value += (random*granularity)-(granularity/2)

Quote: Alvaro : the first solution you suggest won't work for me :-( the reason is that, first of all, I don't have and cannot have an opening book : I've invented the game and I still don't know how to play it well. worse : no one knows how to play it any better since I'm the only one to know the game ! so I don't know any convenient openings (at least not now). So no randomnes can come from the opening. Second, the reason I am implementing the randomnes is that I want the computer to play against itself ... both oponents will therefore be using the same "intelligence" ... the only thing that could make things different during two games would be to force the computer to think for a random amount of time, which doesn't seem fair and won't garanty the outcome to be fair play.

Quote: Your second suggestion, however surprising, is very attractive ! And easy to implement :-) I'm surprised it works !! Do you suggest something like

value += (random*granularity)-(granularity/2)

? I'll try that.

Quote: Concerning your third suggestion, I have to say I didn't understand it ... that's because I'm new to AlphaBeta and search windows ... But do you think it will give better results than what I propose to do using Talyssan's solution with PVS ?

Quote: I don't mean that writing an opening book doesn't require work. You need to learn how to play the game and then decide what are good opening moves. If the initial position of your game is always the same, it doesn't make sense to start thinking on that position when you are playing "live". You should have thought about what you want to do in that position beforehand, and store the results in an opening book.

Quote:
The main transposition table is an array of hash elements. Each hash element looks something like this:

#define    hashfEXACT   0#define    hashfALPHA   1#define    hashfBETA    2 typedef struct tagHASHE {    U64 key;    int depth;    int flags;    int value;    MOVE best;}   HASHE;

Quote:
The final field in my hash element is the "best" move encountered last time I searched this position. Sometimes I don't get a best move, like if everything failed low (returned a score <= alpha), but other times there is a definite best move, like when something fails high (returns a score >= beta).

If a best move is found, it will be searched first.

Some sample code, overlaid onto an alpha-beta function :

int AlphaBeta(int depth, int alpha, int beta){    int hashf = hashfALPHA;     if ((val = ProbeHash(depth, alpha, beta)) != valUNKNOWN)        return val;    if (depth == 0) {        val = Evaluate();        RecordHash(depth, val, hashfEXACT);        return val;    }    GenerateLegalMoves();    while (MovesLeft()) {        MakeNextMove();        val = -AlphaBeta(depth - 1, -beta, -alpha);        UnmakeMove();        if (val >= beta) {            RecordHash(depth, beta, hashfBETA);            return beta;        }        if (val > alpha) {            hashf = hashfEXACT;            alpha = val;        }    }    RecordHash(depth, alpha, hashf);    return alpha;} int ProbeHash(int depth, int alpha, int beta){    HASHE * phashe = &hash_table[ZobristKey() % TableSize()];     if (phashe->key == ZobristKey()) {        if (phashe->depth >= depth) {            if (phashe->flags == hashfEXACT)                return phashe->val;            if ((phashe->flags == hashfALPHA) &&                (phashe->val <= alpha))                return alpha;            if ((phashe->flags == hashfBETA) &&                (phashe->val >= beta))                return beta;        }        RememberBestMove();    }    return valUNKNOWN;} void RecordHash(int depth, int val, int hashf){    HASHE * phashe = &hash_table[ZobristKey() % TableSize()];     phashe->key = ZobristKey();    phashe->best = BestMove();    phashe->val = val;    phashe->hashf = hashf;    phashe->depth = depth;}

Quote: Original post by sgalzinI don't understand where the code corresponding to "search the best move first" is. Actually, I'm not sure what to do with the RemeberBestMove() method : should I just save the current move into a variable, say bestCurrentMove ?