Sunday, 1 July 2012

Training Progress Report

I have finished my current round of endgame training with: Pandolfini’s Endgame Course, Albert & Krogius’ Just the Facts!, and Chessimo Endgames 01.

I have now been through Pandolfini three times.  Most of the book is easy, and I can do the simple king and pawn endings in my sleep.  There are some tricky sections at the beginning, notably K+B+N vs. K and K+R+B vs. R, and I am still having difficulty with a few of the rook endings.  (K+B+N vs. K is rare in practice, but the winning technique is straightforward to learn.  K+R+B vs. R is also rare, and difficult even for GMs, according to Albert & Krogius.  Rook endings are the most common of all endings, and therefore much more important than either of these.)  I did the first two passes through Pandolfini in odd moments, and the third pass took me three days.  I am due for another revision after about two months.

I have been through the core of Albert & Krogius four times.  This book is harder than Pandolfini, so I needed an extra pass through.  I covered the first and the last two chapters less thoroughly than the others.  The first chapter is rather elementary, and the last two have some very complicated examples with rather thin annotation.  The fourth pass took me 2 weeks (2*5 days).  I did all the exercises on the fourth pass.  They are mainly tactical, rather than tests of endgame knowledge.  I found some of the exercises very hard, but got the majority completely right.  This book is also due for another revision after about two months.

Revising Chessimo was easier than I had expected, and I completed the task in a week (5 days).  The next revision is due after about 4 months.  Chessimo let me select each Unit in turn, and go through the first twenty exercises with no trouble.  (When I first returned to Chessimo it asked: “You have not been training for while, would you like to do a quick review,” but I preferred to drive the process manually.)  I feel a bit more positive about Chessimo, having found that the revision is easy.  Revising Albert & Krogius was much harder, but I expect that harder is better here.

I believe that the three types of training complement each other well.  Pandolfini provides drilling in the basic stuff, and all the examples are easy enough to treat as problems. Albert & Krogius provides wide coverage of the easy and the harder stuff, but with relatively few examples of each type.  Chessimo Endgames 01 gives hands on practice on 480 examples, but only tests one variation within each exercise.  I also did a fair number of endgame exercises when I revised Ivashchenko's Chess School 1b, and will no doubt revisit Coakley's Winning Chess Exercises for Kids, which also includes endgame exercises, at some stage.  I am beginning to tire of endgames, so it is time to move on.  I have plenty of good endgame books for future study, notably:

Pandolfini’s Chess Challenges, 111 Winning Endgames.
Chernev’s Practical Chess Endings
Averbach’s Chess Endings Essential Knowledge
Mednis’s Practical Endgame Lessons
Griffiths’ The Endings
Keres’ Practical Chess Endings

I believe that my endgame training has been worthwhile.  I have done a good job of learning the basic endings, and the main winning methods for the more complex ones. My ability to visualise and calculate simplified positions has also improved.  Unfortunately, my board vision for middle game positions has gone to pot!  I can easily remedy that though. I currently have two background tasks: working through Weteschnik’s Understanding Chess Tactics (yet again), and solving some problems of the day:

My next project is O’Kelly’s Improve Your Chess Fast.  This is an old favourite of mine.  It is O’Kelly’s get you to 2000 correspondence course in book form.  It comprises a series of lessons which provide well annotated games (and some endgames), and some good but badly checked problems at the end of each lesson.


I have solved the (solvable) problems many times before, and worked my way though the whole book a few times over the decades.  When I worked in Europe as a young man, I took a copy of this book with me, along with a copy of Modern Chess Openings and a pocket chess set.  Despite not having played chess for over a decade, I was soon playing at about the 2000 level.  The book, probably did not do what it said on the back cover for many of its readers, but it seemed to work for me!  I am a lot older now, but perhaps it still has some magic left.  Second hand copies are available for pennies.  Perhaps it needed a story about secret Belgian training methods that would have trained a string of World Champions, if only they had not been so secret!


  1. I have had the B+N ending (defending and attacking combined) about fifteen times in slightly more than 50,000 games. I stopped teaching as part of my regular curriculum after reading Silman's Complete Endgame Course. If one wants to learn this checkmate, however, Pandolfini's text is the way to go. In one of the roughly fifteen games, I ended with these two pieces due to underpromotion. It was a training game against Chessmaster.

    I've seen O'Kelley's text in bookstores, before Amazon drove most of the good ones out of business, but never owned a copy. On your recommendation, I'll be more attentive the next time it comes across my path.

  2. I still can not understain the point ,the main goal of ur site is design and prove optimal ways for chess training...them u go back to Okelly book,to pandolfinii later to some griffini,someone even propose some silman...what is this?
    There are so many aspect of chess training that are not mentioned here,what about the objetivity
    What about the motivation,the idle time( main idea in botvinnik methods ) the UCIs,and so many many other aspects,ur brilliant math mind that should focus on the main aspects is distracted by so many little things,is a crime
    Ur mind should be challenged by something like this: could be built a system training that ensure a 2000 elo players boost his level to 2400 in 1 year?could i do that? Yeah exactlysimulating the progress of players like kasparov or carlsen,as long as players have brains ,the rest is just the system training to put pressure on it,and results will come...

  3. There are lots of training methods out there, but no proof for any of them, except:

    Study lots of games.
    Solve lots of problems.
    Play lots of games.

    Perhaps we can add read lots of books. Do we also need to be young and talented? It is relatively easy to get to the level of the average club player, 1800 say, depending on the rating system and where we live. Beyond that, for adults, we have no real evidence. Actually, there is evidence, but it is negative. Most adult players train and gradually get worse!

    I read chess books, and reviews are easy to write and bring in lots of traffic. Some of those people take an interest in the other stuff.

  4. PART 1:

    I think chess and making progress (improve the level of play) is much more complex than this (short advice) my friend.

    (1) Study lots of games, (2) Solve lots of problems and (3) Play lots of games... of course these are the (necessary?!) elements to achieve success at chess. Yes, I agree, but please notice there are more important (detailed and various) questions like:

    a) how many games to study? (a day, week, month, year)
    b) what kind of games? (type, length, strenght of the players, date of play, own games or others)
    c) what does it mean to study? (review, replay, repeat, memorize, analyse)
    d) how to study the games? (with a chessboard/screen, against a computer, friend, teacher, masters, alone)

    a) how many problems? (a day, week, month, year) to solve
    a) what kind of problems to solve? (type, level of difficulty, date of publication, number of moves deep)
    c) what does it mean to solve? (see the correct/best move, see all the variations, see the possible refutations, check out all the candidate moves, reject the worse moves)
    d) how to solve the problems? (with a chessboard/screen, against a computer, friend, teacher, masters, alone)

    a) how many games should we play? (a day, week, month, year)
    b) what kind of games? (type, length, strenght of the players, on the board games, virtual/Internet games)
    c) what does it mean to play games? (play for fun, always do your best, try out the ideas, play always against stronger/weaker opponents, play only against humans/enginges, play virtual/Internet games)
    d) how to play the games? (with a chessboard/screen, against a computer, friend, teacher, masters, at odds, writing down the moves, used time)

    I hope you see how difficult problem it is to IMPROVE your chess. There are many traps and pitfalls on the road to mastery. These questions are just to point out that WITHOUT very solid fundamentals we would NOT be able to improve above 1800. Or put it another way: our efforts would be wasted much more than the expected (and possible) results.


  5. PART 2:

    I noticed that most amateurs (I mean 85-90% of interested population!) have stopped improving when they reach 1800-1900 rating points. And yes - you are absolutelly right as you said: " Beyond that, for adults, we have no real evidence. Actually, there is evidence, but it is negative. Most adult players train and gradually get worse!"

    The more I think about "doing chess (tactical) puzzles" - the more useless I see this way of making improvement. Of course I do not say you have to stop doing it or quit it. However I STRONGLY recommend to rethink the system we (you) use.

    Your (I mean general group of players reading a lot of books and tactics) efforts are extremally low-efficient my friends. It reminds me digging a really BIG hole with the use of a tea-spoon! No matter if you repeat the "digging process" 10.000 or 50.000 times - you would do probably 5-10% of the efficient work you could do if you have changed your approach and "hire/build a big digging machine".

    The diminishing results ("to train and gradually get worse!") do not come from doing more puzzles, even if it could be explained like that. In my opinion not-efficient system of training KILLS all the efforts and being exhausted causes the effect of "being worse when training more" simple relationship.

    I am ABSOLUTELLY (yes, 100% of certainity) convinced that with optimal system of work (training and playing) most players (including adults) could have reached AT LEAST about 2100-2500 level of play (instead of 1800-1900) with proper plan and its implementation. I do not say it is "easy or fast way" to mastery, but just try to imagine how could you calculate the trivial math task (below):


    .... in NO MORE than 10-12 seconds.

    Most people need for solving that task about 45-60 seconds - and they do not always get the correct (final) score.

    Do you get what I mean my dear friends? Yes, I mean your efforts (scores) are about 4-5 times worse than they should be (not at math, but at chess - as we dissuss it here).

  6. 1800-1900 is a little above average for experienced adult players:

    What you are saying is that most people do not rise much above the average which is correct, but tells us nothing in itself.

    AoxomoxoA studied the Chess Tempo statistics and concluded that no adult player had significantly improved beyond the first few thousand problems. If he is right, it is not a case of finding the most efficient tactics training method - they all fail.

    Finding the optimal training method would be good, but first we need to find one that works for adult non-beginners.

    There is hope. Blackburne in the nineteenth century, and John Shaw more recently.

  7. There are many training methods - some of them are better then others.

    Adult players rarely break 1900-1950 rating (level) because they do not train efficient and/or they do not always accept and obey some (strict) rules.

    For example: one of the simplest sin is not to check out the move before (!) it is played. Breaking 1900-barrier requires not hard work, but after all consisent play (work) according to some strict rules. If you obey the rules - you are much closer to get oustide of amateur "ceiling" (that most adult players cannot break down - most often 1800-1900 rating pool).

    IF I list some rules - those guarantee the better results - would it be useful? Or should I obey to the rules and break 2000 barrier by myself - to convince anyone that even ADULTS may be able to get better - but only if they obey the rules. I am close to 40 and now I am just in the range of 1920-1950 (FICS rating).

    Please tell me what should be done/written to convince some of our friends (especially from this blog) - to believe that breaking 1850-1900 barrier is rather a matter of ATTITUDE and not very hard work. Of course without working on chess - improvement might not be possible. However I am sure that efficient work allows to get outside of amateur "ceiling".

    I would be glad to hear your comments and opinions.

  8. What would most convince people would be a randomised controlled trail. Pick a random sample of players from the same rating pool. Get them to do X. Leave the others alone. Find that the players who did X improved by an average of Y rating points relative to the others (i.e. the control group). This is not practical. Even if you offered to pay the selected players large amounts of money, some would still refuse.

    Next best is a non-randomised controlled trial. Get some volunteers from the same rating pool. Get them to do X. Leave the others alone. Find that the players who did X improved by an average of Y rating points relative to the others. This is not as good, because the players who do X are a self selecting group. They may be different in some way from the others, e.g. more motivated.

    Weteschnik coached a team of players and they went up to leagues, according to his book, but perhaps they would have done that anyway. His book is said to be an improved version of his course material. Its a good book. I have worked my way through it several times, but I do not think that it made a significant difference to my chess skills. Would it work if I gave a coarse based on the book to a local club team. I doubt it. Weteschink had a short burst of performance and earned an FM title, but he is not currently a particularly strong player. It would be more convincing if we had independent witnesses and full details, and he was able to repeat his results. It would be even more convincing if others got the same results with the same course.