Hi,

My abstract strategy game AI implements negascout in an iterative deepening framework, using zobrist keys and killer moves. I use bitboard representation for my board.

However, when running a version of the game on Android devices, I can only search to a depth of 4, because starting with 5 the search starts taking more time (depth of 5 takes 3-4 seconds per move).

I would like to at least go to depth 5 so I was thinking if it would make sense to store the score of previous game moves in a sort of table and then quickly access them. If the score is not found, the normal eval function kicks in.

Does this make sense, and more importantly, will it bring performance benefits?

Thanks

You are not being precise enough about the mechanism you are proposing, but it could either be an opening book or rote learning. These things do help, but it's also possible that you are missing some low-hanging fruit to make your search much faster (better move ordering, better hashing, reduction of unpromising moves...). It's also possible that your game is not well suited for alpha-beta search, no matter what you do, and a completely different approach might be needed, perhaps MCTS.

It's very hard to provide more relevant advice without knowing anything about your game.

My move ordering sorts by: transposition move, capture move, primary killer, secondary killer in descending order. It also adds the history heuristic to each move before sorting. I currently don't implement move reduction. Maybe I should consider this too.

my game consists of a 6x4 board and 8 pieces per player. All pieces are of the same value (ie. no king, queen, etc). And the game is just moving pieces from one square to another. Something similar to checkers.

I added LMR into my current code. Does this make sense?

                int LMR_DEPTH = 3;
                int LMR_MOVENUMBER_MINIMUM = 3

		for (int i = 0, n = moves.size(); i < n; i++) 
		{
			moveCount++;
			Move move = moves.get(i);

			//Begin Late Move Reduction search
			boolean reduced = false;			
			if (movesSearched >= LMR_MOVENUMBER_MINIMUM 
					&& depth >= LMR_DEPTH
					&& ply < depth)
			{
				depth--;
				reduced = true;
			}		
			//End Late Move Reduction search
			
			//PV has been found
			if (alpha < bestScore) 
			{ 
				alpha = bestScore; 
				foundPV = true; 
			} 
			
			//Make the move
			board.make(move, true);
			
			//Search
			if (foundPV) 
			{
				//Zero window
				val = -negaScout(-alpha - 1, -alpha, ply + 1, depth);
				if (val > alpha && val < beta) 
					val = -negaScout(-beta, -alpha, ply + 1, depth);
			} 
			//PV search
			else	
				val = -negaScout(-beta, -alpha, ply + 1, depth);
			
			//Begin Late Move Reduction research
			if (reduced && val >= beta)
			{
				depth++;
				val = -negaScout(-beta, -alpha, ply + 1, depth);
			}
			//End Late Move Reduction research
			
			//Undo the move
			board.undo(move);
			
			//Cut-offs here...
		}

I'm sorry I don't have time to look at your code right now. But let me just say that I found easier to implement LMR on top of PVS (I believe, "NegaScout" is essentially another name for PVS).

What's your typical branching factor? If your game is anything like checkers, you should get way more than depth 4 in a few seconds.

Can you post a link to the rules?

Have you done any profiling on the code to see if there are particular areas that are bottlenecking?

Move the constructor for the working variable 'move' outside of the loop

depending on what all the constructor does along with the allocation/deallocation each iteration - any/all of that is wasted processing