chess engine

for the last few months, I have been trying to create a moderately strong
chess engine that would run on Winboard or Fritz. My latest effort was not a
happy one. Can you please help me?

"Christopher Potter" <chris.potter5@bigpond.com> wrote in message
>
> for the last few months, I have been trying to create a moderately
> strong chess engine that would run on Winboard or Fritz. My latest
> effort was not a happy one. Can you please help me?
>
What is Winboard or Fritz? A chess engine should be platform-independent.
There must be lots of engine written in C which you can grab from the net.
However if you want your own engine, the heart of it is a function to
evaluate the strength of a position. Obviously checkmate is the strongest
position, whilst you can calculate material strength, centre squares
controlled, and so forth to make the function more sophisticated.

If your strength evaluator was sophisticated enough, you could simply
evaluate every legal move and have the best answer. In fact, it is not
possible to assess positions that well, so you need a tree of possible
moves.
Since there are twenty to forty legal moves in each position, a greedy
algorithm cannot get very far before it runs out of time. Depending how
strong you want your program to be, you can prune unpromising moves and only
investigate the paths that contain moves that would actually be played -
i.e. those leading to the stronger positions.

"Malcolm" <malcolm@55bank.freeserve.co.uk> wrote in message
news:bjnp7l$uob$1@newsg1.svr.pol.co.uk...
>
> "Christopher Potter" <chris.potter5@bigpond.com> wrote in message
> >
> > for the last few months, I have been trying to create a moderately
> > strong chess engine that would run on Winboard or Fritz. My latest
> > effort was not a happy one. Can you please help me?
> >
> What is Winboard or Fritz? A chess engine should be platform-independent.
> There must be lots of engine written in C which you can grab from the net.
> However if you want your own engine, the heart of it is a function to
> evaluate the strength of a position. Obviously checkmate is the strongest
> position, whilst you can calculate material strength, centre squares
> controlled, and so forth to make the function more sophisticated.
>
> If your strength evaluator was sophisticated enough, you could simply
> evaluate every legal move and have the best answer. In fact, it is not
> possible to assess positions that well, so you need a tree of possible
> moves.
> Since there are twenty to forty legal moves in each position, a greedy
> algorithm cannot get very far before it runs out of time. Depending how
> strong you want your program to be, you can prune unpromising moves and
only
> investigate the paths that contain moves that would actually be played -
> i.e. those leading to the stronger positions.
>
>

But from what i've heard, pruning actually weakens the engine