Latest Updates:
Hot Topic (More than 10 Replies) Endgame Tablebases and the Anand-Gelfand (Read 8345 times)
Vass
God Member
*****
Offline


I Love ChessPublishing!

Posts: 1105
Joined: 06/22/11
Re: Endgame Tablebases and the Anand-Gelfand
Reply #12 - 06/18/13 at 18:56:40
Post Tools
The endgame tablebases, starting from a specific position (as the ones you point out), are just a tree-like expanding structures. Therefore, only an exclusively targetted, specially designed software can give you the answers of your questions in a proper manner. Up to this day, there is no such a software available, as long as I know.  Undecided
  
Back to top
 
IP Logged
 
tony37
God Member
*****
Offline


I Love ChessPublishing!

Posts: 616
Joined: 10/16/10
Re: Endgame Tablebases and the Anand-Gelfand
Reply #11 - 06/18/13 at 17:27:27
Post Tools
I know where to find the tablebases but "win in 63" doesn't tell me if and when a capture happens, and in a 2 bishops versus knight endgame the knight tends to get captured towards the end, but how to know whether there's a way to prolong this 51 moves? (knowing that longest distance to mate and longest distance to capture may be other routes)

edit: another example is 2 knights versus pawn (some may think that stuff never happens but I already analysed this twice in my correspondence games), it can take more than 50 moves to win so how to know if there is a way for the defending side not to make a pawn move for at least 51 moves?
  
Back to top
 
IP Logged
 
Vass
God Member
*****
Offline


I Love ChessPublishing!

Posts: 1105
Joined: 06/22/11
Re: Endgame Tablebases and the Anand-Gelfand
Reply #10 - 06/18/13 at 16:47:05
Post Tools
tony37 wrote on 06/18/13 at 15:46:08:
is there a way to look up in the tablebases whether a position is winning taking into account the 50-move rule?
I was thinking specifically about 2 bishops versus knight


Yes, there is!
http://www.k4it.de/index.php?topic=egtb&lang=en
or
http://www.shredderchess.com/online-chess/online-databases/endgame-database.html
  
Back to top
 
IP Logged
 
tony37
God Member
*****
Offline


I Love ChessPublishing!

Posts: 616
Joined: 10/16/10
Re: Endgame Tablebases and the Anand-Gelfand
Reply #9 - 06/18/13 at 15:46:08
Post Tools
is there a way to look up in the tablebases whether a position is winning taking into account the 50-move rule?
I was thinking specifically about 2 bishops versus knight
  
Back to top
 
IP Logged
 
TonyRo
God Member
*****
Offline


I'm gonna crack your skull!

Posts: 1777
Location: Cleveland, OH
Joined: 11/26/07
Gender: Male
Endgame Tablebases and the Anand-Gelfand
Reply #8 - 06/01/12 at 17:37:26
Post Tools
I would imagine that a computer built specifically for chess would load the Nalimov 6-mans onto an SSD, dramatically increasing the access speed no?



Edited:
Title of post edited to reflect the current name of the thread. ~SF May 31, 2012
« Last Edit: 06/02/12 at 04:53:49 by Smyslov_Fan »  
Back to top
 
IP Logged
 
brabo
God Member
*****
Offline


Welcome chessfriend

Posts: 982
Joined: 02/02/07
Re: Endgame Tablebases and the Anand-Gelfand match
Reply #7 - 06/01/12 at 13:05:38
Post Tools
Vass wrote on 06/01/12 at 12:26:49:
@brabo
Well, one can work out a comromise with the 5-men Gaviota tablebases - also available for download online, fit for Houdini & Critter engines and working faster than the Nalimov ones.
Robobases/Triplebases are another option (fit for other engines). They can be loaded directly in RAM and this makes them the fastest ones available, but that's another story.  Wink

Sure I've been thinking about that possibility. However I doubt that it is really worth to insert the 5-men tablebases if you anyway check manually the 6-men tablebases once 6 pieces are left on the board (like I do). True, the engine will check 5-men tablebases even if there are still much more pieces on the board but what is the real value of that? Normally a transition from 6 to 5 pieces is taking a lot of moves unless it is a simple endgame to understand. If it indeed takes a lot of moves then the popping up of the 5-men tablebasehits in the sidevariations of the ongoing engineanalysis have no real value as the critical 5-men tablebasehits are only popping up behind the horizon.

Anyway this is a very specialized domain of computing which would need severe testing to know the exact optimal configuration. Any volunteers for doing such indepth testing? Probably not so we probably best stick to the methods we trust the most.
  
Back to top
WWW  
IP Logged
 
Vass
God Member
*****
Offline


I Love ChessPublishing!

Posts: 1105
Joined: 06/22/11
Re: Endgame Tablebases and the Anand-Gelfand match
Reply #6 - 06/01/12 at 12:26:49
Post Tools
@brabo
Well, one can work out a comromise with the 5-men Gaviota tablebases - also available for download online, fit for Houdini & Critter engines and working faster than the Nalimov ones.
Robobases/Triplebases are another option (fit for other engines). They can be loaded directly in RAM and this makes them the fastest ones available, but that's another story.  Wink
  
Back to top
 
IP Logged
 
brabo
God Member
*****
Offline


Welcome chessfriend

Posts: 982
Joined: 02/02/07
Re: Endgame Tablebases and the Anand-Gelfand match
Reply #5 - 06/01/12 at 08:53:10
Post Tools
Alias wrote on 06/01/12 at 08:23:44:
brabo: I haven't bought the Nalimov tablebases. Just knowing that I can find particular positions on the on-line versions is fine for me.

I tried to google translate your blog. It was quite difficult to follow. I haven't played any CC games in a few years but in my own experience the way to use computers hasn't changed so much even though the programs and hardware have changed dramatically. Someone more experienced, like eg Larry Kaufman, probably has a more modern view on using programs than mine.

1) You don't need to buy the Nalimov tablebases to store them on your PC. You can download all of them for free on http://kirill-kryukov.com/chess/tablebases-online/

2) There are quite some pros and cons (also partly explained on the above link) to store the tablebases on your own computer.
Big advantage:
- While an engine is calculating, it automatically checks if an analysed position hits a tablebase. This can lead to thousands of automatic tablebasehits in an analysis of a specific endgame. This way you have 100% certainty that the millions of positions an engine is calculating, are automatically checked if a tablebasehit is made. So normally leading to higher quality of analysis which is in correspondence chess especially important.

Big disadvantages:
- You need to store a gigantic amount of data on your pc which can cause some nasty sideeffects.
- I have the impression that the engine seriously slows down by checking the tablebases as it has to continuously swap between memory and disk. So it often takes much more time to go to the same level of depth for an engine.
- If you are quite far from the 6 piececonfigurations, the tablebasehits are quite meaningless as the mainline is cleary something else. Again leading to unusefull consumption of processingpower and time.

My conclusion today is that storing the tablebases aren't necessarily bringing higher quality analysis if time is restricted. For OTB analysis I accept fully the loss of quality in the transition phase to the nalimovdatabases. I am however not sure if I would accept this for correspondencechess as it is often exactly in that phase that the men are separated from the boys.
  
Back to top
WWW  
IP Logged
 
Vass
God Member
*****
Offline


I Love ChessPublishing!

Posts: 1105
Joined: 06/22/11
Re: Endgame Tablebases and the Anand-Gelfand match
Reply #4 - 06/01/12 at 08:42:17
Post Tools
Nalimov tablebases are a must for the modern correspondence chess. Having them on the hard disc is absolutely necessary. A correspondence chess player has to show the path to Nalimov tablebases on the computer hard disc to engines when analyzing even in the deep middlegame. The modern engines analyze some forced variations till the deep endgame and Nalimov tablebases help them to assess the outcome rightfully. No online link will be of help to do so.  Wink
  
Back to top
 
IP Logged
 
Alias
God Member
*****
Offline



Posts: 1504
Location: East of the river Svartån
Joined: 11/19/04
Re: Endgame Tablebases and the Anand-Gelfand match
Reply #3 - 06/01/12 at 08:23:44
Post Tools
brabo: I haven't bought the Nalimov tablebases. Just knowing that I can find particular positions on the on-line versions is fine for me.

I tried to google translate your blog. It was quite difficult to follow. I haven't played any CC games in a few years but in my own experience the way to use computers hasn't changed so much even though the programs and hardware have changed dramatically. Someone more experienced, like eg Larry Kaufman, probably has a more modern view on using programs than mine.
  

Don't check me with no lightweight stuff.
Back to top
 
IP Logged
 
brabo
God Member
*****
Offline


Welcome chessfriend

Posts: 982
Joined: 02/02/07
Re: Endgame Tablebases and the Anand-Gelfand match
Reply #2 - 06/01/12 at 07:51:05
Post Tools
Alias wrote on 06/01/12 at 07:39:10:
I once emailed the link to a webpage with the Nalimov table bases to a CC opponent who kept on playing in a drawn position.

What do you recommend to analyse? Just check manually a position on such site or download the gigantic databases of Nalimov? I wrote very recently an article about analysing OTB games on my blog (again in Dutch unfortunately so you need e.g. googletranslate) http://schaken-brabo.blogspot.be/2012/05/analyseren-met-de-computer.html
in which I recommend not to download the databases but maybe you have a different idea.
I can also imagine that correspondence should follow a bit different method of analysis than for on the board chess.
  
Back to top
WWW  
IP Logged
 
Bibs
God Member
*****
Online



Posts: 2237
Joined: 10/24/06
Endgame Tablebases and the Anand-Gelfand match
Reply #1 - 06/01/12 at 07:07:06
Post Tools
Thanks for that Brabo.
IM or not IM, short or tall, rich or poor. Some websites you know, some you don't.
Didn't know that, top site, cheers muchly.
  
Back to top
 
IP Logged
 
Alias
God Member
*****
Offline



Posts: 1504
Location: East of the river Svartån
Joined: 11/19/04
Endgame Tablebases and the Anand-Gelfand
06/01/12 at 07:39:10
Post Tools
I once emailed the link to a webpage with the Nalimov table bases to a CC opponent who kept on playing in a drawn position.




Edited:
Title of thread edited.   ~SF May 31st, 2012.
« Last Edit: 06/02/12 at 04:52:19 by Smyslov_Fan »  

Don't check me with no lightweight stuff.
Back to top
 
IP Logged
 
Bookmarks: del.icio.us Digg Facebook Google Google+ Linked in reddit StumbleUpon Twitter Yahoo