SOMA

[rebel777]

Hi Gerd,

I am not aware of SOMA, let alone his work. Would be fun indeed to know what he invented.

About ancient history, did you know that Don Beal in Cologne 1986 played with Bcp that used null-move in QS? I wish I had paid more attention back then when he explained

Ed

[Carey]

Hi Gerd,

I am not aware of SOMA, let alone his work. Would be fun indeed to know what he invented.

About ancient history, did you know that Don Beal in Cologne 1986 played with Bcp that used null-move in QS? I wish I had paid more attention back then when he explained

Ed

I have a copy of Don Beal's BCP source labeled as 1986. I've never studied the source so I don't know if that has the null move in QS stuff, though.

He originally gave it to me to put on my classic chess programs web site but the site is currently down. (I had hosted it on Google's web site but they've gotten rather snippity and have decided not to allow javascript. Since my web site was based on TiddlyWiki, which was 100% javascript, my site was not longer allowed.)

Anyway, since you did mention it, if anybody wants a copy, I can email it to anybody that asks.

Carey

[Gerd Isenberg]

Hi Gerd,

I am not aware of SOMA, let alone his work. Would be fun indeed to know what he invented.

About ancient history, did you know that Don Beal in Cologne 1986 played with Bcp that used null-move in QS? I wish I had paid more attention back then when he explained

Ed

Hi Ed,

SOMA was a early chess program by M.Smith from 1961?! According to Bartel, Kraas, Schrüfer 1985 Das grosse Computer Schach Buch, it had an one-ply search without quiescence.

Jeff Rollason wrote in his SUPER-SOMA paper (2000): SUPER-SOMA - Solving Tactical Exchanges in Shogi without Tree Searching:

The core of the proposed work is to attempt to predict exchanges across the whole board. There is already an algorithm called SOMA that does this (Michie, D - 1966), but only considers each square in isolation. This does not allow whole board situations to be assessed. SUPER-SOMA needs to find a mechanism to cross-reference these potential exchanges in such a way that sequences of moves from different parts of the board can be predicted.

I didn't find any other references in the Web about SOMA (the program as well the algorithm) so far.

On Don Beal, yes I was aware of his paper on Null Move, but didn't know exactly that it was used in Bcp's QS. When he explained it in Cologne, you were probably a bit preoccupied, if it was before the final round

Gerd

[rebel777]

On Don Beal, yes I was aware of his paper on Null Move, but didn't know exactly that it was used in Bcp's QS. When he explained it in Cologne, you were probably a bit preoccupied, if it was before the final round

Busy with other things indeed. I still remember cause Frans Morsch kept on telling me during the tournament, "Ed, there is big potential in null-move, there must be something, I can't put my finger on it, but there is something...". And I did not listen. And it was exactly Frans who years later as first one had success with null-move and dominated a couple of years the cc-scene until others catch up. Null-move back then was used by some in the last plies, but Frans discovered that the true strength of the idea comes from the recursive use in the whole tree.

Ed

[Bill Rogers]

It sounds to me that SOMA is very much like what the Spaklins used in their very first chess program, Sargon. It was a static exchange evaluator and they published it in thier book. It is basically what I used in my first chess program which was a one ply search only program although I did not go as deep as the Spacklins did in their program.
In essence what they did was to keep finding attacker and defenders on one square until no more could be found.
Bill

[Carey]

Gerd,

You probably already have this info, but I came across a couple pages about it in David Levy & Monty Newborn's "How Computers Play Chess".

They call it an "analyzer" since it didn't do much searching.

It talks a bit about the eval, how mobility & attacks were calculated & scored, etc.

It only briefly mentions the SwapOff routine and doesn't mention any specifics.

They also give one game it played against Machiavelli (both done by hand simulation), which was an analyzer by D. Michie and S. Wylie some years before.

Not really a lot of info, however, obviously David Levy & Monty Newborn were familiar with it and had references that discussed it.


As for mentioning Rolf Smith & Frank Ceruti... I haven't been able to get in touch with Mr. Smith in a couple years. He had found an early printed copy of his program, but I didn't hear from him after that. No replies to any of my emails. But I haven't tried again in a year or so.

[Gerd Isenberg]

On Don Beal, yes I was aware of his paper on Null Move, but didn't know exactly that it was used in Bcp's QS. When he explained it in Cologne, you were probably a bit preoccupied, if it was before the final round

Busy with other things indeed. I still remember cause Frans Morsch kept on telling me during the tournament, "Ed, there is big potential in null-move, there must be something, I can't put my finger on it, but there is something...". And I did not listen. And it was exactly Frans who years later as first one had success with null-move and dominated a couple of years the cc-scene until others catch up. Null-move back then was used by some in the last plies, but Frans discovered that the true strength of the idea comes from the recursive use in the whole tree.

Ed

and Frans later told it Chrilly ...

[Gerd Isenberg]

It sounds to me that SOMA is very much like what the Spaklins used in their very first chess program, Sargon. It was a static exchange evaluator and they published it in thier book. It is basically what I used in my first chess program which was a one ply search only program although I did not go as deep as the Spacklins did in their program.
In essence what they did was to keep finding attacker and defenders on one square until no more could be found.
Bill

Sure, but there is still the question about the program of M. Smith called SOMA from 1961. And the SOMA algorithm by Donald Michie from 1966! Looking for references.

[Gerd Isenberg]

Gerd,

You probably already have this info, but I came across a couple pages about it in David Levy & Monty Newborn's "How Computers Play Chess".

They call it an "analyzer" since it didn't do much searching.

It talks a bit about the eval, how mobility & attacks were calculated & scored, etc.

It only briefly mentions the SwapOff routine and doesn't mention any specifics.

They also give one game it played against Machiavelli (both done by hand simulation), which was an analyzer by D. Michie and S. Wylie some years before.

Not really a lot of info, however, obviously David Levy & Monty Newborn were familiar with it and had references that discussed it.


As for mentioning Rolf Smith & Frank Ceruti... I haven't been able to get in touch with Mr. Smith in a couple years. He had found an early printed copy of his program, but I didn't hear from him after that. No replies to any of my emails. But I haven't tried again in a year or so.

Hi Carey,
no, I don't have that book. You think M. Smith (SOMA program 1961, name literally from Bartel, Kraas, Schrüfer 1985 "Das grosse Computer Schach Buch", 'M' a typo or missing r for Mr.) might be Rolf C. Smith, co-author of SCHACH, which competed the early ACM-events from 1970? Yes, makes some sense, if I read the description of SCHACH. Could really be that Rolf C. Smith had an earlier program SOMA.

Schach was written in Fortran and already introduced the concept of Static Exchange Evaluation for Move Ordering and Pruning:
The backbone of SCHACH is the concept of piece board control, defined as all squares on which a piece exerts direct or indirect influence (can move to in a capture mode). Utilizing this concept we have found that a pseudo-dynamic position projection can be effected in a static environment on a local scale.

Can you (or anybody else) identify these people from Genie vs Schach at the 2nd North American Computer Chess Championship in Chicago, Illinois?

(not the German program Schach btw.)

Cheers,
Gerd

[Carey]

Sure, but there is still the question about the program of M. Smith called SOMA from 1961. And the SOMA algorithm by Donald Michie from 1966! Looking for references.

For Michie, how about checking "On Machine Intelligence", 1974, Donald Michie.

In the Spracklen's article on writing Sargon, they say about that book:

Michie's book provides an excellent treatment of exchange evaluation. He uses the concept of an exchange polynomial for accurately determining the outcome of battles engaged on the board. The basic approach we used in XCHNG, the Sargon exchange evaluator, turned out to be surprisingly similar. Sargon's approach, however, is far less computationally complex. We highly recommend this reference to anyone planning to write a chess program without look-ahead.

The date is later, of course, but it sounds a lot like it would be related to the original SOMA stuff.