The pre-flop deception level lets you choose how much randomization it will use in pre-flop actions. If you set the slider towards honest it will raise with strong hands, call with moderate strength hands, and fold everything else. With high deception (tricky) it will have a higher chance of calling strong hands and even some very weak hands, and will raise some moderate hands.
The second slider controls what types of actions the simulated opponents will do during the trails. If set to Call, the simulation becomes a basic all-in equity simulation. If the slider is moved towards Predict then it will use its opponent models to try and predict what each opponent would do in each simulated situation.
Improved Opponent Modeling in Poker.
Aaron Davidson, Darse Billings, Jonathan Schaeffer, and Duane Szafron,
Proceedings of the 2000 International Conference on Artificial Intelligence (ICAI'2000), Las Vegas, Nevada, pp 1467-1473, 2000.
http://www.cs.ualberta.ca/~darse/Papers/ICAI00.htmlUsing Probabilistic Knowledge and Simulation to play Poker.
Darse Billings, Lourdes Peņa, Jonathan Schaeffer, and Duane Szafron,
Proceedings of the Sixteenth National Conference of the American Association for Artificial Intelligence (AAAI-99), 1999.
http://www.cs.ualberta.ca/~darse/Papers/AAAI99.html