|after every gauntlet||7||28%|
|after a new book enters||0||0%|
|Weekly - Monday||2||8%|
|Weekly - Sunday||10||40%|
|Every 2 weeks||2||8%|
|Every Month - starting 1st July||1||4%|
|Every Month - starting 25th July (=1 month after started)||0||0%|
|Do not upload at all||3||12%|
More days in the top three: Poker 1.3
More days in the top five: Poker 1.3, Bertran Blitz 1.0, Yograj Y2.4 V1.1
P.S. Could You please upload the games weekly on Sunday ;)
most first places: Yograj (3times)
most ranks under Top 2: Yograj, Poker (3 times)
most ranks under Top 3: Poker (5 times)
most ranks under Top 4: Poker, Bertran Blitz (5 times)
most ranks under Top 5: Poker, Bertran Blitz (5 times)
most first places in a row: Yograj (3times)
most 2nd places: Poker (2)
most 3rd places: Titielmago, Poker (2)
most 4th places: Bertran Blitz (2)
most 5th places: Bombook (2)
biggest win against an opponent inside top 5: Yograj (10-0 vs Mybook15)
highest total score: Yograj(29/40 = 72,5%!!)
highest score against a leader: Titielmago with 5,5 - 4,5 (vs Poker) and Poker with 5,5 - 4,5 (vs Bertran Blitz)
best average rank: Yograj (2) (2nd poker (2,2) 3rd Bertran Blitz (2,6) 4th Titielmago (2,67) 5th Mybook 15 (4))
ETB-Tour Gauntlet 2009
MoonriderIII - poker 1.3 4.5 - 5.5 +2/-3/=5 45.00%
MoonriderIII - bertran blitz 1.0 2.0 - 8.0 +0/-6/=4 20.00%
MoonriderIII - yograj 2.4 V1.1 2.0 - 8.0 +0/-6/=4 20.00%
MoonriderIII - titielmago 6.1 4.0 - 6.0 +1/-3/=6 40.00%
MoonriderIII - hurricane 3 3.5 - 6.5 +1/-4/=5 35.00%
ETB-Tour Gauntlet-5 2009
OM Blitz Book 2.0 - Poker 1.3 4.0 - 6.0 +0/-2/=8 40.00%
OM Blitz Book 2.0 - Bertran Blitz 1.0 4.5 - 5.5 +0/-1/=9 45.00%
OM Blitz Book 2.0 - yograj 2.4 V1.1 4.0 - 6.0 +1/-3/=6 40.00%
OM Blitz Book 2.0 - titielmago 6.1 5.0 - 5.0 +3/-3/=4 50.00%
OM Blitz Book 2.0 - Hurricane 3 4.0 - 6.0 +2/-4/=4 40.00%
ETB-Tour Gauntlet-6 2009
Troy 2 - Poker 1.3 3.5 - 6.5 +1/-4/=5 35.00%
Troy 2 - Bertran Blitz 1.0 6.5 - 3.5 +3/-0/=7 65.00%
Troy 2 - yograj 2.4 V1.1 6.0 - 4.0 +3/-1/=6 60.00%
Troy 2 - titielmago 6.1 4.5 - 5.5 +0/-1/=9 45.00%
Troy 2 - Hurricane 3 4.0 - 6.0 +0/-2/=8 40.00%
adding Troy 2 results:
1. Poker 1.3 27,0/40
2. Titielmago 6.1 25,5/40
2. Hurricane 3 25,5/40
4. Troy 2 24,5/40
5. Bertran Blitz 1.0 24,0/40
6. Yograj 2.4 V1.1 23,5/40
removing Yograj 2.4 V1.1 results:
New Top 5:
1 2 2 4 4
1. Poker 1.3 ********** 5,5 - 4,5 4.5 - 5,5 5,5 - 4,5 6,5 - 3,5 22,0/40
2. Bertran Blitz 1.0 4,5 - 5,5 ********** 5,5 - 4,5 7,0 - 3,0 3,5 - 6,5 20,5/40
2. Titielmago 6.1 5.5 - 4.5 4,5 - 5,5 ********** 5.0 - 5.0 5.5 - 4.5 20,5/40
4. Hurricane 3 4,5 - 5,5 3,0 - 7,0 5,0 - 5,0 ********** 6,0 - 4,0 18,5/40
4. Troy 2 3,5 - 6,5 6,5 - 3,5 4,5 - 5,5 4,0 - 6,0 ********** 18,5/40
i didnt have a problem downloading titielmago..
if you want to stay on top without creating a book on your own you always have to be up to date. best is if your the first who knows whats currently the best. my results are pretty close at the moment.. its hard to say whichone is best and many are good to use.
if you want an insider tip about what public book to use on playchess at the moment, just send me a pm. if i seriously recommend anything here it wont be an insider tip anymore.
Yograj 2.4 V1.1 3383
Bertran Blitz 1.0 3351
Titielmago 6.1 3327
Hurricane 3 3323
Poker 1.3 3312
Troy 2 3308
Bombook 1.1 3307
Om Blitz Book 2.0 3289
MyBook 15 3279
Bulldozer 2.7 3256
Perfect 2009 3257
Moonrider III 3208
sorry for the delay:
Rybka 3 32-bit titielmago40 = Titielmago 6.1
Rybka 3 32-bit Troy = Troy 2
Rybka 3 32-bit Yograj 1.1 = Yograj 2.4 V1.1 (=not the public version which has the same name, ask sunny why)
Rybka 3 32-bit Bertran 10 = bertran blitz 1.0
Rybka 3 32-bit Moonrider = Moonrider III
thats it, rest should be clear
> Yograj 2.4 V1.1 (=not the public version which has the same name, ask sunny why)
Book is same as public version with minor tweaking. If you download the public version and the private version there very slight changes. I just opened some more variations in private version. For author who are new to book making can change minor variations and see the result will be even much better.
ETB-Tour Gauntlet-9 2009
Mybook 15.1 - Poker 1.3 6.5 - 3.5 +4/-1/=5 65.00%
Mybook 15.1 - Bertran Blitz 1.0 4.5 - 5.5 +2/-3/=5 45.00%
Mybook 15.1 - titielmago 6.1 4.5 - 5.5 +0/-1/=9 45.00%
Mybook 15.1 - Hurricane 3 5.5 - 4.5 +1/-0/=9 55.00%
Mybook 15.1 - Troy 2 5.5 - 4.5 +1/-0/=9 55.00%
adding Mybook 15,1 result:
1. MyBook 15.1 26,5/50
2. Bertran Blitz 1.0 26,0/50
2. Titielmago 6.1 26,0/50
4. Poker 1.3 25,5/50
5. Hurricane 3 23,0/50
5. Troy 2 23,0/50
removing Troy 2 results (because of the 4-6 direct result against Hurricane 3)
New Top 5:
1 2 3 4 5
1. Mybook 15.1 ********** 5,5 - 4,5 4,5 - 5,5 6,5 - 3,5 5,5 - 4,5 22,0/40
2. Bertran Blitz 1.0 4,5 - 5,5 ********** 5,5 - 4,5 4,5 - 5,5 7,0 - 3,0 21,5/40
3. Titielmago 6.1 5,5 - 4,5 4.5 - 5.5 ********** 5.5 - 4.5 5.0 - 5.0 20,5/40
4. Poker 1.3 3.5 - 6,5 5,5 - 4,5 4,5 - 5,5 ********** 5,5 - 4,5 19,0/40
5. Hurricane 3 4,5 - 5,5 3,0 - 7,0 5,0 - 5,0 4,5 - 5,5 ********** 17,0/40
ETB-Tour Gauntlet-10 2009
HS-Masterbook - Poker 1.3 3.0 - 7.0 +0/-4/=6 30.00%
HS-Masterbook - Bertran blitz 1.0 2.5 - 7.5 +0/-5/=5 25.00%
HS-Masterbook - titielmago 6.1 4.0 - 6.0 +2/-4/=4 40.00%
HS-Masterbook - Hurricane 3 3.5 - 6.5 +0/-3/=7 35.00%
HS-Masterbook - Mybook 15.1 4.0 - 6.0 +1/-3/=6 40.00%
ETB-Tour Gauntlet-11 2009
blitzgamebook - Poker 1.3 3.5 - 6.5 +1/-4/=5 35.00%
blitzgamebook - Bertran blitz 1.0 4.5 - 5.5 +1/-2/=7 45.00%
blitzgamebook - titielmago 6.1 3.5 - 6.5 +2/-5/=3 35.00%
blitzgamebook - Hurricane 3 4.5 - 5.5 +2/-3/=5 45.00%
blitzgamebook - Mybook 15.1 5.5 - 4.5 +1/-0/=9 55.00%
ETB-Tour Gauntlet-12 2009
Fx3 - Poker 1.3 4.0 - 6.0 +1/-3/=6 40.00%
Fx3 - Bertran blitz 1.0 4.0 - 6.0 +0/-2/=8 40.00%
Fx3 - Titielmago 6.1 5.0 - 5.0 +1/-1/=8 50.00%
Fx3 - Hurricane 3 4.0 - 6.0 +1/-3/=6 40.00%
Fx3 - Mybook 15.1 5.5 - 4.5 +3/-2/=5 55.00%
Om Gambit Book
GabrielNathans Miniature ChessBase Openings Book (will call it 'Gabriel-mini')
Amazing (still waiting for the download link...is there somebody scared? ;) just kidding)
Books i hope to find time to test one day..allthough i dont expect a good result...im sure they dont make it to top 5:
Compmaster 3.0 Beta
Rybka 3 without opening book
> Amazing (still waiting for the download link...)
Probably not until you test with book learning(And other things), if you do not test with book learning turned on then there is no reason to have the book tested. Now it is obvious why in your test Yograj beat MyBook +10/=0/-0, MyBook never adapted to it's mistakes and kept losing the same way. In Sedat's test's with book learning the score was +5/=5/-0 in favor of MyBook 15. You can not run a test without book learning turned on for books that are made to use book learning. I know other people who think the way you are and it just does not show the book's real strength.
Edit: I need to say that with the amount of books and games that Jonas is testing he has probably chosen the most efficient way to test books, but it needs to be written somewhere in this thread the exact settings Jonas is using to test. I am not saying that Yograj would not have still won the match between MyBook at this time control with book learning, but anytime a book beats another book 10 games straight especially with a book the size and strength of MyBook a flag has to go off in everyone's head. MyBook 15 plays the 6.Bg5 Najdorf but if it losses enough times it will switch to the 6.Be3 lines if book learning is enabled. I have explained to Jonas a good way of testing books(via PM) and still using book learning but it would make his job a lot tougher.
believe me i chose the best method for book testing. learning on destroys the books and destroys my results. (like it did in my previous tournaments)
learning on is good if you only test 1 book. but i run gauntlets.
i want the same conditions for every book so if i want to use learning on i'd have to test book A vs book B then overwrite the books with the previous book files to remove all they learned and then test book A vs Book C and remove again and so on. else they learn from bad books and change their good lines against better books.
that is not 'thougher' thats simply impossible when testing hundrets of books. there are also more reasons why learning: off is better.
obviously i had a different version of yograj or a different version of mybook15 than sedat used in his tests.
>(And other things)
what other things?
> how often do you want me to post the exact settings? you can read them on my first post in this thread.
You do not post the exact settings, you just post the time control and learning off. How about things like.. Type of CPU and clock speed, hash size, permanent brain, and EGTB's. Look at the conditions section on Sedat's website and see the "exact settings" he informs others about.
> that is not 'thougher' thats simply impossible when testing hundrets of books.
Maybe you should try learning from Sedat. Sedat used book learning and was able to test 21 books this pass but go back and look at SCCT IV and previous SCCT Tournaments.
> learning on destroys the books and destroys my results.
No my friend, learning off destroys your results. You are not testing the books the way the authors intended them to be tested, therefore your results are meaningless and unreliable. Filter your games for doubles, have you even checked that yet? With book learning off your games are going to repeat themselves way more often then with book learning. Results like your Yograj 10-0 score over MyBook speak for themselves, Yograj is a great book but even Sunny would have to admit that performance is not reflected in even his own private tests.
> obviously I had a different version of yograj or a different version of mybook15 than sedat used in his tests.
I doubt that very much. Sunny has said that the version Sedat, you and myself tested was the same version. MyBook 15 is "MyBook 15". My own tests with the public version of MyBook 15 and private Yograj are much different then yours.
> what other things?
Well for one thing, you seem to be online a lot. Do you have two PC's, or are you using the one that you are testing on to surf the forum during your testing? I also do not like the "format" of this eternal top five thing which ties into the whole book learning thing.
> I also do not like the "format" of this eternal top five thing which ties into the whole book learning thing.
What I like about the format is the originality, if Jonas went and made a clone of Sedat's tourney and got the same results all the effort would just be redundant, here at least we get a different angle. So, I think different approaches should be encouraged even if they're not optimal.
i have written the exact conditions a hundret times. most of them you can read on my homepage and on the thread ETB-Book Tournament (Info + Registration) and i gave informations about my hardware a hundret times. yes, its weak and thats why i use 4+1. blame italian81 he uses almost same hardware testing 1+0.
learning is off because i want to give an exact result on how strong the books are after downloaded and not after they learned against many books that others will never play against. the result gets totally useless if learning is on.
lines also change due to variety. learning is not a good method to increase book strengh. if a book has no variety and if its main lines are bad it is easy possible that it gets a bad result in my tests. i recommend everyone to turn it off if using a book and then either tune it by hand or get the next best book if its lines doesnt work anymore.
sedat doesnt care about those things. first he doesnt need to care that much about it because he gives a deadline on registering and then starts testing. he dont have to care which book is tested first. i test one book after another. his results are less interesting for people who want to use the best book on chessbase. in fact noone should care if a book gets leader because it scored well against tons of weak or old books noone ever uses on playchess.
maybe your book really was best at the time you sent it to sedat. but would it also win against mybook 15.1 or other new versions i already test (Venezuela 1.3, the new bouddha, Titielmago 6.1) and other versions that now come out? no... because they can tune against you having the games from sedats test. so what you would have to do is release a new version to stay on top.
here you could prove it. if your book isnt top anymore i would find out just a few days after it happened. and you can send me your next version.
for many reasons:
1. those books shouldnt change places of the top books. noone is using them.
2. those books shouldnt affect their learning.
3. after running gauntlets of weaker books all tested books got weaker in later tests because learning was on. so it was best to get tested at last.
4. after my test was finished many new versions came out
so i had to solve these problems to make games more usefull to book authors and to make results more usefull for users who play on playchess:
1. learning: off so it doesnt destroy any results
2. get results faster because new books are released every day
3. higher time control
4. make it eternal so it doesnt mather when you join
5. only use results of the best books
6. automatically more games from strong books
7. less games from weak books
> 1. learning: off so it doesnt destroy any results
I will tell you what. Make a book, send it to Sedat or someone for private testing. When you spend hours making sure you book will play more then one variation you will understand why book learning is needed. In a lot of books they will play variation A 95% of the time and variation B 5% of the time. Maybe against a certain book variation A losses and variation B wins, with book learning off you are making book that play this way lose 95% of the time in that situation. With learning on they will only lose once or twice and then win a few games until the other book is forced to play something else. You have no variety without book learning and I bet hundreds of your games are doubles, I will check for you....
i already got your argument and you repeat it.. i still use learning off for reasons i already mentioned.
what you have to do is find the exact mistake and fix it...instead of removing the whole line because thats what learning does.
simple example: you have 1.e4 and 1.d4 in your book and it plays e4 and looses with that. would you say 1.e4 was the mistake? the book doesnt play it again because it lost.
> learning is not a good method to increase book strengh....i recommend everyone to turn it off if using a book and then either tune it by hand or get the next best book if its lines doesnt work anymore.
It is official, you have no idea what you are talking about. This whole paragraph in your above post is going to let everyone know how ignorant you are to book testing and book making. I wonder if authors will still send you private books?
>sedat doesnt care about those things. first he doesnt need to care that much about it because he gives a deadline on registering and then starts testing. he dont have to care which book is tested first. i test one book after another. his results are less interesting for people who want to use the best book on chessbase. in fact noone should care if a book gets leader because it scored well against tons of weak or old books noone ever uses on playchess.
Whoa son, you are talking shit about the #1 book tester in the world. It is obvious you have no idea how he tests book or anything about his previous tests.
>maybe your book really was best at the time you sent it to sedat. but would it also win against mybook 15.1 or other new versions i already test (Venezuela 1.3, the new bouddha, Titielmago 6.1) and other versions that now come out? no... because they can tune against you having the games from sedats test.
Here are results from a 4+1 gauntlet that I ran last night, as it happens I was testing one of the books you mentioned. Amazing is always updated so yes this version is better then Sedat's
Rybka 3 Amazing 3200 - Rybka 3 Fx3 3200 12.5 - 7.5 +6/=13/-1 62.50%
Rybka 3 Amazing 3200 - Rybka 3 Fx7 3200 12.5 - 7.5 +6/=13/-1 62.50%
Rybka 3 Amazing 3200 - Rybka 3 Hurricane 3 3200 11.0 - 9.0 +2/=18/-0 55.00%
Rybka 3 Amazing 3200 - Rybka 3 MyBook 15.1 3200 13.5 - 6.5 +7/=13/-0 67.50%
Rybka 3 Amazing 3200 - Rybka 3 Bombook 1.1 3200 12.5 - 7.5 +5/=15/-0 62.50%
the paragraph had nothing to do with ignorance if i tell that learning is not a good method in improving book strengh... AFAIK there are better ways.
i know exactly how sedat is testing books... and i also know that he is the #1 book tester. i did never doubt that
never ment to offend you and make you angry
whatever... if your book is that good why dont you send it to me and prove it? im still a fair tester and if your book has such a good result i would be happy.
you have nice results there and in sedats tests and i never said your book is weak.
> whatever... if your book is that good why dont you send it to me and prove it? im still a fair tester
Well I am waiting on the challenge from Yuri, so I do not want games or any information about how the latest version plays to be leaked or released. I know you are a fair tester and I would trust you, but no matter what without book learning anything can happen. If another book keys in on a losing line my book along with any other books they would be completely defenseless but with book learning the book can adapt and play better moves instead of losing the same way over and over and over. I will not have Amazing tested without book learning.
>would much rather have a nice discussion instead.
ok, me too.. never meant to be aggressive.
and im not ignorant: propably you can change my mind and ill switch back to learning on again. in fact nelson was the one who convinced me to use learning off. i used almost the same arguments as you do now. but my opinion atm is still learning should be off
and to that 10-0 result. until now it only happened once. so it doesnt happen 95% of all time...not 1 other result, if you look at all the gauntlets i ran, turned out like that. its not my fault that it happened. i also couln'd trust my eyes. it was just an amazing hit by sunny and sure there was luck.
Edit: I can not send you "Amazing" but there is somebody out there wanting his book tested. Check your PM :)
please read this thread beginning with the first post of nelson and you might laugh http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?tid=11143
> please read this thread beginning with the first post of nelson and you might laugh
The argument he was making is right. There is a difference in testing all the books at different times and letting them build up learning verses letting them use learning but only for each individual gauntlet. I keep a set of default zipped books and each test I run I use learning but when I start a new test I reset the books back to the default version.
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill