![[de]](/mwf/v12/flags/de.png "Germany")
I don't consider myself to be an 'offbeat player', so I have to disagree that these are just offbeat lines. Blumenfeld has recently been played by Nisipeanu and Volokitin, and by many more top players in the past. The 1.Nf3 f5 2.d3 line was also employed by very strong players, see, e.g., the quite spectacular game Carlsen-Dolmatov some years ago.
But I just wanted to give some examples. You seem to focus on very recent games, and also to heavily rely on 'engine room theory'. This method appears to have its pluses (I wouldn't have believed this to be possible some years ago, but I was probably wrong. Nowadays, it can be a bad mistake not to be familiar with engine games). But it might also have its downsides.
Still, please don't take this as a criticism. I really appreciate your book, and I am aware that there is a lot too much theory to check it all.
Thank you very much, regards,
Jens
But I just wanted to give some examples. You seem to focus on very recent games, and also to heavily rely on 'engine room theory'. This method appears to have its pluses (I wouldn't have believed this to be possible some years ago, but I was probably wrong. Nowadays, it can be a bad mistake not to be familiar with engine games). But it might also have its downsides.
Still, please don't take this as a criticism. I really appreciate your book, and I am aware that there is a lot too much theory to check it all.
Thank you very much, regards,
Jens
![[nl]](/mwf/v12/flags/nl.png "Netherlands")
True, each method has its plusses and its minuses :-). I try to combine both top GM games with engine games,
but somehow the engine has to understand what it is playing. I think engine room theory is very much underrated,
the simple proof of that is that in the Freestyle tournaments incredible novelties have been played and 2600 GM's
never made it to the final. I was happy to outplay a 2575 GM in the final of the 3rd Freestyle, because I had a killer
line in the Poisoned Pawn he didn't know :-).
In any case, I have a very high respect for the people in the engine room. Some of them find novelties that GM's
can only dream of :-). But true, you should not believe everything they play, there is now a trend to play 1.h3 and
throw the opponent out of book, while having 30 moves or more in your own book. This is not really interesting for
me and I think also not for the majority of the Rybka customers.
No problem, I don't see your posting as criticism and I welcome a discussion any time!
Kind regards, Jeroen
but somehow the engine has to understand what it is playing. I think engine room theory is very much underrated,
the simple proof of that is that in the Freestyle tournaments incredible novelties have been played and 2600 GM's
never made it to the final. I was happy to outplay a 2575 GM in the final of the 3rd Freestyle, because I had a killer
line in the Poisoned Pawn he didn't know :-).
In any case, I have a very high respect for the people in the engine room. Some of them find novelties that GM's
can only dream of :-). But true, you should not believe everything they play, there is now a trend to play 1.h3 and
throw the opponent out of book, while having 30 moves or more in your own book. This is not really interesting for
me and I think also not for the majority of the Rybka customers.
No problem, I don't see your posting as criticism and I welcome a discussion any time!
Kind regards, Jeroen
I have a very high respect for the people in the engine room. I am glad you said that as I was just about to delete your account on Playchess ;-)
Good Moaning,
Harvey
I am glad you said that as I was just about to delete your account on Playchess ;-)
:-) :-) Which one!?
:-) :-) Which one!?
the powers I have now can find them all by ip etc.... I will change the names to characters from allo allo :)
Yuck! Please no Gruber or madame Edith in that case :-) :-)
Fallen Madonna with the big Beebies?
Nighthawk sounds good you have 1 similar to that already!?
but Madame Edith would be my 1st choice :)
Nighthawk sounds good you have 1 similar to that already!?
but Madame Edith would be my 1st choice :)
Nah, not Nighthawk! If I had a choice, then I would go for 'Crabtree', 'Rene from the cafe' (because of the waitresses,
that goes without saying!) or 'Mr Alphonse'.
If you pick madame Edith, I will delete that account right away :-)
that goes without saying!) or 'Mr Alphonse'.
If you pick madame Edith, I will delete that account right away :-)
![[us]](/mwf/v12/flags/us.png "United States")
lol
Just to inform you: today I put the requested lines in the Rybka3 book. I just hope this will not encourage a lot of others to make a wish list :-).
BTW, I still think these lines are (at least slightly) offbeat, as they are not played by strong GM's on a regular basis. If you compare them with
the number of games f.e. played in the Slav, I cannot call them main lines :-).
Kind regards, Jeroen
BTW, I still think these lines are (at least slightly) offbeat, as they are not played by strong GM's on a regular basis. If you compare them with
the number of games f.e. played in the Slav, I cannot call them main lines :-).
Kind regards, Jeroen
Thank you very much, that's great! Let's just say we are in complete agreement except for the fact that we would give a slightly differing definition of 'offbeat line'.
Regards,
Jens
Regards,
Jens
![[br]](/mwf/v12/flags/br.png "Brazil")
It is not possible that an author of books to review all the lines. And even if he could,
in a short time someone will discover that some moves ahead, the engine will not find a
good answer to a move that is not necessary the best.
I find very difficult the work of the author of books, because they have to think about
games with several time controls, and with thousands of lines.
Jeroem already written here that he needs 24 hours to review certain lines, and I do not doubt.
I already looked at some positions more than four hours to come to the conclusion that they are
not good, and anyway there is a possibility that I was wrong.
Paulo Soares
in a short time someone will discover that some moves ahead, the engine will not find a
good answer to a move that is not necessary the best.
I find very difficult the work of the author of books, because they have to think about
games with several time controls, and with thousands of lines.
Jeroem already written here that he needs 24 hours to review certain lines, and I do not doubt.
I already looked at some positions more than four hours to come to the conclusion that they are
not good, and anyway there is a possibility that I was wrong.
Paulo Soares
Yes, that is true. Besides, if I take - let's say - 3 months to cover all the sidelines,
then suddenly my main lines are 3 months old and out of date :-)
then suddenly my main lines are 3 months old and out of date :-)
![[ch]](/mwf/v12/flags/ch.png "Switzerland")
Not if you implement the correct line directely. You dont have to change what is good/perfect/works.
I have some variations in my own private book that I know I will never need to change because there is no improvements to do in these lines.
But I agree that there is so many lines that book editing is an endless job.
I spend something like 4 Year editing my private book with the lines I play in tournament so I have double benefit as I also learn.
regards
I have some variations in my own private book that I know I will never need to change because there is no improvements to do in these lines.
But I agree that there is so many lines that book editing is an endless job.
I spend something like 4 Year editing my private book with the lines I play in tournament so I have double benefit as I also learn.
regards
Bouddha, If your "perfect lines" not look for an endgame, you may lose a
game. Already happened to me, especially in games of 3 min.
Man, I am impressed by the creativity of those who plays in Playchess.
Paulo Soares
game. Already happened to me, especially in games of 3 min.
Man, I am impressed by the creativity of those who plays in Playchess.
Paulo Soares
What seems correct now, can be out of date 3 months later :-). Everything evolves, also opening theory.
So what is best at this moment, will have a decent chance of 'not being the best' in 2009.
Yeah, editing takes time, patience and a lot of work. I'll be happy when it is over :-)
So what is best at this moment, will have a decent chance of 'not being the best' in 2009.
Yeah, editing takes time, patience and a lot of work. I'll be happy when it is over :-)
But it will never be over... :-)
comming back to variations, there are many in my book which I know there is no point in editing them. I have been working so many hours with all chess programs digging in the position that finding an improvement novelty in thoese just doesn't sound possible.
comming back to variations, there are many in my book which I know there is no point in editing them. I have been working so many hours with all chess programs digging in the position that finding an improvement novelty in thoese just doesn't sound possible.
And when you try in 2012 with 64 core machines, running engines that are 200 elo better than the ones you used?
Agreed that there can be some exeptions.
When you fully analyzed a line not just by letting the engine think on a position but by going through variations and creating new lines, you somehow in some line exhausted the "good" possibilities.
There are many lines I think no future engine will improve.
There are some where a new and stronger engine would help.
When you fully analyzed a line not just by letting the engine think on a position but by going through variations and creating new lines, you somehow in some line exhausted the "good" possibilities.
There are many lines I think no future engine will improve.
There are some where a new and stronger engine would help.
Hi Jeroen, you have private books for Rybka. Will you empty your pockets with all of the special opening variations into your new book. Just curious ! I would expect your answer will be no. (don't blame you for that in any case)
Kind regards
Wayne
Kind regards
Wayne
No, there will always be a strict seperation between my private tournament book and the official Rybka book.
Now and then I reveil some secrets when I think I will not be able to use them anymore and then I put those
lines in the commercial book as well.
But the other situation also appears: when doing analysis on the commercial book, I very often run into some
crazy, beautiful and/or strong line that is not in my tournament book yet. So both books profit from the
development of their counterpart :-)
Now and then I reveil some secrets when I think I will not be able to use them anymore and then I put those
lines in the commercial book as well.
But the other situation also appears: when doing analysis on the commercial book, I very often run into some
crazy, beautiful and/or strong line that is not in my tournament book yet. So both books profit from the
development of their counterpart :-)
2 Questions :
a) How do you decide what goes in the public book for which we pay ? is it old or weak staff ? (I am pinching you in purpose, please dont be offended)
b) what is the percentage time you work on your private book and on the public book ?
regards
a) How do you decide what goes in the public book for which we pay ? is it old or weak staff ? (I am pinching you in purpose, please dont be offended)
b) what is the percentage time you work on your private book and on the public book ?
regards
Difficult to say which percentage goes to what book. At the moment it is 100% Rybka3.ctg.
The tournament book has been worked on for 5 years now. The public book was made in - let's say - three or four months.
That could give you a little insight in the differences :-). Still you cannot compare both very well, as both books have a
different approach.
What goes into the public book is easy: all the best lines I find during testing, analysis and priority setting, based on the
set of games I selected.
Kind regards, Jeroen
The tournament book has been worked on for 5 years now. The public book was made in - let's say - three or four months.
That could give you a little insight in the differences :-). Still you cannot compare both very well, as both books have a
different approach.
What goes into the public book is easy: all the best lines I find during testing, analysis and priority setting, based on the
set of games I selected.
Kind regards, Jeroen
Jeroen, say no more! Only a total imbecile would expect you to publicly reveal lines which you sweated over for untold hours and carefully salted away for the day when they could be used to great advantage.
That said, would you please send me privately everything you've amassed over the years on the Poisoned Pawn Variation of the Sicilian ?
Thanks in advance,
David
That said, would you please send me privately everything you've amassed over the years on the Poisoned Pawn Variation of the Sicilian ?
Thanks in advance,
David
would you please send me privately everything you've amassed over the years on the Poisoned Pawn Variation of the Sicilian ?
Does your inbox accept 30 MB attachments!? (just kidding!)
Does your inbox accept 30 MB attachments!? (just kidding!)
> The result of Blitz 40/4 says nothing at all.
No, it says a lot. Nowadays, at this time control with normal computers, the depths are HUGE. Also, normally engines handle "normal short" time controls quite good. Time losses are rare. Did you never compare 40/4 and 40/40 lists? The differences are small.
Extreme time controls like 1+0 or 100/1m may be a problem, or much less reliable. 40/4m is entirely ok, giving an average of 6+ seconds per move (10 moves per minute). I would prefer 3+2 or 4+1 for blitz tests though, because I am not interested in potential time trouble effects which are insignificant "at home" and don't exist in analysis.
Yes the test indeed was with 2 CPUs for each engine.
Really? I didn't know that.
So i want to see the games that the test on http://www.hiarcs.com/hiarcs_games.htm was based.
Anyone knows where to find them?
>also the hiarcs book seems to be highly optimized against Rybka II ctg.
Really? I didn't know that.
So i want to see the games that the test on http://www.hiarcs.com/hiarcs_games.htm was based.
Anyone knows where to find them?
I also asked for the games at CCC. Needless to say I am very curious at the number of
book wins Hiarcs 12 got in that match :-).
book wins Hiarcs 12 got in that match :-).
I still don't get it:
•In http://www.hiarcs.com/hiarcs_games.htm it said:
Deep HIARCS 12 v Rybka 2.3.2a UCI +32-30=28
•Peter Grayson that seems to made that test according to you(since you replied in my post about the requested games) reports in this link http://www.hiarcs.net/forums/viewtopic.php?t=1128 :
Score so far:
Rybka 2.3.2a mp (x64) +21 Elo +19 -15 =33
Hiarcs 12 MP -20 Elo +15 -19 =30
These 2 results result in a contradiction, because of the number of draws.
So those 2 matches can't be the same!
So i'm still looking for the games of the page: http://www.hiarcs.com/hiarcs_games.htm. This page says:
Deep HIARCS 12 v Rybka 2.3.2a UCI 51-49 (+32-30=28)
If you add 32 plus 28/2 we have 32+14=46
Also 30 plus 28/2 gives 30+14=44
So we have a score of 46-44 OR the +32-30=28 is wrong and 51-49 is correct.
Note that in the page you give it has Rybka 2.3.2a to win against Hiarcs 12 and not losing as the initial games had.
•In http://www.hiarcs.com/hiarcs_games.htm it said:
Deep HIARCS 12 v Rybka 2.3.2a UCI +32-30=28
•Peter Grayson that seems to made that test according to you(since you replied in my post about the requested games) reports in this link http://www.hiarcs.net/forums/viewtopic.php?t=1128 :
Score so far:
Rybka 2.3.2a mp (x64) +21 Elo +19 -15 =33
Hiarcs 12 MP -20 Elo +15 -19 =30
These 2 results result in a contradiction, because of the number of draws.
So those 2 matches can't be the same!
So i'm still looking for the games of the page: http://www.hiarcs.com/hiarcs_games.htm. This page says:
Deep HIARCS 12 v Rybka 2.3.2a UCI 51-49 (+32-30=28)
If you add 32 plus 28/2 we have 32+14=46
Also 30 plus 28/2 gives 30+14=44
So we have a score of 46-44 OR the +32-30=28 is wrong and 51-49 is correct.
Note that in the page you give it has Rybka 2.3.2a to win against Hiarcs 12 and not losing as the initial games had.
Peter's test is a totally new test against Rybka 64bit. He was just trying to duplicate marks results and chose to try 64bit Rybka instead of 32. He is not a Hiarcs Beta tester. I replied elswhere on this forum that Mark will upload the games v Rybka next time he updates his website.
Best Wishes,
Harvey
Best Wishes,
Harvey
OK thanks. That solved all my questions, nice :-)
[EDIT] Not all since the +32-30=28 should yield a 46-44 score and not a 51-49 as it says. Perhaps you should correct that in:
http://www.hiarcs.com/hiarcs_games.htm
[EDIT] Not all since the +32-30=28 should yield a 46-44 score and not a 51-49 as it says. Perhaps you should correct that in:
http://www.hiarcs.com/hiarcs_games.htm
well spotted i will point it out!
![[mx]](/mwf/v12/flags/mx.png "Mexico")
"Still" removed from the title as requested by Soren Riis.
About the issue, I would have expected HIARCS 12 to at least take number 2, but I'm still happy with it and with the HIARCS Paderborn 2007 version that comes with it.
I've analyzed certain positions with the three HIARCS (including 11.2) and all of them give useful different move choices. When different HIARCS complement each other, it's a good thing (Sort of like Rybka can be complemented with Winfinder.)
About the issue, I would have expected HIARCS 12 to at least take number 2, but I'm still happy with it and with the HIARCS Paderborn 2007 version that comes with it.
I've analyzed certain positions with the three HIARCS (including 11.2) and all of them give useful different move choices. When different HIARCS complement each other, it's a good thing (Sort of like Rybka can be complemented with Winfinder.)
Hi,
Since H12 was released I have spent a few hours watching games played on the playchess server at various time controls to try and help me make up my mind if I should buy yet another HIARCS 32bit engine ( Latest one I have is 11.2 :( ) ... I know it was not a very scientific approach but I only saw H12 win 1 game!! some draws and plenty of losses ... Of course books etc can make a big difference but by watching plenty of games you start to get a feel and it feels like H12 will not even make it to number 2 on the rankings
Regards
Since H12 was released I have spent a few hours watching games played on the playchess server at various time controls to try and help me make up my mind if I should buy yet another HIARCS 32bit engine ( Latest one I have is 11.2 :( ) ... I know it was not a very scientific approach but I only saw H12 win 1 game!! some draws and plenty of losses ... Of course books etc can make a big difference but by watching plenty of games you start to get a feel and it feels like H12 will not even make it to number 2 on the rankings
Regards
![[gr]](/mwf/v12/flags/gr.png "Greece")
I think for a while that actually we don't have a number 2 at all.
Zappa Mexico is very strong and interesting engine but in blitz really sucks and practically all the amateur engine tournaments are in blitz.
Naum 3 is more or less a weak and passive Rybka to tell the truth and Toga is strong and free but against Rybka perform under his elo constadly.
New Fruit's are interesting IMO but not any close to Rybka for sure. So i find the situation quite boring now, especialy because
the new Rybka seems to be even stronger but should still retain the flows engines always have today.
Zappa Mexico is very strong and interesting engine but in blitz really sucks and practically all the amateur engine tournaments are in blitz.
Naum 3 is more or less a weak and passive Rybka to tell the truth and Toga is strong and free but against Rybka perform under his elo constadly.
New Fruit's are interesting IMO but not any close to Rybka for sure. So i find the situation quite boring now, especialy because
the new Rybka seems to be even stronger but should still retain the flows engines always have today.
> Latest one I have is 11.2 :-(
Yes, if you're looking only for strength maybe you should stick with HIARCS 11.2, but for those that still don't have a HIARCS, 12 is a good buy.
I can't say that I'm surprised. Already with HIARCS 11 marketing seemed to be better than the engine...
This result http://www.hiarcs.net/forums/viewtopic.php?t=1128 does seem to show that Marks results are not made up :)
Well, you have a book that is highly optimized against Rybka II ctg . If you would use another book, lets say sheebar.ctg, you wouldn't score that well. So you chose the best circumstances you could imagine to play against Rybka, which is good for marketing but very bad to get an idea of the playing strength...
If we spend a long time making the book why shouldn't we use it? I am sure Jeroen's R3 book will also be very good. I am sure you will not criticise people using that against other engines in tests. R3 + book when released will be added to Marks list.
In actual fact the book was not specifically optimised against Rybka but as a majority of Playchess games involve Rybka then of course a lot of the database it was made from are Rybka games.
In actual fact the book was not specifically optimised against Rybka but as a majority of Playchess games involve Rybka then of course a lot of the database it was made from are Rybka games.
If you build a new book based of games against mainly one book (in this case Rybka II.ctg) it's not fair to make a test after that against Rybka II.ctg . I don't doubt that your book is strong, but it's not as strong against other books like Sheebar as against Rybka II.ctg I would think. I would be interested in te result of a macth Rybka (with Sheebar.ctg) against Hiarcs (with it's own book).
Sadly I am not interested in that as when I buy Rybka I get Rybka book and Hiarcs book with Hiarcs. But if someone else wants to run it under the same conditions that would be fine with me.
I did not say we built a book based on games against Rybka 2 ctg I said we built a book including many games from playchess where there are lots of Rybkas and many not using Rybka 2 ctg.
I did not say we built a book based on games against Rybka 2 ctg I said we built a book including many games from playchess where there are lots of Rybkas and many not using Rybka 2 ctg.
Mark decided to download the Sheebar book and run the test with learning reset as this will not be part of the ratings list. Hiarcs+Hiarcs book won against Sheebar+Rybka. The score was +42-31=27 55.5%
If anyone wants the games send me a pm with an email address and i will send them - the last game was very nice.
If anyone wants the games send me a pm with an email address and i will send them - the last game was very nice.
ok, that sounds impressive, so we have to consider the hiarcs book to be extremely strong indeed. 100 games are a bit too less to come to any conclusion about the difference to Rybka+Rybka II.ctg, but that she scores even worse than with RybkaII.ctg is a suprise for me. I hope we will have the book in ctg or (even better) hsh format soon :)
![[is]](/mwf/v12/flags/is.png "Iceland")
Harvey, could you please post the games from the Rybka-Hiarcs match here on the forum?
Hi Dadi,
Mark says he will put the games, v Rybka, on his website next time he updates it.
Best Wishes,
Harvey
Mark says he will put the games, v Rybka, on his website next time he updates it.
Best Wishes,
Harvey
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill