Ø  Player receives Blackjack – The player should flip the cards and you can count them.

Ø  Player doubles – Player shows first two cards and you count them. You can infer nothing about the down card if it is down.

Ø  Player splits – All cards seen and counted.

Ø  Player busts – All cards seen and counted.

Ø  Player stands pat – You see no cards.

Ø  Player draws one or more cards and does not bust – you see the hits, but not the two original cards.

The Process

OK, we need to think about the last two cases. If the player stands pat, he is more likely to have high cards, particularly if the dealer has a seven through ace. If the player draws and does not bust, he more probably has low cards, particularly if the dealer has a low card up. Let’s look at some stats created by the CVData Counting by Inference strategy generator for Hi-Opt II:

Rows are by dealer upcard. The four Columns contain data for a player pat hand, and three-, four- and more-card final hand lengths. Unbusted hands where the original two cards are not seen are the only hands included. The data in the cells are the counts of the cards that are hidden. So, if the player has a pat hand, and the dealer has an ace up, then the two hidden cards have a Hi-Opt II count of -2.3 on average. (Hi-Opt II counts tens as -2.) Not a surprise given that the player would not have stood pat against an ace with low cards. If the player stops at three cards against a dealer 6, on average the hidden cards have a count of 2.2. Let us simplify this and create the CVData strategy table seen on the next page.

 © 2009 Norman Wattenberger