New Poll: What is your opinion ?

[Sedat Canbaz]

Ok, I admit I didn't expect such a good result for Rybka (considering smp adds a bit of non-determinism by itself).
One more reason not to enter any tournament without a good book.

Yes...and I don't think that only Rybka 3 can beat Stockfish 5

For example, if I will start to work very very hard over this book,
then there is a big possibility even your engine (Cheng) to beat easily Stockfish (without book)

Of course I don't want to spend more time over optimizing this book,
Because it requires really much efforts...no time for all sorry...
Even if I will be successful about Rybka to be 100 Elo over Stockfish...is that very important ? I think no...

In other words,
There is nothing impossible !!

I think that it is easier to get good result at fast time control.
If you play stockfish with no book against Rybka3 with the same book at 30 minutes per game then I expect stockfish to get more than 60%.

Yes...probably your estimation can be right, because at slow time controls we can not be sure exactly... !
But when I say right, I mean especially for the current beta AntiStock.ctg book (because I worked over this book only approx. 3-4 hours)

But what about? if I start to optimize my book for 30 minutes games,
Let's say after 3-5 months of hard working over the book lines...

And after such period of time,
Do you expect that 60 % will be in favor for Stockfish ??
Then I guess...the results will be 60% in favor for Rybka

[Sedat Canbaz]

Btw,
I noticed also that mostly of my chess activities are underestimating
Or ending with negative opinions (including this testing too)

Why nobody can't say simple: 'thanks' or 'congratulations' or 'well-done'

I know those guys are great experts...but not sure on papers or on real Arena ?!

And I have no patience to see theirs books as participant in SCCT !)

[Leto]

I'm surprised you were able to get such results so quickly. Congrats Sedat. I guess Stockfish isn't as good at the opening phase as I thought.

[Sedat Canbaz]

I'm surprised you were able to get such results so quickly. Congrats Sedat. I guess Stockfish isn't as good at the opening phase as I thought.

Hello Leto,

Thank you for your positive comment!

Frankly even me I was surprised,
After 3-4 hours of working, I expected to see approx. 100 Elo difference

But however,
This would be injustice if we say only Stockfish suffers...
I mean all engines (Houdini, Komodo, Gull, Rybka etc...) suffer when they play without book

And once more I say:
Stockfish is a wonderful chess engine, a piece of diamond and plus free !!!


Best,
Sedat

[Uri Blass]

Ok, I admit I didn't expect such a good result for Rybka (considering smp adds a bit of non-determinism by itself).
One more reason not to enter any tournament without a good book.

Yes...and I don't think that only Rybka 3 can beat Stockfish 5

For example, if I will start to work very very hard over this book,
then there is a big possibility even your engine (Cheng) to beat easily Stockfish (without book)

Of course I don't want to spend more time over optimizing this book,
Because it requires really much efforts...no time for all sorry...
Even if I will be successful about Rybka to be 100 Elo over Stockfish...is that very important ? I think no...

In other words,
There is nothing impossible !!

I think that it is easier to get good result at fast time control.
If you play stockfish with no book against Rybka3 with the same book at 30 minutes per game then I expect stockfish to get more than 60%.

Yes...probably your estimation can be right, because at slow time controls we can not be sure exactly... !
But when I say right, I mean especially for the current beta AntiStock.ctg book (because I worked over this book only approx. 3-4 hours)

But what about? if I start to optimize my book for 30 minutes games,
Let's say after 3-5 months of hard working over the book lines...

And after such period of time,
Do you expect that 60 % will be in favor for Stockfish ??
Then I guess...the results will be 60% in favor for Rybka

I do not expect 60% for stockfish if you use more time.
My point is only that I believe that the importance of book is relatively smaller at longer time control like the time control that nakamura used against stockfish in the match.

[Sedat Canbaz]

Ok, I admit I didn't expect such a good result for Rybka (considering smp adds a bit of non-determinism by itself).
One more reason not to enter any tournament without a good book.

Yes...and I don't think that only Rybka 3 can beat Stockfish 5

For example, if I will start to work very very hard over this book,
then there is a big possibility even your engine (Cheng) to beat easily Stockfish (without book)

Of course I don't want to spend more time over optimizing this book,
Because it requires really much efforts...no time for all sorry...
Even if I will be successful about Rybka to be 100 Elo over Stockfish...is that very important ? I think no...

In other words,
There is nothing impossible !!

I think that it is easier to get good result at fast time control.
If you play stockfish with no book against Rybka3 with the same book at 30 minutes per game then I expect stockfish to get more than 60%.

Yes...probably your estimation can be right, because at slow time controls we can not be sure exactly... !
But when I say right, I mean especially for the current beta AntiStock.ctg book (because I worked over this book only approx. 3-4 hours)

But what about? if I start to optimize my book for 30 minutes games,
Let's say after 3-5 months of hard working over the book lines...

And after such period of time,
Do you expect that 60 % will be in favor for Stockfish ??
Then I guess...the results will be 60% in favor for Rybka

I do not expect 60% for stockfish if you use more time.
My point is only that I believe that the importance of book is relatively smaller at longer time control like the time control that nakamura used against stockfish in the match.

Uri,

I don't know how much you are experienced in book making or in book competitions...

But I am one of the people who has the biggest collection of the strongest private books and so far I've tested probably more than 1000 books

And that experience shows me that there is nothing impossible !
There are such superior book lines, where they are just killer moves...
These great books are playing for win or at least the game is ending in a draw

Why I believe that we can get approx. 60 % at 30 minutes per game,
Simply because we can use the new ideas of the strongest available top engines

Do you know why Perfect books were always at best performance (during my testing on Playchess or in any other book competitions?)
I worked very very hard...not many people can imajine
I found new ideas that never played before...and of course the most opponents were helpless...and the success was in the door !

And now I don't want prove again...plus I have no much free time
But I am quite sure in that...there are top book makers who can manage to archive 60 % in favor for Rybka

And last,
If in blitz is achievable...at slow time control has possibility too!
Maybe we can not see exactly same performance, but there is a BIG possibility the results even at slow time control to be in favor for Rybka

[Uri Blass]

Ok, I admit I didn't expect such a good result for Rybka (considering smp adds a bit of non-determinism by itself).
One more reason not to enter any tournament without a good book.

Yes...and I don't think that only Rybka 3 can beat Stockfish 5

For example, if I will start to work very very hard over this book,
then there is a big possibility even your engine (Cheng) to beat easily Stockfish (without book)

Of course I don't want to spend more time over optimizing this book,
Because it requires really much efforts...no time for all sorry...
Even if I will be successful about Rybka to be 100 Elo over Stockfish...is that very important ? I think no...

In other words,
There is nothing impossible !!

I think that it is easier to get good result at fast time control.
If you play stockfish with no book against Rybka3 with the same book at 30 minutes per game then I expect stockfish to get more than 60%.

Yes...probably your estimation can be right, because at slow time controls we can not be sure exactly... !
But when I say right, I mean especially for the current beta AntiStock.ctg book (because I worked over this book only approx. 3-4 hours)

But what about? if I start to optimize my book for 30 minutes games,
Let's say after 3-5 months of hard working over the book lines...

And after such period of time,
Do you expect that 60 % will be in favor for Stockfish ??
Then I guess...the results will be 60% in favor for Rybka

I do not expect 60% for stockfish if you use more time.
My point is only that I believe that the importance of book is relatively smaller at longer time control like the time control that nakamura used against stockfish in the match.

Uri,

I don't know how much you are experienced in book making or in book competitions...

But I am one of the people who has the biggest collection of the strongest private books and so far I've tested probably more than 1000 books

And that experience shows me that there is nothing impossible !
There are such superior book lines, where they are just killer moves...
These great books are playing for win or at least the game is ending in a draw

Why I believe that we can get approx. 60 % at 30 minutes per game,
Simply because we can use the new ideas of the strongest available top engines

Do you know why Perfect books were always at best performance (during my testing on Playchess or in any other book competitions?)
I worked very very hard...not many people can imajine
I found new ideas that never played before...and of course the most opponents were helpless...and the success was in the door !

And now I don't want prove again...plus I have no much free time
But I am quite sure in that...there are top book makers who can manage to archive 60 % in favor for Rybka

And last,
If in blitz is achievable...at slow time control has possibility too!
Maybe we can not see exactly same performance, but there is a BIG possibility the results even at slow time control to be in favor for Rybka

I have no experience and no knowledge in book making
and I have no idea how people can make books for your tournament
without writing a special program(maybe there is already a special program that is ready to make books that people use but I do not know).

I agree that it is probably possible to get 60% by rybka3 with no book against stockfish also with longer time control but you need to devote more time.

I think that it is even possible to get 100% by every engine assuming the book can include many million of lines because you only need to build a killer book that is based on full games that stockfish with no book lost.

Stockfish with no book may be non deterministic but it has limited number of options to respond to the killer book and even if you need to memorize 10 millions games then it is clearly not above the memory of computers.

Algorithm to build a killer book also should be simple.
You can play stockfish with no book at 3 minutes per game against a clearly superior opponent so practically stockfish lose every game or almost every game(for example stockfish with 30 minutes per game)

After every game that stockfish with no book lose you decide to add the winning line to the book of the opponent and hopefully after enough games you do not get more new games and stockfish simply lose by the same way that it lost earlier.

.

[Sedat Canbaz]

I have no experience and no knowledge in book making
and I have no idea how people can make books for your tournament
without writing a special program(maybe there is already a special program that is ready to make books that people use but I do not know).

I agree that it is probably possible to get 60% by rybka3 with no book against stockfish also with longer time control but you need to devote more time.

I think that it is even possible to get 100% by every engine assuming the book can include many million of lines because you only need to build a killer book that is based on full games that stockfish with no book lost.

Stockfish with no book may be non deterministic but it has limited number of options to respond to the killer book and even if you need to memorize 10 millions games then it is clearly not above the memory of computers.

Algorithm to build a killer book also should be simple.
You can play stockfish with no book at 3 minutes per game against a clearly superior opponent so practically stockfish lose every game or almost every game(for example stockfish with 30 minutes per game)

After every game that stockfish with no book lose you decide to add the winning line to the book of the opponent and hopefully after enough games you do not get more new games and stockfish simply lose by the same way that it lost earlier.

.

I am happy that finely you started to see the reality

And I will be very glad if you will start to see also in that:
Stockfish should be rated at 3500 Elo (at slow time controls + decent fast hardwares) !

Just I'd like to inform you:
1)So far I did not see any book to be based on 10 millions games
*Note:SCCT private books are based average on 100.000 games
2)AntiStock.ctg (used for the current test) is based on 95569 games
4)And I can build another new book which will be less than 10.000 games
And this book with a big possibility (after 20-30 days) I can get 50 % at 30 min per game
5)To get 50% or better performance (exception a lot of own efforts, time etc..) requires a lot of new strong ideas as well
6)Only after finding the new superior ideas...then we can say EASY to beat...but till to this date...it is very HARD )) !!
etc...

And I am asking you:
-If it was so easy... why both strong human chess players could not beat Stockfish (without book) ??

-Ok...now you will say blitz slow..etc...
If so easy...even at blitz...I have no patience to see 3 hours of working + your book + Rybka to destroy Stockfish

BTW,
As far as I know...you agree with me that Stockfish can be performed at blitz around 3500 Elo, right ?

But unfortunately (in the same time) at slow time controls,
You say Stockfish should be at 3200 Elo, right ?

And if so...you are again wrong !!
Because the Elo difference is not 300 Elo...
As far as I know: between slow time control and blitz is approx. 50-100 Elo

Another a little note (with respect to you):
You are much experienced in comments than me, but I have a little advantage than you:
I am much experienced in practice than you !! ))

So what is more important ? ))

And the below saying quite suits:

.

[Uri Blass]

1)It is not easy for humans to memorize enough lines in order to beat stockfish without book even with the help of rybka3 with inferior hardware
and they do not need only to memorize the games but also to generate them first.

2)I did not claim that it is easy to build a book that beat stockfish.

3)The reason that I believe 3500 is not realistic for computers at long time control is that it is obvious that computer opponents do not play for a draw against superior opponents when humans can do it and playing for a draw is not only about contempt but about choosing opening lines or decision if to simplify to the endgame or not to do it.

[A Distel]

I have no experience and no knowledge in book making
and I have no idea how people can make books for your tournament
without writing a special program(maybe there is already a special program that is ready to make books that people use but I do not know).

I agree that it is probably possible to get 60% by rybka3 with no book against stockfish also with longer time control but you need to devote more time.

I think that it is even possible to get 100% by every engine assuming the book can include many million of lines because you only need to build a killer book that is based on full games that stockfish with no book lost.

Stockfish with no book may be non deterministic but it has limited number of options to respond to the killer book and even if you need to memorize 10 millions games then it is clearly not above the memory of computers.

Algorithm to build a killer book also should be simple.
You can play stockfish with no book at 3 minutes per game against a clearly superior opponent so practically stockfish lose every game or almost every game(for example stockfish with 30 minutes per game)

After every game that stockfish with no book lose you decide to add the winning line to the book of the opponent and hopefully after enough games you do not get more new games and stockfish simply lose by the same way that it lost earlier.

.

I am happy that finely you started to see the reality

And I will be very glad if you will start to see also in that:
Stockfish should be rated at 3500 Elo (at slow time controls + decent fast hardwares) !

Just I'd like to inform you:
1)So far I did not see any book to be based on 10 millions games
*Note:SCCT private books are based average on 100.000 games
2)AntiStock.ctg (used for the current test) is based on 95569 games
4)And I can build another new book which will be less than 10.000 games
And this book with a big possibility (after 20-30 days) I can get 50 % at 30 min per game
5)To get 50% or better performance (exception a lot of own efforts, time etc..) requires a lot of new strong ideas as well
6)Only after finding the new superior ideas...then we can say EASY to beat...but till to this date...it is very HARD )) !!
etc...

And I am asking you:
-If it was so easy... why both strong human chess players could not beat Stockfish (without book) ??

-Ok...now you will say blitz slow..etc...
If so easy...even at blitz...I have no patience to see 3 hours of working + your book + Rybka to destroy Stockfish

BTW,
As far as I know...you agree with me that Stockfish can be performed at blitz around 3500 Elo, right ?

But unfortunately (in the same time) at slow time controls,
You say Stockfish should be at 3200 Elo, right ?

And if so...you are again wrong !!
Because the Elo difference is not 300 Elo...
As far as I know: between slow time control and blitz is approx. 50-100 Elo

Another a little note (with respect to you):
You are much experienced in comments than me, but I have a little advantage than you:
I am much experienced in practice than you !! ))

So what is more important ? ))

And the below saying quite suits:

.

You must have a prodigious amount of mind...