Temposchlucker

I was pretty succesful with the different bishop endings. Usually I could calculate everything fairly well. But a knight is a whole different animal. I have done way more than my fair share in knight exercises. Microdrills, Troyis,  Maurice Ashley etc.. Yet I always had the feeling that my knight vision, especially in time trouble, is still way below what it should be. Today I started with KNpKN exercises and boy, do I score badly! To get an impression: It turns out that this endgame is a perfect knight vision exercise. Try your hand yourself: Black to move. You can find the problem here . [FEN "8/8/2N5/8/p7/1knK4/8/8 b - - 1 1"] It can take a while before I can win those endgames. But to take an advance on it I exercised the KQKN endgame which will usually be the ensueing endgame. In the beginning I sucked, but after some time I developed the vision that is needed to bring this endgame to a good end.

When I wrote the previous post and read the comments on it, I finally realized that I actually have enough knowledge to play the KRpKR ending well. I know the boundaries of my knowledge, that is to say I have a clear idea about what I don't know yet, but it takes another 2 weeks to master that. There is no reason at all to flick in that study now. At the moment I know the technique how to promote a non rook pawn and how to promote a rook pawn. I know how the Lucena, Philidor and Vancura relate. That's why I started with bishop endgames yesterday. It turns out that I have much more feeling for that endgame allready. I have a clear idea what I'm after, and the amount of techniques are already known. Most of them that is. I score very well at CT, and find those problems much easier to calculate. So hopefully tomorrow I can move on with knight endgames. I can't wait.

To write the previous post helped me to get from the passive "Troyis-mode" into the active thinking mode. With Troyis mode I mean the way how I learned to play Troyis. On autopilot by just doing. The lazy Troyis mode only works if the amount of themes you are trying to master is low. Preferably one. Otherwise the density of problems of the same theme is too low and you forget before you repeat. So things look "new" for eternity. The creation of the previous post forced me to create a framework. When I now encounter a problem, I can relate it to the framework. Which helps to let it stick long enough until the next repetition of a problem of similar theme. I could find the general strategy (Lucena) for a KRpKR ending in Troyis mode in about two weeks. The strategy to promote a rook pawn can't be found in Troyis mode, since it is too complicated and counter intuitive. I looked at a few video's and now the general idea of promoting the rook pawn is c...

Because of my latest post mr. Z accused me of lazyness. He sensed that quite right. In stead of taking the lead and working out the themes, I immersed myself in a trial and error approach towards this endgame. In that way, I make the same errors over and over again due to the lack of "density" of problems belonging to the same theme. That leads to a lack of focus due to the fact there are too many themes to handle at the same time. In this post I will try to correct that. On another note, it is not so easy to learn this endgame since the computer feels obliged to play perfect moves. A perfect move is the move that delays the mate the most. But humans can throw less optimal moves at you which you still have to know how to defuse them. Which can't be exercised against a computer with perfect play. The amount of moves to mate is quite irrelevant. What's important is that the methods you develop are easy to remember. No matter the amount of moves, as long it provides a ...

It took me two weeks to get the hang of the general technique to promote a pawn in a KRpKR ending, starting at the 4th rank. I can now play it a tempo. I developed a technique that isn't necessarily the fastest, but a sure way to escort a pawn to promotion. The necessary positions for this standard method cannot always be reached though. Now I'm trying to get the hang of the exceptions. That's no easy task. I made a problemset with my failures at CT and wrote down the themes of the exceptions. It is impossible to find all positions, since there are too many exceptions. For instance there are 209 different positions of mutual zugzwang alone. But the amount of themes they can be classified by seems to be fairly limited. Sofar I have found the following themes: Rook pawns need a totally different approach. Sometimes there is a mate at the edge of the board. Both the attacker and the defender can fall victim. Beware of skewers and pins. King against rook, king against pawn...

I haven't repeated my tactical problemset with 2000+ rated problems since nov, 22th. I used to have a succesrate of 100% and a solving time under 40 seconds for every problem by then. Yesterday I repeated the set with a succesrate of 91%. But the average solving time has increased to about 100 seconds. Which means that it takes much more time to recall the solution. This seems to indicate that forgetting is preceded by an increase of recalling time. Or maybe we simply call "forgotten" the state where the recalling time has become unacceptable. The balance of recalling by move sequence vs. by geometrical pattern (diagram) is changed from 60:40 to 50:50. This might be a weak indication that move sequences are more prone to fading a way in memory than geometrical patterns. I will train with the set untill the succesrate is 100% and the maximum solving time is below 40 seconds again.

I finished the first 6 chapters of Silman's complete endgame course. That doesn't mean that I master all the positions that are needed already. But the reading made it possible to map the endgame problems of CT on the chapters of the course. CT gives a clue about the frequence of occurrence of certain endgames. When I combine the information of SCEC and CT I get to the following list of material distribution: KpK KpKp KQKp KRpKR currently in progress KBpKp KNpKp KppKp KRKp KppKQ The topics in italics are already known pretty well by me. At the moment I'm working on KRpKR endings. The order of the list is based on the frequency of occurrence in actual games in combination with growing complexity. At the moment I'm reading chapter 7. The first 6 chapters provide the building blocks. The 7th chapter is meant for expert level. Although it is a bit above my head, it indicates whereto matters are going to develop. In chapter 7 you learn to decompose a complicate...

Now my games become longer, I get rook endings on the board. To understand rook endings, the first task is to get the hang of KRp vs KR endings. The Philidor- and Lucena positions form only the beginning of this type of ending. One has to learn to make progress when the pawn is still on the fourth rank all the way to promotion. To learn that from books is a very daunting task. Virtualy impossible. I made a filter at CT with 1600 KRpKR problems. That should be enough to get the hang of it. I have done the first 100 problems, and I scored 70%. Of course that must become 100% before I can move on to the next stage. What emerges is a general method to make progress towards promotion. Already I approach that systematically. But every now and then, there is a special element in the position that prevents such straight forward approach. At first sight it can appear to be rather esoteric. By doing a lot of exercises, I hope to discover some similarities in the positions that aren't st...