A little test

Above test games were played with time control 40 moves in 40 minutes, on Quad HP9300 with 3 cores, no pondering. Note that Houdini.2 was a modified version, whereby I had improved some values with the Houdini-config tool. And with the Houdini-2 engine, I used my ChessPartner (CP) userbook (about 16Mb), whereby with the new Houdini-3 engine, I used the 'Perfect2012' book by Sedat Canbaz version for ChessPartner(CP). Although these 'perfect' books by Canbaz already are quite good, with previous tests with more games, I concluded my book is about 50 points better. Nb the worst books, like the default CP book is worse, at least 50 rating points worse than 'perfect'.
So whats' the fuzz about new engine releases ? 50 points rating improvement isnt that much, it can also be achieved with better books, hardware and so on. But yes you can use all together as well obviously if you're really fanatic.. And better engines -like Houdini3-certainly help me to improve my opening books. So now eagerly waiting for the Komodo5-mp release; end 2013 maybe ?? :)
(will they catch up with Houdini ?)
And when i get some more time i'll see if i can compare my books with the strong Hiarcs14 book, which is in Chessbase (.ctg) format. First step is to compare the 'Perfect' with the Hiarcs14 book, if Hiarcs also is about 50 points better then i will conclude my CP book is equally good as the Hiarcs book. But maybe later i'll see if i can make a .ctg book myself.

Houdini hype, new theory ?

Recently a new update of the top-engine Houdini-3 was released, and i have analyzed the endnodes of my medium-sized Bookbuilder book with this engine. Subsequently backsolved, and the new book can be downloaded as part of the trial-package for Bookbuilder at www.superchess.com
See link to download, and then -on new page- download 'demo'-version which is fully functioning program for a period of 14 days. Whereby I updated many booklines, i don't claim this is the ultimate theory now, but for intermediate players it gives a good repertoire. For the rest did quite some work on the E-book Chess opening theory, renewed, an innovative approach, for all levels!

Playing with 1.d4 Nf6 2.c4 e6 against the Queens-Indian with 3.Nf3 b6 (3...d5 4.Nc3 also is good for White) 4.g3! (main line) leads to strong play for White, with my recent analysis; but Black will be able to equalize, if he/she knows precisely what to do. Which will be described in detail in our E-book about opening theory. So the Catalan with 3.g3 also remains a good choice for advanced players, and we found many novelties which are specified in the latest version of our book. Black can also deviate with 2...c6! going for the Slav or 2...g6 (Gruenfeld), but the 'best' lines found with computer analysis also are given in the E-book.

For those who register for the Bookbuilder program the latest version of the E-book will be sent. With lots of novelties, years ahead of current published theory!
Until end of 2012 with a reduction, thereafter i probably release a Bookbuilder-pro package in addition to the normal Bookbuilder, which will contain my ultrastrong handtuned Chess-Partner opening book, which contains years of work and has been checked against other theoretical works such as the Hiarcs14 book, several GM repertoire (paper)books from Quality Chess, and so on. But the price for the 'pro version will be higher, obviously Good luck with your chess,
jef

PS obviously only with minimax you don't get a really 'playable' computer chess book with statistics, but i have used the info about the best lines to develop a book for the Chess Partner interface, and there the book choice is made by a combination of learned score, statistics, and times played. PS coming weeks I am going to check/improve the book by analyzing (EPD) nodes in the book and then importing them again via the special book correction module in Bookbuilder3.6 (adding the position after one move more, and then backsolving the whole lot again.