Monday, September 17, 2018

The scientific approach part 2

In my previous article I gave a glimpse of about what is needed to become a world-class-player. This information is very hard to find. Only by an extensive research through the games of former world-topper Vladimir Epishin I was able to discover how enormous his opening-knowledge is. I summarized it by comparing the number of Vladimir's openings with my own's. It were about 25x more and besides Vladimir often was also much more versed of the theory which is something really astonishing.

Though I do work quite a number of hours myself at chess. So an interesting question is how Vladimir succeeds to maintain and remember so many different openings. Well the simple answer is that he only plays a selected number of openings during a period and he only maintains and studies those. So it is definitely not the case that I could encounter all possible 72 positions in our game. However only Vladimir knows which openings in his repertoire are active. I have to guess which means I only play safe by checking everything.

At that time I didn't get enough time for such elaborated study but there exist a couple of rules which help to improve the odds of a preparation. 60% of Vladimir's openings were single experiments. I guess professionals can become bored of the usual traditional stuff and like to spice up from time to time the opening. An extra bonus is that the opponent will be surprised.
So only 12 of the 72 opening-positions were played by Vladimir in more than 5 games. Of course those 12 get a higher priority. Anyway even more important than the frequency is when an opening was played last. Vladimir plays no opening all his career.
80% of the openings Vladimir played, were used maximally 2,5 years. That is a very short period of time which obviously was influenced by the many one-time openings. Eventually we discover that only for a limited amount of openings that their lifespan overlap. I made an overview per year of the number of overlaps to better illustrate this.
Obviously there are less overlaps in the initial and the latest years. The peak is in 2008 with no less than 17 overlaps. The average is 8,2 overlaps. In other words if you would know which 8-10 openings are today overlapping then you could just limit the preparation on those.

For sure the best is to start with the most recently played openings if time is lacking. However it is wrong to assume that by looking solely at the openings played in the last 2,5 years that you have an 80% success-rate to prepare the right opening of the game. Beside the lifespan of an opening, also the frequency of the opening must be taken into account.
So in about 50% of his games this former top-player chooses for brand-new openings of his repertoire. I guess that preparations at his playing-level can be so detailed that it becomes almost mandatory to continuously reinvent yourself. We also see that we need to return 5 years back to get an 80% hit-rate of preparing the right opening of Vladimir's repertoire. The pareto-principle is definitely also valid for chess. Finally we also remark that in about 10% of the games, Vladimir likes to reuse an old love. Vladimir surely knows about the benefits of using old wine in new skins see part 1 and part 2.

So Vladimir cleverly uses all the assets of his arsenal of openings. It is still a lot of work to study and maintain but not something impossible for a professional. Anyway his approach stands diametrically opposed to how I play chess. Scoring is for me less important. On the other hand I can enjoy more the historical aspect of an opening (which probably explains partly why chess960 is still a small niche today despite some serious tries to get the public more excited about it like last by attracting former world-champion Garry Kasparov, see the current ongoing St. Louis-chessfestival).

That means I am for sure not a polygamist as a chess-player. I am a serial monogamist. Besides a few exceptions, I don't play several openings in parallel in the same position. So contrary to Epishin, you will see that the lifespans of my openings are almost never overlapping. If I make the same exercise for my own repertoire as I did for Epishin's repertoire then this becomes clearly visible. I checked each of my games played in the last 5 years how old my chosen openings were.
Contrary to Epshin many of my old games in the database (10 years and even older) can often still be used in the preparation. It is something which some opponents knowing me well, gratefully take advantage of. On the other the fact that I am using in 30% of my games fresh openings, shows that I do work at my openings. Openings which have shown weaknesses, are replaced immediately.

That last aspect was once more explicitly shown in the opening which occurred in the 8th round of Open Gent. It concerns a rather long rare line of the Dutch stonewall which I encountered for the first time in a correspondence-game. We have to go back 20 years in time for that game when I was renting a small student-room at the Paardenmarkt in Antwerpen. I just ended school and started working (besides today I still have the same employer).
[Event "EU/M/1234"] [Site "?"] [Date "1998.??.??"] [Round "?"] [White "Verhoef, H."] [Black "Brabo"] [Result "1/2-1/2"] [PlyCount "99"] [EventDate "1998.??.??"] [WhiteElo ""] [BlackElo ""] [ECO ""] [CurrentPosition "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1"] 1.d4 f5 2.g3 Nf6 3.Bg2 e6 4.c4 d5 5.Nh3 c6 6.Qc2 Bd6 7.Bf4 Be7 8.Nd2 O-O 9.O-O h6 10.Bxb8 Rxb8 11.Nf4 Bd6 { (In this correspondence-game I played for the first time this rare opening. Ever since I kept it in my repertoire despite that it is not so easy for black to get sufficient counter-play. This game can be found in the Ultra-corr database.) } 12.Nd3 b6 13.Nf3 Rf7 14.b4 Ba6 15.c5 Bf8 16.Nfe5 Rc7 17.Rab1 Nd7 18.Rfc1 Nxe5 19.Nxe5 Be7 20.cxb6 Rxb6 21.a3 Bd6 22.Nd3 Bxd3 23.Qxd3 Rcb7 24.Ra1 Qa8 25.Qd2 a5 26.bxa5 Ra6 27.e4 fxe4 28.Bf1 Rxa5 29.Rxc6 Bxa3 30.Bh3 Kh8 31.Bxe6 Rb2 32.Qc3 Rb8 33.Rc5 Rxc5 34.dxc5 Qc6 35.Bxd5 Qxc5 36.Qxc5 Bxc5 37.Bxe4 Rf8 38.Ra2 g5 39.Kg2 Kg7 40.f3 Rd8 41.h4 gxh4 42.gxh4 Rf8 43.Kh3 Rf4 44.h5 Be7 45.Ra7 Kf8 46.Kg2 Bg5 47.Bg6 Kg8 48.Rd7 Kf8 49.Kf2 Kg8 50.Ke2 1/2-1/2

The game is stored today also in Ultracorr-x. An OTB-game of mine with the same opening which managed to get in the standard databases, is the one below against Emmanuel Bricard, at that time still an international master but today he is a grandmaster.
[Event "Open Plancoet 6de ronde"] [Site "?"] [Date "2003.??.??"] [Round "?"] [White "Bricard, E."] [Black "Brabo"] [Result "1/2-1/2"] [ECO "A90"] [WhiteElo "2456"] [BlackElo "2274"] [PlyCount "108"] [EventDate "2003.??.??"] [CurrentPosition "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1"] 1.d4 f5 2.g3 Nf6 3.Bg2 e6 4.c4 d5 5.Nh3 c6 6.O-O Bd6 7.Bf4 Be7 8.Qc2 O-O 9.Nd2 h6 10.Bxb8 Rxb8 11.Nf4 Bd6 { (This game is the only with this opening which was inserted in the normal databases like the mega-database. From then onward any player could prepare against this opening of my repertoire.) } 12.Ng6 Rf7 13.Nf3 Rc7 { (In my then made analysis were Nd7 and b6 recommended as more accurate.) } 14.Rac1 b6 15.Nfe5 Bb7 16.Qa4 a6 17.Qb3 b5 18.c5 Bxe5 19.Nxe5 a5 20.h3 Re7 21.g4 Nd7 22.f4 Nxe5 23.dxe5 g6 24.Rf3 Rg7 25.Rg3 Qe7 26.Qa3 Ra8 27.Kh2 Kh7 28.Rg1 b4 29.Qe3 Rag8 30.Bf3 Ba6 31.a3 b3 32.a4 Bc4 33.h4 Qf8 34.Qf2 Qe7 35.Qd4 Qf8 36.e3 Kh8 37.Kg2 Kh7 38.Kf2 Kh8 39.Be2 Bxe2 40.Kxe2 Qa8 41.Qc3 Qa6+ 42.Kf3 Qa7 43.R3g2 Kh7 44.Rc1 Rb8 45.Qd3 h5 46.gxh5 gxh5 47.Rg5 Rxg5 48.hxg5 Kg6 49.Kg3 Qb7 50.Kh4 Qb4 51.Qd1 Rh8 52.Rc3 Qe4 53.Kh3 d4 54.Qxd4 Qf3+ 1/2-1/2
15 years later the Belgian FM Marc Lacrosse chose for exactly the same line as Emmanuel. Maybe Marc had prepared an improvement which isn't so hard to do. Anyway I deviated first with my improvement and I got very quickly a comfortable position. Probably it is karma for some people but against Emanuel I threw away a half point by proposing a draw in a winning position. This time Marc gave me a half point back by resigning in a drawn position.
[Event "Open Gent 8ste ronde"] [Site "?"] [Date "2018.??.??"] [Round "?"] [White "Lacrosse, M."] [Black "Brabo"] [Result "0-1"] [ECO "A90"] [WhiteElo "2120"] [BlackElo "2310"] [PlyCount "88"] [EventDate "2018.??.??"] [CurrentPosition "1rbq2k1/p4rp1/1ppbpnNp/3p1p2/2PP4/5NP1/PPQ1PPBP/R4RK1 w - - 0 14"] 1.d4 f5 2.g3 Nf6 3.Bg2 e6 4.c4 d5 5.Nh3 c6 6.Qc2 Bd6 7.O-O O-O 8.Bf4 Be7 9.Nd2 h6 10.Bxb8 { (In our previous game played in 2004 Marc chose for the weaker Rac1.) } 10...Rxb8 11.Nf4 Bd6 12.Ng6 Rf7 13.Nf3 b6 { (In 2003 I played the inferior Rf7 against the French IM Emanuel Bricard. Maybe Marc hoped for that move but I still remembered my old analysis.) } 14.Rac1 Rc7 15.c5 { (After the game Marc lamented about this move and he is right. Black has now a comfortable game. Critical is Nfe5.) } 15...bxc5 16.dxc5 Bf8 { (I knew Be7 is better when the rook is already at c7 but I wrongly thought it isn't valid when white's knight is at g6.) } 17.Rfd1 Qe8 18.Nxf8 Qxf8 19.Ne5 Nd7 20.Qc3 Qf6 21.Nxd7?! { (White's last moves were not the very best but only now black gets the advantage. With the ugly looking f4 white could've kept things under control.) } 21...Qxc3 22.Rxc3 Rxd7 23.b3 Kf7?! { (A more concrete approach with e5 or a5 is stronger.) } ( 23...e5!? 24.f4 e4!? 25.Bh3! Rdb7 26.Kf2 a5 27.Rd4 Rb4 28.e3 Kf7 29.Bf1 Ra8 30.h4 ) 24.e3?! { (A more accurate sequence is first Rd4.) } ( 24.Rd4! a5 25.f4 Rb4 26.e3 Rdb7 27.Bf3 a4!? 28.Bd1! axb3 29.Bxb3 e5 30.Rxb4 Rxb4 $13 ) 24...a5 25.Rd4?! { (This is now less strong because of e5. White should've played here Bf1.) } 25...Rb4? { (Again I postpone the concrete e5 for something useful but less strong. In the end I lose opportunities as in chess you need to take sometimes risks.) } 26.f4 Ra7 27.Kf2 Ba6 28.Bf1? { (This logical looking exchange brings troubles for white. Black gets full control about the queen-side and will later miss the bishop for the defense.) } 28...Bxf1 29.Kxf1 Kf6 30.Ke2 g5 31.Kd3 e5?! { (So I do eventually play e5 but as often I don't understand the position. A much better plan is Rab7 to first avoid a3 and next to open a second front with h5.) } 32.fxe5+ Kxe5 33.Rc2 Re7 34.Rf2 Rb5?! { (It is again the same as here once more the concrete h5 is more accurate.) } 35.Rc2 Kf6 36.Rc3? { (Marc also starts to wait and see which isn't a surprise when you are running out of time. Searching counter-play with h4 is stronger.) } 36...Rb4 37.a3 Rxd4+ 38.exd4?! { (The engines consider Kxd4 as more resilient but also after exd4 the win is not easy.) } 38...f4 39.g4 h5 40.h3 hxg4 41.hxg4 Ke6?? { (I felt intuitively that the endgame is won but I couldn't find the winning move as I was playing almost exclusively by increments. The precise Re4 forces the win.) } 42.Rc2 Kd7 43.Re2 Rxe2 { (I knew that the pawn-endgame is a draw but I couldn't find anything better. The engines confirm this evaluation but I still get an unexpected present. ) } 44.Kxe2 a4 { (Here Marc resigned. To be sure that I heard this rightly, I asked him to repeat this. After it I requested the arbiter to record this result. Only then I showed to my bewildered opponent that the final position is a dead-draw. No doubt the lack of time played an important role here but Marc also admitted that he was too pessimistic as often the case.) } ( 44...a4 45.b4 Kc7 46.Kd3 Kb7 47.Kc3 Ka6 48.Kd3 Kb5 49.Kc3 f3 50.Kd3 f2 51.Ke2 Kc4 52.Kxf2 Kxd4 53.Ke2 Kc4 54.Kd2 d4 55.Kc2 d3+ 56.Kd2 Kd4 57.Kd1 Kc4 58.Kd2 Kd4 59.Kd1 Kc4 60.Kd2 $11 ) 0-1
In part 1 I explained how I as amateur can achieve openings with a much richer complexity in my games by using a scientific approach compared to somebody whom prefers to variate a lot so choosing a more creative road. This article to some extent confirms this but also shows another side of this scientific approach. By playing the same opening for several decades you get also a historical ingredient in the games. Call me a stupid nostalgic player but I like to write some history even at the expense of some ratingpoints.

Brabo

No comments:

Post a Comment