12 April, 2011

Endgame Adventures In Team League 4545 : Where we ended in T47

Partly, I wrote this post to offset against this one on T47 openings ( I wouldn't want to be classified as an openings freak :)  and to continue my exploration of T47, but also because I was interested in just what sort of endings are played in TL games ( especially compared to 'professionals') and maybe just a bit on trying to identify what endings I should concentrate on, to shortcut trying to work on them all.

Presentation was an issue, since, although endgames are classified, they aren't as straightforward as openings ( ie not recorded in the game notation ), but I used a 'standard' classifier in my database, so at least I present some sort of consistency.

Straight into things.

Below is a table of endings, showing type, games occurring and percentages, with an additional column given for the 2010 Gibraltar Tournament ( 8th GibTelCom Masters ) to compare with TL games ( it was a relatively random choice ).

It should be noted that one game can fall into many categories, since what begins, for example, as a Rook ending, may, via exchanges, switch to a pawn only ending. It's not, therefore, that one game has only one possible ending. Confusing possibly, but I think its correct

Type Description Games T47 % GibTelCom2010%
I Pawn Endings 294 22 10
II Rook Endings 374 28 45
III Bishop Endings 162 12 11
IV Knight Endings 127 9.5 9
V Bishop vs Knight Endings 124 9 11
VI Rook vs Bishop Endings 68 5 5
VII Rook vs Knight Endings 29 2 2
VIII Queen vs Pawn Endings 46 3.5 2
IX Queen vs Queen Endings 47 3.4 2
X Queen vs Rook Endings 19 1.5 < 1
XI Queen vs Knight Endings 10 1 < 1
XII Queen vs Bishop Endings 8 1 < 1

Immediately, two things are noteworthy to me.

One is that Rook Endings are the most likely to arise in both TL and Professional play : they are the highest percentage. The second is that in TL play, Pawn endings are almost as likely to occur as Rook endings ( 28% for Rooks and 22% for Pawns), something that is definitely not true for Professionals, where 45% of endings are Rook-based.

My guess would be that Professionals actually know their Rook endings, so that they resolve the game in that phase and don't need to exchange down to pawn-only endings. This probably applies to other endings too, and must be a major point differentiating the two camps.

How about looking at a few "truisms"...

Are opposite bishop endings drawn  ? Well, in T47 the statistics aren't conclusive, but imply that it could be true-ish ! 22 games of this type and 10 are drawn.

Here's one that is definitely the case...
52. Kb3
 After 52 moves, there isn't much movement left.

But here..
I think I would play on. After 55. Kxg5 Kxg7, I think there are some possibilities for the Black King to make the cross-board journey and support his pawns, especially if White gives in to the temptation of the King-side Black pawns, and lets face it, we amateurs are often pawn-greedy !

Are Queen and pawns vs Queen and pawns a draw ? It seems to be true. In T47, 16 of those, and 8 are drawn, including this one, proof that just because the most powerful pieces are left doesn't mean that a result is guaranteed.
41. Qe4...Draw agreed

In all sorts of endings, there are lots of examples of agreed draws that should be wins, but I think a lot of these are down to time issues. A relatively short time frame of 45/45 means that what could be the most interesting, thought-provoking, and indeed most "able-to-be-worked-out" part of the game, is reduced to a scramble, and in such occasions draws in won positions are often agreed or actually are the result of not working out the correct move to play in the 10 seconds available, and making the drawing, rather than winning move : its understandable !

Success in endgames other than the obvious and/or basic, is an area where the better prepared and practised player will benefit, much more so than openings, where mistakes are often recoverable. This is as true in Master play as in amateur.

But, since we (amateurs) don't know our endings in detail, we don't know how to win particular ones, but we do know how to win the basic one, namely King and pawns, hence we exchange to the minimum possible, and fight it out there.

In endgames, you can never get away from specifics, so finally, as a "bad" example of the supposed “rule” of simplifying when ahead to win, what happened here....?

White played b5, and immediately transformed a likely win,to a draw, apparently forgetting his basic K + P endgame theory, since after Qxc5, Kxc5, Kc7, the Black king is in front of the white pawn and will prevent it queening !

I started out by suggesting that I wanted to try to identify the most common endings and see if I could short cut a way to prepare for them.

This isn't possible : there is not a short cut.

Certainly, I have identified that knowing Rook endings could provide benefits [ see this as a starter ] and if you know your pawn endings as well, together these cover 50% of T47 endings.

However, really it goes to show that in chess, be it professional or amateur, there are few short cuts : experience, practice and hard work are the only ways to be prepared and improve.

Time to break out the endgame books !  


ScotchYeti said...

Interesting. There is a good statistical overview in the book from Mueller+Lamprecht, based on the Megabase 2001. The most common endgames were rook+minor piece vs rook + minor piece (15%) followed by rook vs rook (8%).

I find it quite difficult to study rook endgames. The basic positions are clear but when it comes to manoeuvering in almost equal positions you have to spend a lot of time to study them. As a happy patzer I rather do another round of tactics. In the last 10 games I had only 3 endgames so the motivation is not really high to do more.

In case you want to study rook endings, look at games from Rubinstein:
http://www.chessgames.com/perl/chesscollection?cid=1009298 or http://www.chessgames.com/perl/chesscollection?cid=1005118

Signalman said...

I will "doff my hat" to M & L as being far more accurate, since they probably invested more time in it than me !

It is difficult to assess since you really need to go through all games, rather than use a classifier, at least to do it correctly. My only defence is that I used the same classifier to do the comparison.

All I can do/see is that not only do we amateurs not know our endgames too well, but we are limited in exactly the ones we do know !

An excellent point that only 30% of your games reached an ending...doesn't really inspire one to make an effort on them, and I too find them hard. Its much nicer to finish off in the middle game !

Rubinstein is also highlighted for his rook endings by Marin in his "Learn from the Champions" book.