I haven't gone through my 6-year old statements, but wonder where I said exactly that.

There is a world of difference between an "algorithm" that can made legal moves, and one which makes good moves.

<Computer chess is quite simple. At high level the algorithm is basically For each legal move from the current position:
Calculate material balance
Repeat to desired search depth.>

Such a program could just as well play h3,a3,b2 as the first, second, third move.

<If you are unconcerned with optimizations, it could surely be coded in about a hundred short lines.

If you are unconcerned with readability, it could be stuffed into eight or ten lines of obfuscated perl.>

You start your posting with talking about how long _time_ it takes to write programs. Then you talk about compressing/obfuscating them.

Any programmer knows that this takes more time to do, not less - so it appears not to be relevant.

<Such a chess program would be fairly easy to beat positionally>

Such a program would be incredibly easy to beat for a human that had played chess for less than a week. It would not make a single sensible move, apart from saving material if it could see it being lost.

<...since it would ignore important but less tangible factors such as piece activity, pawn structure, etc. While true positional understanding is beyond the reach of computers, a very convincing approximation can be achieved by throwing in a few simple heuristics (like a minus score for backward pawns), and simple brute-force searching.>

Ah, we are going a bit beyond the handful of lines, aren't we?

Trying to make anything worthwhile will take considerable time. Chess programmers have in total dozens of man-years on developing programs.

<Computer chess is considered very old hat by computer scientists.>

An truly odd statement. Can you bring any quotes?

<Everyone and his dog has written a chess engine>

Despite spending 5 years in AU, Denmark, in the Computer Science faculty, I don't know any one who actually carried out this task, despite a number of them being most interested in chess. The general opinion was, that to get it anyway further than legal moves would eat enormous time.

<..and judging by the attitudes of certain posters here, i may have to write one myself to prove the point.>

Which you didn't exactly get done, I assume? Even less so one that actually made sensible moves.

- - -

This thing started with my claim that Turing has not written an algorithm where you'd feed a position into, and that it would return these moves, simply by doing calculations as computers can do.

Anyone who knows the very least about Turing's work, and computer science history, would write it off as, yes, ludicrous.

<1. Mobility: For the queen, rooks, bishops, add the square roots of the number of moves that the piece can make, counting a capture as two moves.

2. Piece safety: For the rooks, bishops and knights add 1 point if there is one defender and 1.5 if there is more than one

3. King mobility: For the king use the same method of scoring as for the piece, but do not count castling

4. King safety: Deduct points for the king's vulnerability, defined as the number of moves that a queen could make were it on the square of the king

5. castling: add 1 point if castling is still legally possible after this move. Add another point if castling is immediately possible or if a castling move has just been made

6. Pawn credit: score .2 points for each rank advanced and .3 points for each pawn defended by one or more non pawns

7. Check and mate threats. Score 1 point for the threat of mate, and .5 points for a checkmaterial values used to each of the pieces were:pawn =1, knight=3, bishop=3.5 rook =5, queen=10 >

With the knowledge he had at the time, would he be able to determine whether the W move were chosen solely by an algorithm or was there some human intervention?

http://www.spiegel.de/static/flash/...

- has additional annotations.

It is not certain if these annotations are provided by Turing himself or by a third party at some later point, although the tone is reminiscent of Turing.

The German chess piece names there used are -

K = König = King
D = Dame = Queen
T = Turm = Rook
L = Läufer = Bishop
S = Springer = Knight

The article was in a pretty obscure book, but it is reproduced on a page at chessbase where Matthias Feist writes about programming a 'Turing Engine' for Fritz using Turing's position evaluation rules.