Edsel Apostol wrote: If I understood what you want to convey you're basically saying that those moves with a presumed bigger subtree should be put in the end of the movelist? But isn't it the kind of moves with the bigger subtree that will have a higher chance of improving the best score or alpha of the current position? Also those moves that are easily refuted means that they are foolish moves, isn't it only logical to search them last?
Suppose you have 30 moves to try fro a ply, of theese you have three moves A, B, C.
Move A is the best. Moves B and C _seems_ good moves either but are not the best.
The remaining moves are easy to prune.
If I think of a 'magic' algorithm for ordering the moves, then this magic algorithm should order in this way:
A, all remaining moves, B, C
So that the best is tried immediately and then the remaining are quickly pruned and then at the end are tried the most troubled ones, i.e. the moves that you need a lot of time to search but don't get you to anywhere.
Of course nobody has this magic alghortim, but we can think of something that approximate this in practical terms.