7-men : Breaking news

Endgame analysis using tablebases, EGTB generation, exchange, sharing, discussions, etc..

7-men : Breaking news

Postby kronsteen » Thu Sep 10, 2009 9:23 pm

Very interesting news In this thread in rybkaforum :
http://rybkaforum.net/cgi-bin/rybkaforu ... l?tid=8477
kronsteen
 
Posts: 88
Joined: Fri Aug 01, 2008 11:20 am
Location: France

Re: 7-men : Breaking news

Postby crplr » Mon Sep 14, 2009 5:34 pm

Comparison of DTM and DTC tables for kqbnkqn looks as natural. It seems that it is serious work. The question is only how he receives results so quickly.
crplr
 
Posts: 5
Joined: Fri Sep 04, 2009 4:58 pm

Longest 7-man win ...

Postby guyhaw » Sat Sep 19, 2009 11:50 am

There's a lot of rubbish in the rybkaforum thread cited - but the idea of videoing an endgame and overlaying a sound track to communicate extra inforamtion is interesting.

Marc's EGT is to DTC so the fact is that 'Depth to Conversion' = 517 moves (given that the two sides are respectively minimising and maximising this figure).

The fact that one can tack on a 7-move DTMate-optimal 6-man phase does NOT mean that DTM = 524 moves in the original endgame.

What is quite likely is that Black hangs in there too long, trying to defer the capture of its man to the last minute, when in fact - to make sure White caught a later bus - it should have let the defence of its material go somewhat earlier. The only limit on DTM for the 7-man endgame is 517m +maxDTM for the subsequent 6-man endgame(s).

Equally, it may be that if White had gone for mate directly, it could have mated in less than 524 moves, but it would still have required 517 moves or more, so there's not much downside here. There's a classic case in KBBKNN: the maxDT50 win in KBBKNN involves White rejecting the offer of a Black Knight because this will lead to a KBBKN phase which is longer than 50 moves: the mate in fact comes in about 24 moves and is an elegant quadrille worth watching on John Tamplin's website.

g
guyhaw
 
Posts: 489
Joined: Sat Jan 21, 2006 10:43 am
Location: Reading, UK

Re: 7-men : Breaking news

Postby mbourzut » Sat Sep 19, 2009 2:08 pm

The work by user Volodya on the Rybkaforum looks legitimate and very interesting.
mbourzut
 
Posts: 30
Joined: Fri Mar 03, 2006 7:48 pm

Re: 7-men : Breaking news

Postby guyhaw » Mon Sep 21, 2009 4:46 pm

Interesting - are you saying that the 2nd version of the 'mate in 545m' makes apparent sense when viewed in DTC terms?

I'd certainly expect the DTC-minimaxing and DTM-minimaxing lines to go along in coincidence for quite a long time.

Good to have more expertise on the block if that's the case.

We're hanging by a single MB/YK thread at the moment, even if that thread is very strong.

g
guyhaw
 
Posts: 489
Joined: Sat Jan 21, 2006 10:43 am
Location: Reading, UK

Re: 7-men : Breaking news

Postby Ed Trice » Sat Oct 30, 2010 2:37 pm

Mark showed me this ending way back in 2006.

http://www.gothicchess.com/endings/chess_kqnkrbn/game.htm

It's a 517 move conversion.
5.0 Ghz Intel Gulftown Supercomputers
http://www.liquidnitrogenoverclocking.com
User avatar
Ed Trice
 
Posts: 58
Joined: Thu Oct 28, 2010 1:13 am
Location: Henlopen Acres, Delaware

Announcement of the 'dtc = 517' result

Postby guyhaw » Mon Nov 01, 2010 12:10 am

Marc' 'showed' everyone inteested the result - in that it was communicated by the relevant forums on or before 26th May 2006, q.v. e.g. http://www.xs4all.nl/~timkr/chess2/diary_16.htm

g
guyhaw
 
Posts: 489
Joined: Sat Jan 21, 2006 10:43 am
Location: Reading, UK

Re: 7-men : Breaking news

Postby Ed Trice » Mon Nov 01, 2010 4:32 am

Maybe I should rephrase:

Marc emailed me the position and the moves, and asked if I would link it into one of my javascript pages on the Gothic Chess site, and I said "Sure, no problem."
5.0 Ghz Intel Gulftown Supercomputers
http://www.liquidnitrogenoverclocking.com
User avatar
Ed Trice
 
Posts: 58
Joined: Thu Oct 28, 2010 1:13 am
Location: Henlopen Acres, Delaware


Return to Endgame Tablebases

Who is online

Users browsing this forum: No registered users and 1 guest