Casino Verite Software

The Casino Verite' software in both H17 or S17 calls for the player to hit a 12 v 4 (and lists it as a $.02 error if one stands). Clearly, 12v4 is a very close call. The Expected Return for Every Play in Appendix 1 shows w/an S17 Infinite deck that the ER Standing is -0.211 and the ER Hitting is -0.2135.
Fact is, Mr. Wattenberger truly knows his stuff so I immediately conclude that I must be missing something.
Now I understand that any negative count would drive a player to hit a 12 v 4; and, I like the Wizards suggestion 'Oh, there is 12 vs 4. Scan the table and hit it if there are more face cards out than 6-9s.'
I just don't get the difference in approaches between the software and the charts that I use from the Wizard.
Can anybody explain the difference?

Looking at the composition-dependent return tables for 6-decks (S17), it seems that hitting is the better choice for 10,2 vs 4, while standing is the better choice for all other 2-card compositions (9,3 8,4 and 7,5) as well as presumably all 3-card and more compositions. I think overall, taking in to account the frequency of the compositions, standing is the better play. It seems that many have come to this conclusion, including the Wizard with his basic strategy. I am not sure why Casino Verite is bucking the trend, but perhaps they only factor in the 10,2 composition.
Also I think you are referencing me in that quote, here is the thread for the benefit of other readers:
3 card or more 16 vs dealer 10, 9, 8, or 7
The wizard suggests a similar method for 16 vs 10. Note that both of these 'table counts' will only reduce the house edge by a small amount.

The cheat sheets are based on sums of player cards, not individual compositions.
------------------------
Appendix 9 on the blackjack section of the Wizard of Odds has each composition dependent expected value.
1 deck, dealer hits on soft 17(Dealer 4) : The expected values for each move are:
Player 10+2=12,stand=-0.206776, hit=-0.194247 , player should hit [probability = 0.386124%]
Player 9+3=12, stand=-0.184834, hit=-0.225837 , player should stand [probability = 0.096531%]
Player 8+5=12, stand=-0.175045, hit=-0.292257 , player should stand, probability the same
Player 7+6=12, stand=-0.152149, hit=-0.277486 , player should stand, probability the same
Player sum=12, stand=-0.191305, hit=-0.224653 , player should stand
You have to multiply 10+2 results by 4/7 because it is more 4 times more likely than the other combination 9+3, 8+5, and 7+6. Multiply the other ones by 1/7 and add them up. Over the range of results where the sum=12 the player should stand.
If you look up 12 vs 2 for different sets of rules you will find that often you should hit 10 & 2 vs a 4. See Fine points of basic strategy (Appendix 3) Two cards exception rule #5.
------------------------
10,2 vs 4 hit
J, 2 vs 4 hit
Q, 2 vs 4 hit
K, 2 vs 4 hit
9, 3 vs 4 stand
8, 4 vs 4 stand
7, 5 vs 4 stand
6, 6,vs 4 split
So there are 4 combinations to hit, and 3 combinations to stand.
The software must be using composition dependent calculations, and not dependent on a simple sum. It may go further by using the cards that have already been played.

The only major thing missing is the 3-card and up compostitions (eg 3,4,5 or 2,6,4,2). And although they are not listed, I am almost positive every single one of those will favor standing. This tips the scales in standings favor on the total-dependent basic strategy. Overall, without knowing any other information except the hand total and the upcard, standing is the marginally better play.
To sum up, the only situation where hitting is better is 10,2. But the EVs behind all the other compositions together (of which there are many) outweigh the 10,2 hand.
I think its very likely that casino verite's charts and numbers are based on specific counts. Technically, at a count of exactly 0 (neutral), the player should hit. We can see that ALL compositions that favor standing put the running count at +2 or above. For example, the hand 9,3 gives a count of +1 (actually +2 with the 4 upcard). Now if we are told first we actually have a count of exactly 0, it means there must be at least two high cards out on the table to compensate for the upcard and a low card which must complete the players hand of 12. Or alternatively, the count must have been negative going in to the hand. Thus, in this situation, it is always better to hit, and not doing this has a cost.
Therefore if we have a hand of 12 vs 4, it is more likely than not that the count is positive rather than neutral and we should stand.
I think this is a case of neither figures being wrong - they are both right. For a basic strategy player, you should stand. For a counter faced with the decision at a perfectly neutral running count, you should hit.

Excellent analysis (pocketaces, please excuse my mis-attribution) and usable/practical information.
Thank you.

'It is impossible to begin to learn that which one thinks one already knows.' -Epictetus

I have it set to REKO (6 Deck) with a running count so that all impacted indexes read 22. Despite the fact it reads 22, all practice drills require a hit showing a $.02 error if you stand. For what it's worth, 12 v 2, and 12 v 3 read the exact same.
The Basic Strategy table reads to stand with a 12 v 4.
Now, this is no biggie--if anything, this discussion has deepened my understanding of a composition dependent approach with the 12 v 4.
Thanks for asking
By the way, the software is excellent (I practice daily) as is the on line literature.