opening books

[james uselton]

Sorry, but I have a novice computer question. Opening books must vary in quality. Some are stronger than others and one must be the best of all.
Why dont people use the best opening book whatever that book is? By the way, Which opening book is the best?

[Graham Banks]

Sorry, but I have a novice computer question. Opening books must vary in quality. Some are stronger than others and one must be the best of all.
Why dont people use the best opening book whatever that book is? By the way, Which opening book is the best?

Depends entirely on what you want to use it for.

[Bill Rogers]

Another thing to consider is that not all opening books are free or even for sale. A programmer might make a very strong book that enables his program to win lots of games. In that case he might not want to share at all.
I am in the process of making my own opening books by using moves made by champions only. I will not have any draw games in my data base only winners some for white and some of black.
Bill

[Mark Mason]

James,

There are a number of good books publically available - you can download most of them from:

http://www.frayerchess.com/

Also there is Sedat Canbaz Perfect 14 just released :

http://scjewellery.com/Perfect14_book.7z

However, as Bill has said the strongest books are the ones people have made themselves, usually investing hundreds of hours in analysing and hand tuning. These books are private and usually the authors do not choose to share them.

Mark

[Anil]

I would like to add that the book which is considered to be the strongest today may be overthrown by another book tomorrow. Hence, the fluctuation in the strongest book category.

Some people may create opening books specifically to play against other strong opening books after a lot of analysis and tuning. The book may not be the strongest overall, but it may be able to defeat the strongest available book (which it is tuned to defeat).

[Sedat Canbaz]

James,

However, as Bill has said the strongest books are the ones people have made themselves, usually investing hundreds of hours in analysing and hand tuning. These books are private and usually the authors do not choose to share them.

Hello friends,

Here is available a Book Ranking by Klaus Ritter:
http://chessstars.de/pageID_5104739.html

So what are the names of these users or books ? Where can check the results of these super strong private books...on Playches or still
the results are still not published ?

I have a comparison ranking Playchess table,where i selected only the Rybka 3 x64bit against Rybka 3 x64bit games of the users of Playchess server,played min 100 games per user on this august.Especaially between 15.08-23.08 2008,Rybka3 using Perfect13x/14 book has a record:
182 wins,149 draws,only 23 losses ( +72,46 % ) of 354 games played (note:only against Rybka3 engines)


And now,we can not expect the same performance,just becouse now mostly of the users of Playchess have the Perfect14 book,in other words the mystery is solved...The strongest secretly moves lines will not perform as before,there is also possibility that many users will use the Perfect14 book and the ratings will be balanced...so i can not promise that the Playchess's rating of each user of Perfect14 book will be at 64% winpersantage as before


I am pleased to announce,the Perfect13x/14 book has the best performance for august month on Playchess sever:

Code: Select all

Rank PlaychessUser      Score/Games        +

01 SedatChess ----------420.0/662 ------ 63.45%
02 Formula-Open --------360.5/577 ------ 62.5%
03 Bushel --------------175.0/285 ------ 61.42%
04 Big Rabbit ----------650.5/1072 ----- 60.97%
05 Mistral33 -----------68.0/112 --------60.71%
06 Lickenzendickens---- 148.0/244 -------60.67%
07 PortCitySlim ------- 89.0/147 --------60.56%
08 Major Tom ---------- 643.5/1067 ----- 60.31%
09 FindecanoElanese ----162.5/272 -------59.88%
10 Devintam Danguje --- 324.0/548 -------59.17%
11 Balabachi -----------83.0/141  -------59.17%
12 Zombik ------------- 441/751----------58.82%
13 Morula ------------- 95.5/165 --------58.13%
14 Seven of 9 ----------92.0/160 --------58.0%
15 DR3 Esquire ---------131.5/230 -------57.47%
16 Pacificrabbit -------311/550 ---------56.81%
17 Moheevi_chess -------368.5/651------- 56.81%
18 Phil Harris2 --------966/1712 --------56.49%
19 Phil Harris ---------1130/2011 -------56.21%
20 Everyone ------------141.0/251 ------ 56.17%
21 Nijinsky ------------439.5/781 -------56.0%
22 Neysene -------------380/681 ---------55.86%
23 Sandwichman ---------126/230 ---------54.94%
24 EIVIssa -------------388.5/712 ------ 54.64%
25 CompMaster ----------380.5/695 -------54.49%
26 Coleus --------------78.5/145 --------54.34%
27 Betty Boop ----------88.5/163 --------54.34%
28 Ulta-d --------------85.0/158 --------54.06%
29 Takker --------------338/626 ---------54.05%
30 Thamover ----------- 768/1428 --------54.05%
31 Guest42 ------------ 434.0/803 -------54.05%
32 Haagshart ---------- 655.0/121 -------54.05%
33 Coffindodger --------73.5/136 --------54.05%
34 KIngsDead -----------87.0/161 --------54.05%
35 Pakman --------------76.5/142 --------54.04%
36 N 1001---------------184.5/344 -------53.76%
37 Bookmaker -----------528/1005 --------52.63%
38 Demchuks ------------92.5/177 --------52.35%
39 Acid Moon -----------193.5/378 -------51.28%
40 ABT tuning ----------285.5/562 -------51.02%
41 Iron Ball -----------248.5/492 -------50.76%
42 Juantamad -----------188.0/371 -------50.76%
43 Slent mode ----------556.5/1107 ------50.50%
44 Lasker#77------------93.0/187 --------49.75%
45 Bommetje ------------159.5/321 -------49.75%
46 Vousatej ------------482.5/975 -------49.50%
47 Agent_Smith9999 -----202/410 ---------49.50%
48 OlegGR --------------85.0/173 --------49.26%
49 Eight to eight ------172.0/353 -------48.75%
50 Max -----------------54.5/116 --------47.5%
51 Seppi ---------------480/1026 --------46.94%
52 Jernjoffen ----------275.5/593 -------46.51%
53 SIR JMA -------------171.0/378 -------45.45% 




And here is my private book test:



                             Perfect's Opening Book Rating:
              ----------------------------------------------------------




    Program                          Elo    +   -   Games   Score   Av.Op.  Draws

  1 Perfect 14                     : 3324   20  20   560    64.3 %   3222   49.6 %
  2 Perfect 13x                    : 3295   20  20   490    61.1 %   3216   58.2 %
  3 Rybka3                         : 3250   20  20   520    56.2 %   3207   55.0 %
  4 Compmaster 2b                  : 3241   20  19   580    55.0 %   3206   52.4 %
  5 Magister 2                     : 3228   24  24   420    51.9 %   3215   48.6 %
  6 My Book 9.1                    : 3218   24  24   360    49.9 %   3219   55.3 %
  7 Perfect 13                     : 3217   30  31   200    42.5 %   3270   60.0 %
  8 Salvo 4                        : 3215   28  28   290    49.8 %   3216   52.8 %
  9 Salvo 4.1                      : 3213   30  30   230    50.4 %   3210   54.8 %
 10 My Book 8.1                    : 3212   31  31   220    53.9 %   3185   55.0 %
 11 Ceko 2.0                       : 3207   20  20   490    52.1 %   3192   56.5 %
 12 Ghost 14.2p                    : 3196   26  26   280    50.4 %   3194   60.0 %
 13 Takker2008                     : 3194   21  21   490    50.1 %   3193   52.4 %
 14 Sicilian2                      : 3189   32  32   200    49.0 %   3196   57.0 %
 15 Ghost 14.2t                    : 3186   26  26   330    47.7 %   3202   51.2 %
 16 Salvo 3                        : 3185   27  27   300    49.7 %   3187   51.3 %
 17 SicilianExtreme                : 3181   24  24   390    43.3 %   3228   51.8 %
 18 Magister                       : 3180   26  26   320    48.8 %   3189   55.0 %
 19 Playchess                      : 3176   21  21   490    47.4 %   3194   52.4 %
 20 Ghost 14.2k                    : 3175   29  29   220    49.1 %   3181   60.0 %
 21 My Book 9                      : 3175   24  24   340    47.8 %   3190   58.5 %
 22 Awesome                        : 3156   21  21   470    45.2 %   3189   55.5 %
 23 HS-for-Rybka                   : 3136   25  25   350    42.0 %   3192   53.1 %
 24 RybkaII                        : 3133   31  31   220    42.5 %   3185   54.1 %
 25 Zappa Perfect 14               : 3113   39  39   150    46.0 %   3141   50.7 %
 26 Hiarcs12 Perfect 14            : 3084   41  41   150    43.3 %   3131   46.7 %
 27 Zappa Perfect 13               : 3023   43  44   150    30.0 %   3170   41.3 %
 28 Hiarcs12 H12MPBook             : 2994   47  48   150    27.7 %   3161   34.0 %


==================================================================================

SCCT (4m+2s) Book Ratinglist:

Rybka 3 x64 using Perfect 14.ctg    : 3324 elo
Rybka 3 x64 using Perfect 13.ctg    : 3217  elo
*Difference:107 elo

Zappa Mexico X64 using Perfect 14.ctg  : 3113 elo
Zappa Mexico X64 using Perfect 13.ctg  : 3023 elo
*Difference:90 elo


Hiarcs 12 using Perfect 14.ctg  : 3084 elo
Hiarcs 12 using own H12MPBook   : 2994 elo
*Difference:90 elo

My recommendations on improving Perfect14 book to get high rating on Playchess sever:
1.Use Rybka 3 x64bit with 128 MB Hastablesize for 3 min - 5 min time controls
2.You need fast Hardware,at least quad-core processor,8 core...are will be very good choise
3.Disable loosing weak lines with red colour
4.Daily update the book lines or at least in a week with new and stronger moves
5.If the lost game's lines seems strong,then analyse the same lost position With Rybka 3 engine at
slower time control to find stronger move and add that new move to the book,
but dont forget to enable the new moove with green colour
6.And then,go to the Edit/Openings book menu and "Reset weights"

And here are the games played by Rybka3 using Perfect13x/14 book on Playchess sever
http://scjewellery.com/Perfect_R3x64.rar

Note:some games are played M.Servet Kultur using Perfect13x book

Best Wishess,
Serdat Canbaz

[Marc Lacrosse]

(...)
I am in the process of making my own opening books by using moves made by champions only. I will not have any draw games in my data base only winners some for white and some of black.
Bill

Hi Bill.

I suppose that like many people you decided to discard draw games for making the book hoping to get a more lively opening repertoire.

Please do have a look at this recent post of mine on rybkachess :

ML on rybka chess forum 27.08.2008

Hi Nick,

There is a very interesting paradox in this particular topic (keeping or discarding drawn games for book building).

I am sure that many people hope to get a more fighting, less drawish book if they discard drawn games and only keep decided ones for building their book.

This is completely erroneous !

In fact it will happen that the book built without drawn games will favor the most drawish lines in many cases !

Let's see with a little example. Imagine a position with four resonable continuations for which we have 400 games (100 games with move A, 100 with move B and so on)

Imagine that full results of these four continuations are (from the side-to-play point-of-view):

line A : +55 = 0 -45
line B : +30 =50 -20
line C : +20 =70 -10
line D : +10 =90 -00

Although the proportion of draws varies very much, all four continuations get a similar overall result : 55%.

If we take the full set of 400 games, we will have equal probabilities for all four moves.

Now let's see what happens if we discard drawn games and build our book from decided games :

In our selected set of decided games we see :
line A : 55% success (55/100; no game has been discarded)
line B : 60% success (30/50; 50 drawn games have been discarded)
line C : 66% success (20/30; 70 drawn games have been discarded)
line D : 100 % success (10/10; 90 drawn games have been discarded)

As most book heuristics strongly favor the more successful lines, move D will be the preferred move in this book, followed by C, B and A.

If we go back to the full set of games we clearly see that having discarded draws we will favor the most drawish line .
Line D which will be the most played move in our book had 90% draws, C had 70% and so on!

Nice, isn't it ?

Marc

PS I fear that quite a few book-makers are not aware of this simple fact.

[Rubinus]

My scrapp is only remis shortly 20 moves - e.g. exchange french and the like or settle games. And all games less than 15 moves.
My advice is book max. 30 ply - next Rybka 3 is better than igm 2600 - rarity for ECO C88-99, B97 and the like, probably.
1x IGM mistake:
Sax - Psachis, Aruba 1992
1.e4 e6 2.d4 d5 3.Nc3 Nf6 4.Bg5 de4 5.Nxe4 Nbd7 6.Nf3 h6 7.Nxf6+ Nxf6 8.Bh4 Be7 (better is c5) 9.Bc4 c5 10.Qe2 cd4 11.Nxd4?? Qa5?? (after Qxd4 12.Bb5+ Kf8! black to win) 12.c3 -- 1/2

Book for Playchess is other league, but e.g. CompMaster have 550MB and very narrow repertoir, ****.

I will widest variability.

[Ferdy]

(...)
I am in the process of making my own opening books by using moves made by champions only. I will not have any draw games in my data base only winners some for white and some of black.
Bill

Hi Bill.

I suppose that like many people you decided to discard draw games for making the book hoping to get a more lively opening repertoire.

Please do have a look at this recent post of mine on rybkachess :

ML on rybka chess forum 27.08.2008

Hi Nick,

There is a very interesting paradox in this particular topic (keeping or discarding drawn games for book building).

I am sure that many people hope to get a more fighting, less drawish book if they discard drawn games and only keep decided ones for building their book.

This is completely erroneous !

In fact it will happen that the book built without drawn games will favor the most drawish lines in many cases !

Let's see with a little example. Imagine a position with four resonable continuations for which we have 400 games (100 games with move A, 100 with move B and so on)

Imagine that full results of these four continuations are (from the side-to-play point-of-view):

line A : +55 = 0 -45
line B : +30 =50 -20
line C : +20 =70 -10
line D : +10 =90 -00

Although the proportion of draws varies very much, all four continuations get a similar overall result : 55%.

If we take the full set of 400 games, we will have equal probabilities for all four moves.

Now let's see what happens if we discard drawn games and build our book from decided games :

In our selected set of decided games we see :
line A : 55% success (55/100; no game has been discarded)
line B : 60% success (30/50; 50 drawn games have been discarded)
line C : 66% success (20/30; 70 drawn games have been discarded)
line D : 100 % success (10/10; 90 drawn games have been discarded)

As most book heuristics strongly favor the more successful lines, move D will be the preferred move in this book, followed by C, B and A.

If we go back to the full set of games we clearly see that having discarded draws we will favor the most drawish line .
Line D which will be the most played move in our book had 90% draws, C had 70% and so on!

Nice, isn't it ?

Marc

PS I fear that quite a few book-makers are not aware of this simple fact.

Hi Marc,

From your example above, what I will do is to extract only won lines.
Total wins = 55+30+20+10 = 115. Now our new success rate for each line will be:

line A : 55/115 (47.8%)
line B : 30/115
line C : 20/115
line D : 10/115 (8.7%)

So line A is favorable.

Say for example from starting position, without considering traspositions.
e4 : 55/115
d4 : 30/115
c4 : 20/115
Nf3 : 10/115

Best regards,
Ferdy