| View previous topic :: View next topic |
| Author |
Message |
Jorge Pichard
Joined: 09 Mar 2006
Posts: 2557
|
|
| Back to top |
|
 |
Heiner Marxen
Joined: 08 Mar 2006
Posts: 39
Location: Germany, Berlin
|
 Posted: Fri Apr 27, 2012 10:06 pm Post subject: Re: Has anybody tried the Network Truncation Algorithms |
  |
|
| pichy wrote: |
| http://www.antd.nist.gov/wctg/netanal/EvalNetRel.pdf |
Excuse my ignorance... but how does this apply to chess programs?  |
|
| Back to top |
|
|
Mel Cooper
Joined: 16 Dec 2010
Posts: 113
|
| Posted: Sat Apr 28, 2012 7:56 am Post subject: Re: Has anybody tried the Network Truncation Algorithms |
|
|
This is interesting as an analogy, substituting chess concepts for network ones.
However, from p. 9 one can read:
"The networks of interest are tactical radio networks, represented by directed graphs with imperfect vertices (nodes) and edges (links). "
I can substitute an imperfect score with a node defining a chess position for an "imperfect vertex" but what would be an imperfect link?
In chess a move is either legal or not, no fuzziness here and so, IMHO, the analogy and transferability of algorithms devised for network reliability analysis to chess just breaks down.
_________________
Per ardua ad astra |
|
| Back to top |
|
|
Marco Costalba
Joined: 14 Jun 2008
Posts: 2090
|
|
| Back to top |
|
|
Mel Cooper
Joined: 16 Dec 2010
Posts: 113
|
| Posted: Sat Apr 28, 2012 9:30 pm Post subject: Re: Has anybody tried the Network Truncation Algorithms |
|
|
Looks tastier than the network recipe
But don't forget that when she says 450 degrees for 40 minutes in the oven
she means Fahrenheit degrees or circa 220 Celsius 
_________________
Per ardua ad astra |
|
| Back to top |
|
|
Marcel van Kervinck
Joined: 26 Feb 2010
Posts: 348
|
| Posted: Sat Apr 28, 2012 10:50 pm Post subject: Re: Has anybody tried the Network Truncation Algorithms |
|
|
howdini.com, isn't that recipe a rip-off? |
|
| Back to top |
|
|
Ricardo Barreira
Joined: 27 Apr 2010
Posts: 821
|
| Posted: Sat Apr 28, 2012 11:01 pm Post subject: Re: Has anybody tried the Network Truncation Algorithms |
|
|
| melajara wrote: |
Looks tastier than the network recipe
But don't forget that when she says 450 degrees for 40 minutes in the oven
she means Fahrenheit degrees or circa 220 Celsius |
Much closer to 230 than 220. |
|
| Back to top |
|
|
Jorge Pichard
Joined: 09 Mar 2006
Posts: 2557
|
| Posted: Sun Apr 29, 2012 12:35 am Post subject: Re: Has anybody tried the Network Truncation Algorithms |
|
|
| hMx wrote: |
| pichy wrote: |
| http://www.antd.nist.gov/wctg/netanal/EvalNetRel.pdf |
Excuse my ignorance... but how does this apply to chess programs? |
It apply to chess programming when you combine it with a game tree search method which is more brute-force and less selective than other popular approaches. |
|
| Back to top |
|
|
Jorge Pichard
Joined: 09 Mar 2006
Posts: 2557
|
| Posted: Sun Apr 29, 2012 1:31 am Post subject: Re: Has anybody tried the Network Truncation Algorithms |
|
|
[quote="pichy"]
| hMx wrote: |
| pichy wrote: |
| http://www.antd.nist.gov/wctg/netanal/EvalNetRel.pdf |
Excuse my ignorance... but how does this apply to chess programs? |
It apply to chess programming when you combine it with a game tree search method which is more brute-force and less selective than other popular approaches, some chess programs determine the material balance of a position by a static analysis of all possible capture-move sequences. These routines are often referred to as SOMA (“Swapping Off Material Analyzer”). An enhanced SOMA algorithm with chess-specific features. The general operation of SUPER-SOMA is to calculate a table of all single-square threats and then iteratively cycle through this table, ordering the priority of each tactical move, choosing a move and then updating the table.
The idea is best explained by considering captures around one single square. Cycling through the table produces all moves that are involved in the capture. These moves are usually ordered by their material value. Since pins may be involved, the table should
be updated continuously in order to produce legal moves. One further step is to take into account features such as ‘overloading’. If the first capturing piece is an overloaded Knight and the second piece is a free Rook, then the capture sequence may be re-ordered.
The SUPER-SOMA algorithm allows separate exchanges on the board to be linked and prioritized. In complex positions, with many interrelated exchanges on several squares, the cycle is repeated until neither side can play a beneficial move. |
|
| Back to top |
|
|
Lucas Braesch
Joined: 31 May 2010
Posts: 1732
|
| Posted: Sun Apr 29, 2012 1:38 am Post subject: Re: Has anybody tried the Network Truncation Algorithms |
|
|
+1 |
|
| Back to top |
|
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
