Games #6 and #7

Game #6 was the 2nd game I've played on ICC, and quite possibly the best game of chess I've ever played.  Heck, it might be the best game of chess I'll ever play, but I hope not.  It featured, for me, lots of tactical variations to consider in a pretty game where there were lots of losing or drawing moves to be made for both sides.

Game #7 was a weekly slow game played via the DHLC on chess.com.  Another game where I get into time trouble, trade down into a winning endgame and make some inaccuracies turning the win into a possible draw.

On the advice of Laurent, I will try to draw more specific conclusions about my play.

1. I need to look 5-ply ahead at a minimum every time, no exceptions.  Too many times I am making what are good moves, but not considering his potential threats.
2. I'm still having some issues with time trouble.  With all the time I'm using, you'd think I was spending it looking at least 5-ply ahead and doing concrete analysis, but I'm not.  Hopefully when I do my DeGroot exercise with Dan, he can suggest some improvements.
3. I am moving too fast when in a clearly winning game.  Rule #1 of winning a won game is to think defense first.  In game #6 I had 20 different ways to win, but didn't make the 1 move that would've taken away all hope for a save.  I still won the game, but that won't happen every time.  I should definitely know better here.
4. I feel like my tactics and endgame are at least where they need to be if not ahead of where they need to be for my level.  My ability to evaluate a position is probably my weak link right now.  I plan to read Heisman's "Elements of Positional Evaluation" and Shereshevsky's "Endgame Strategy".  The former should help with the early middle game, while the latter should help with the late middle game.

My current stats against class players.

Class E: +3 -1 =0

Class D: +2 -0 =0

Class C: +1 -2 =1

Class B: +1 -1 =0

Still very early to draw any conclusions.  Chess.com released the results of a survey where members were asked to post any chess.com or OTB ratings they had which were current IE >= 25 games played and >= 5 games played in the last 6 months.  Online standard players have a rating ~100 points higher than their USCF rating up until around 1300. Once their online rating gets to ~1600, it's ~100 points lower than their USCF rating.  And around high Class A/expert range, the difference is 200+.

Anyone who said that online ratings are lower or higher than OTB ratings were right. :)

My class D and E stats could be combined into +5 -1 =0 and it would be more like USCF Class E players.

Chess Hero, where have you been all my life?

In a previous blog post about John Bain's Chess Tactics for students, I mentioned I was making a secondary set for diagrams I had the most trouble with.  I loaded a PGN file with just those diagrams into Chess Hero to see how it would compare to Anki.

Overall, I think this will be much faster than using Anki.

1)  I can very easily see my completion times for each problem.  Just gathering the data at the end of a big set usually takes up enough time to annoy me.

2) I don't have to worry about using Windows snipping tool to create diagrams to load into Anki.

3) The solutions are automatically computer checked.  You load an engine into Chess Hero and each problem automatically has the solution built in.  No need to go through and add the solutions to the backs of my flashcards then worry about errata.

Downsides:

1) I have to have a pgn of whatever tactic book I'm working with, and I'll likely need to make a profile for each set that I create out of that master pgn so I can control exactly what problems I get. That's probably still faster overall if the pgn exists on the internet somewhere.  I wasn't able to find an electronic copy of the Woolum book or a pgn file, so I had to make my own pgn file to load into chessbase to have diagrams to snip into Anki.  Don't ask me how much time that took; it was a lot.  Electronic copies and pgn copies of Bain exist, so that was very little inconvenience.

If this software were modified slightly to account for spaced repetitions, it would be perfect.  I know I could just go with something like CT-Art or Chess Tempo and not have to worry about that, but I guess I'm being stubborn about doing tactical problems from specific books and working my way up through those.

Anyway, it's certainly software that any chess improver should look into to see if it fits into their study plan.

3 more games of Chess

Game #3 was my 2nd round match in the DHLC Slow Chess League qualifier tournament.  I had a conversation with Dan Heisman about the ratings at chess.com and he feels they are about equal to a true USCF rating.  At 1613, this would be the highest rated opponent I've ever had.  I made it a point to not worry about the ratings difference too much.  I went in with the mentality that I wanted to win this game.  I did some research on my opponent beforehand and saw that he'd meet d4 with either the Dutch or the KID.  Since I know nothing about either, I decided to go with e4, which he'd meet with e5.  I've always wanted to give the Scotch a try, and I did some light reading of The Scotch Game Explained, by Gary Lane.

Game #4 is my 1st ever game on ICC.  According to Heisman, ICC ratings are about 100-150 points higher than their equivalent USCF rating.  I had been doing pretty poorly with clock management up to this point, and I wanted to emphasize averaging no more then 3-4 minutes at the most in certain positions.  Of course, more critical positions would warrant more time, but I should be averaging about 75 seconds per move in a G45/45.  The good thing about ICC is the time stamping feature.  I made lots of moves in this game in 75 seconds or less, devoting more time to positions I thought warranted it.  There's still room for improvement.

Game #5 is my last round match in the DHLC Slow Chess League qualifier tournament.

Some takeaways:

1. I am getting horrible positions out of the opening.  Most of the time I'm still finding a way to win the game.  I guess the opening doesn't matter after all.
2. I am doing okay at staying safe, but not so great at taking advantage of my opponent's mistakes.  My thought process needs lots of improvement, and I think part of it is getting into a defensive mindset in a cramped opening position. 
3. I need to be more aggressive with my break moves and piece placement in the opening.  I'm getting the slow and safe parts down, but active needs work.
4. I'm making enough mistakes in games to still lose to Class E players, but they haven't been taking advantage.  I need to play up more.  I'm actually rated over 1300 at chess.com now.  Following Dan's advance, most of my games need to be against players rated up to 200 points above me to punish me for my mistakes.  I'm not really afraid of losing and I welcome the learning opportunities.
5. I am overconfident in the endgame.  I've made it a primary emphasis, along with tactics and it's paying off some.  But I need to analyze more and do less hand-waving in endgame positions.
6. I still need to speed up some in less critical positions.

I am now 4-1 against Class E players, with three of my wins coming as Black. I've never won 4/5 against Class E players before. That horrible blunder I made in game #3 was hopefully enough to not make that kind of game changing mistake again.

I am 1-0 against Class D players.

I am 0-1 against Class B players.

I'm still considering myself a Class E player until USCF tells me otherwise.  Hopefully I can get several games against higher rated players in the coming weeks.  I have been afraid to play against them since I considered my playing strength to be equal to my USCF 1000 so I stuck with under 1200 opponents.  Time to kick it up a notch.

John Bain - Chess Tactics for Students: Revision 1 Complete

I started my 1st revision of this book on November 25th and finished it January 16th (52 days).  After excluding the one bad problem, the book contains 389 basic tactics, broken down by motif.  Dan Heisman suggests you should be able to complete 85% in less than 15 seconds.

I divided the problems into 6 sections, exactly the same way as Bright Knight in his Bain Experiment.  Each section should be slightly harder than the last.  Instead of presenting each section, broken down by time solved for each repetition, I'm going to show two different graphs:

Percent completed in under 15 seconds for the 1st pass of each section:

Keeping in mind that A1 was hand-timed, thus faster than the other sections, this shows overall improvement.   Given that I generally improved about 20% from pass 1 to pass 2, I suspect A1 would come in around 35% if it were Anki-timed.

Percent completed in under 15 seconds for all passes:

Note: I left out A1 from this chart since I thought it looked better without the hand-timed results.  As I mentioned above, I generally improved about 20% from pass 1 to pass 2, however I improved from 40% to 81.54% from F1 to F2!

I solved at least 85% of the problems in under 15 seconds by the time I got to pass 6 at the latest, except for section B, where I topped out with 76.92% for some strange reason.  My total under 15 seconds solve rate across all sections was 86.3%. I estimate I would be about 88.4% if all were hand-timed.

I have now made a supplementary problem set, made of all the problems that I failed to get in under 15 seconds in at least two of the last three passes.  This comes out to 58 problems. I will be treating these problems like a 7th section and work them for 6 passes.  When I finish, I should be 2 weeks away from doing a 2nd revision of all the problems.  With this supplementary work, I suspect I should do much better than 86.3% after 1 pass of each section.

Percentages from each motif that are in the supplemental section:

 Keeping in mind that "Knight Forks", "Forks" and "Double Threats" are a similar type of problem and "Discovered Checks", "Double Checks" and "Discovered Attacks" are also similar, they each account for 31% of my supplemental problem set.

I'm already working on the Woolum problem set, and am just over 50% completed.  The problems overall in this book are much harder than Bain and I feel pretty confident I won't be getting anywhere near 85% in under 15 seconds.  I think 75% might be doable though.

As a last thought, I definitely need to tone down the amount of books I'm going through at once for repetitions.  I had Bain, Pandolfini, Woolum and Mattson all going at once.  I think I'm going to stick with no more than 2 books at a time as needed.

Just noticed that today marks the one-month anniversary of this blog.  #BirthdayCake

Bruce Pandolfini - Pandolfini's Endgame Course Revision 1 Complete

I started my first revision of this book on November 27th, and finished it on January 10th (44 days).  The book contains 239 technical endgame positions, generally 1 position and explanation per page.

Section 1 features elementary, checkmates, Queen vs Rook and Rook vs Minor Pieces problems.  The highlight of this section is the Philidor Rook + Bishop vs Rook position, which caused me the most trouble.

Section 2 features King and Pawn Endgames.

Section 3 features piece and Pawn endgames and is the largest section of the book.  Lucena, Philidor and Centurini are all here.

I divided the book into 4 sections of 54-58 problems each:

Set A: The non-trivial problems from Section 1, plus the first part of Section 2.
Set B: The remainder of Section 2.
Set C: The first half of Section 3.
Set D: The second half of Section 3.

This book contains LOTS of typos, but there is only one busted position.  Typos aside, I thought this was an excellent book and any beginner should be able to tackle this book and learn a great deal from it.  It more or less covers the same topics as Silman's Complete Endgame Course up through the Class A section.  If you want a workbook to supplement the material in Silman's book, this is a good place to get it.

And here's the final stats in chart form:

Set A:

Initial score of 77.2% with a peak of 91.2% after pass 3.  A slight dip to 87.7% after pass 6.  My struggles with this set was mainly around the multiple Philidor R+B v R problems.  After some supplemental repetitions play against Houdini, I was able to all but eliminate my weakness there.  The problems I got wrong on the last repetition were mainly random problems that I usually was good on, but may have gone a little too fast and overlooked something.  I tried to grade myself harshly on these.

Initial score of 46.6% with a peak of 96.6% after pass 3.  A slight dip to 93.1% after pass 6.  I forgot to record the results for pass 4. I started off very poorly with this assortment of King and Pawn endgames, but the material is obviously pretty easy to learn and retain after pass 1.  I didn't experience any recurring issues with any particular problems in this set.

Initial score of 66.1% with a peak of 100% after pass 5.  A slight dip to 91.4% after pass 6.  No recurring issues with problems in this set.

Initial score of 37% with a peak of 90.7% after pass 5.  A slight dip to 85.2% after pass 6.  I had multiple recurring issues with problems in this set.  Rook and Pawn endgames can be pretty hard.  And even when you get the solution down that Pandolfini gives, Houdini sometimes has other ideas.  I did multiple supplemental repetitions with Houdini here to help me learn the material, which was reflected in the big gain between pass 4 (72.2%) and pass 5 (90.7%).

Overall, I finished with a final score of 89.4% across all sets, which I am pretty happy with.  I didn't time these, and the solutions can sometimes be pretty involved with lots of alternate defenses to work through.  I have scheduled a second revision starting on February 22nd, 44 days after I finished the last repetition of revision 1.  I plan to do 1 or maybe 2 repetitions for each set, depending on my retention.  I hope to still be in the ~90% range.

I'm now in search of a 2nd endgame problems book.  I'm currently working through James Howell's Essential Chess Endings.  Yes, I paid $80 to get this out of print book.  There's probably similarly good books out there, but Silman once listed this as one of his 3 essential endgame books to learn from.  It's similar to Silman's endgame book, but the explanations and examples are more detailed.  It's not much of a problem book, since there's only 48 exercises.

Potential candidates for my next problem book are:

Larry Evans - Chess Endgame Quiz

200 multiple choice problems.  From the Amazon preview they look pretty basic and possibly a step down from Pandolfini.

Bernd Rosen - Chess Endgame Training

This looks like the Endgame version of Bain's Chess Tactics for Students.  Endgames broken down into 15 chapters, with a short introduction to the theme followed by 16 problems.  There's a final comprehensive quiz at end.  There looks to be 258 problems in this book of similar difficulty as Pandolfini with a few harder ones per chapter.

John Hall - Endgame Challenge

451 graded problems.  I can't really tell from the Amazon preview since they mostly showed the easy and moderate problems.  Seems like a contender.

Ger Van Perlo - Endgame Tactics
Considering the awards this book has won, this should be a very good book.  1300 endgame problems and it looks like it might be a good book to read after I finish with Howell.  The new edition doesn't come out until May.  But I was also planning on reading the Shereshevsky book after Howell.  I guess we'll see.

2 more games of chess

Heisman recommends playing at least 2 slow games per week. I haven't kept up with that pace, but I have managed to get a few more 45/45 games in since my last post.

A few notes:

1. I need to do a better job at development
2. I need to do a better job at controlling the center/respecting a strong pawn center

I also have a question:  After move 23 I have 3 minor pieces and a Rook vs 1 minor piece and 2 Rooks.  I'm also about to win a pawn.  I know that generally you want to exchange pieces when up material, but if I could only exchange off one of my pieces for his, would it be better to trade the minor pieces or the Rooks?  IOW, 2 minor pieces + Rook vs 2 Rooks or 3 minor pieces vs 1 minor Piece and 1 Rook.  Does it matter?  Am I overthinking it?  I have to assume my opponent won't want to trade.  I guess my best strategy is to just untangle a bit and start pushing Queenside pawns and force the trades anyway.

After this game, we decided to play another game in a few days with me as Black again so I could get practice in the Grünfeld.  I spent the next few days booking up and I learned something important; I don't like it!  I went over several games in Aagaard's Starting Out: The Grünfeld and each game I felt like I didn't like Black's position.  However, I felt like that could be due to unfamiliarity and decided to go with the Grünfeld for our return game.  Of course, I spent that time going over the Classical and Modern Exchange variations and he avoided that entirely this time...

A few more notes:

1. Play real chess every move!  Include all possible CCT until quiescence
2. Give critical positions the time they deserve

I still don't like the Grünfeld, so it's shelved for now.  I think I need to stay away from hypermodern openings in general.  For now, I'm sticking with the classical Queen's Gambit Declined.  I'm reading Sadler's "Queen's Gambit Declined" book and I find it to be excellent for a beginner.  The format of the book is something like a cross Between Fine's "Ideas Behind the Chess Openings" and Chernev's "Logical Chess Move by Move".  Each chapter covers a variation of the QGD with several annotated games in the format of a hypothetical student asking questions about the moves with Sadler giving the explanations for the moves.  Multiple annotated games are included each chapter to cover the main choices White or Black has at various points.  Between this and Fine's book as a supplement, I feel like I'm getting a good grounding on the ideas and strategies of a particular opening and what I should be looking for.  I've seen the book recommended many times, generally along the lines of  "the theory is outdated but the base knowledge is excellent".  I definitely feel that is true.  Does anyone know of any other books that follow a similar format that are good for a beginner other than Sadler's other books on the Slav and Semi-Slav?