's just a matter of figuring out how to get them. A lot of times this is the order in which I recognize a pattern, first I see a weakness then I see how my pieces get to it. Unfortunately, to do this efficiently, I have to already know where my pieces are and where they can go. So I need to get both these things in my auto-process, but right now I seem to only do one at a time. I'm certainly not doing both efficiently.

Upcoming tournaments: One day, 3 rounds at G/75 tomorrow. Two day, 5 rounds, next week.

Labels: ,

Ct-Art Exercise 252

I am through exercise 252 in CT-Art, stopping there because I spent far too long on that problem. The solution I calculated actually did win a piece or the exchange, but the correct solution won a queen for a rook. It's a little disappointing that the program doesn't count winning a piece as correct, but the requirement to be perfect will be good for getting better.

Here's the problem in question:
Black to move

My solution was. 1. ... Qe4+, CT-Art gave 2. f3 as a refutation. My planned response to this move was 2. ... Nxd2 3. fxe4 Rc4 where I am attacking the rook and the bishop.

The better solution is 1. Rg3+ fxg3 2. Qe4+ and now black can take the queen free and clear.

CT-Art statistics:

  • level 10: 97% (110 problems)
  • level 20: 94% (142 problems)
  • elo: 2031

It will probably take me through the end of next week to finish level 20. Tomorrow's regular weekly update will include some comments on this weeks work at CTS.

Labels: , , ,

Monday, October 02, 2006

CT-Art level 20

Just a quick update tonight. I did the first 50 problems in level 20. I was a little distracted and didn't get to do them all in one sitting, but I think it took a little under an hour, or about a minute per problem. Most problems were still pretty straightforward, but some really made me think. After 50 problems I'm at 95% success for level 20, slightly less than my level 10 score. There are 286 problems in level 20, so I'll probably be doing these for the rest of the week.

CT-Art Results:
  • level 10: 97% (110 problems)
  • level 20: 95% (50 problems)
  • elo: 1956


The next beginning

I am now in possession of Michael Stean's Simple Chess and the famous tactical training software CT-Art. I have finished section 1 of Winning Pawn Structures and began to read Stean's book today. It's a short book, but I think there is a lot to be learned from it. I also began the long journey with CT-Art.

CT-Art isn't exactly user friendly. I didn't know what to do first upon firing up the program. Should I be solving problems in practice mode or test mode? I solved all of the level 10 problems today in practice mode. I just did them in numerical order, though it seems that this order groups by theme, meaning that there would be a few problems in a row with the same kind of mate, or the same kind of knight fork, or some other motif. I was close, but not quite perfect for level 10. My final percentage was 97% for the 110 problems in level 10.

The level 10 problems were a lot like the Chess Tactics Server (CTS) problems I've been doing lately and typically didn't take more than 10 seconds to see the solution. Needless to say, this isn't helping my calculation skills yet, but my understanding is that the problems get tougher in the next few levels. I'm sure my calculation and tactical acumen will be pushed to (or beyond) the limit soon enough.

Since the large majority of my time has been at CTS, I'm excited to have some other activities to put significant energy towards. Balance is one of the keys to improvement. Simple Chess will add much needed positional ideas and CT-Art will help with calculation skill (at least once I get to a level requiring more calculation).