[UATrewqaz]

This seems plausible, the way I'm seeing it in my head

You enter a range of hands (say 1K, 5K, 10K, whatever)

It then goes through and analyzes the card distribution. It basically then does a rough EV calculation for each hand, based on either predetermined numbers or user custom numbers.

(AKA: For ever AA I expect +2.8 BB, for ever KK I expect + 2.0 BB, etc.) and it would also calulate junk hands being -EV.

It would then show you what your pure EV results are for the set of hands. Thus a set of hands jam packed with big pocker pairs would have a higher EV than one with just a bunch of junk.

This is PREFLOP analysis only though and does not factor in how you play the actual hands of course.

The only other idea I had was a "What's he got?" predictor

Basically it would use opponent modeling combined with current hand activity and board information and show you a % break down of possible holdings.

For example: A 40% VPIP / 3% PFR guy check raises your QQ on a board of TT3.

The program would then find similiar VPIP/PRF people and similiar situations (I have an overpair to a paired board, when checkraised by this type of player what % of the time do they have trips?)