Blackjack Therapy by Clarke Cant
Chapter 11, the Uston+/-, Zen, and Victor APC count indexes.
All indexes reflect my opinions on index selection and a personal preference to drop indexes of less than 50% correlation with effects of removal. I have not recalculated numbers for late surrender in the Zen and Victor APC counts as I have for the Uston +/-.
I have not attempted any i18 condensation either. This is best done individually for each count using the tables in Theory of Blackjack, by Peter Griffin (newer editions preffered for this), and following these rules:
Select an index with the highest product of correlation and importance. Keep multiple indexes that have the same value or can be rounded to the same values. Most tries at more accuaracy yield little more gain. Gain for true count point is best found by dividing the sum of squares of your point count by the inner product, as found in the third chapter. True edge indexes, in my opinion, lead one to understate this gain per true count point, and lead to drastic underbetting of some spreads. Infrequent indexes have an element of cover, in situations where playing just basic strategy makes it more noticeable that you may be a counter. All charts assume the dealer stands on all 17s initially, then give indexes if there is a change when the dealer hits soft 17s. * Indicates consulting a detailed basic strategy might be recomended. DA$ indexes are omitted. Indexes are single deck with some compromise for multideck.
The Uston +/- count performs nearly exactly as well as the hi-lo count. Difference can be good, if the pit boss uses hi-lo….
2 3 4 5 6 7 8 9 X A
count values 0 1 1 1 1 1 0 0 -1 -1
your cards dealer upcards
s18 h h .*
17 -8
16 -11 -11 -13 -17 s h h h 0 7
15 -7 -7 -9 -12 -13 h 14 9 3 8
14 -4 -4 -6 -8 -8 14 15 12 6
13 -1 -1 -3 -5 -5 14
12 3 3 1 -1 0 20
11 -16 -16 -17 d d -12 -8 -6 -4 -1
10 -11 -11 -12 -16 -17 -7 -5 -2 3 2
9 0 0 -3 -6 -7 4 9
8 11 9 6 3 3 13
7 h 17 13 12 14
A9 8 8 6 5 5
A8 6 3 2 0 0
A7 1 -1 -6 -8 -10
A6 * -3 -7 -16 -16
A5 h h -2 -8 d
A4 h h -2 -10 d
A3 h 5 -2 d d
A2 7 3 -1 d d
AA $ $ $ $ $ -13 -12 -12 -11 -9
XX 10 9 6 5 6
99 -2 -1 -3 -5 -3 5 -7 -7 s 2
88 $ $ $ $ $ $ $ $ <5 -14
77 -15 -12 $ $ $ $
66 0 -1 -4 $ $
33 15 4 -3 -7 $ $
22 h * -3 -6 $ $
Insurance 1 deck 1.5; 2 decks 2.5; all shoe games 3
dealer hits soft 17s
your cards upcards
6 A late surrender
s18 7 9 X A h17 A
17 -7 17 <-4 <1
16 -17 2 16 0 -4 -4 -7
15 -15 4 88 1 *
14 -10 6 15 3 0 0 -2
13 -7 10 14 6 2 4 2
12 -3 16 1-deck 77 4 0 2 0
A9 4
A8 -1
A7 -12
A6 -12
AA -10
XX 5
99 -5 0
88 -4
The Zen count here gives the original values and the original indexes. Anyone using the values from Blackbelt in Blackjack has been using these (I did them for myself, and then sent an error sheet to Snyder that had those I felt were wrong in his initial printing.) indexes for some time probably. You might get a kick out of knowing, that they were done using a Timex/Sinclair 1000 computer and the Algebraic Approximation paper written by Snyder, originally.
2 3 4 5 6 7 8 9 X A
1 1 2 2 2 1 0 0 -2 -1
s18 -4
17 -13
16 -16 -18 -21 -25 -21 h h h -1 12
15 -9 -12 -14 -17 -17 23 20 13 5 14
14 -5 -7 -9 -12 -12 24 27 20 12 18
13 -1 -3 -5 -8 -7 25
12 6 4 1 -2 0 37
11 -24 -26 -28 -32 -35 -17 -12 -9 -7 -2
10 -18 -20 -22 -26 -28 -12 -8 -3 5 5
9 1 -1 -4 -9 -11 7 15
8 20 14 11 6 5 29
7 h 28 22 19 21
A9 16 14 12 9 9
A8 13 6 4 2 1
A7 2 -2 -11 -13 -13
A6 * -5 -9 -23 -24
A5 h 6 -2 -11 -26
A4 h 6 -2 -12 d
A3 16 6 -2 -11 d
A2 13 6 1 -6 d
AA -20 -18 -18 -18 -13
XX 17 14 11 8 9
99 -2 -3 -5 -7 -5 11 -14 -13 s 4
88 <21 <7 -24
77 -26 -23 $ $ $ $
66 2 -1 -6 $ $
33 23 5 -5 -11 -13 $
22 h * -4 -7 -12 $
Insurance 1 deck 3; 2 deck 4; shoes 5
dealer hits soft 17s late surrender
6 A 9 X A
S18 12 16 1 -4 -1
17 -11 15 5 -1 3
16 -15 4 14 5
15 -21 7
14 -15 11 I didn’t revise the Zen surrenders for h17,
13 -11 19 from an original program run after getting
12 -4 30 surrender effects of removal.
11 –4
10 3
7 19
A9 8
A8 -1
A7 -17
A6 -23
A5 -25
AA -15
XX 8
99 -8 0
88 -6
The Victor APC first appeared in public, in an article in Blackjack Forum. While it is a level 3 count it is a bit easier than it appears, at first glance to count. All the low cards are either 2 (2 thru 7)or 3 (the 5s), while the high cards are either –3 (X), or –1 (the 9s). It is best for single decks only. In power it is the top of virtually any multilevel count, in most single deck games. A player building their bankroll, and trying for the lowest HD possible, might try it, if they can handle the difficulty. Some unusual plays have fairly high correlation indexes in this count, making it deceptive too.
2 3 4 5 6 7 8 9 X A
2 2 2 3 2 2 0 -1 -3 0
s18 h h -1
17 -57 -55 -60 -58 -63 -47 s s s -16
16 -19 -23 -27 -31 -25 h h 10 -1 19
15 -12 -16 -20 -23 -22 31 31 19 8 23
14 -6 -9 -13 -15 -18 32 38 28 15 27
13 0 -2 -6 -9 -8 52 49 h 30 37
12 10 6 2 -1 0 59
11 -31 -33 -36 -41 -41 -20 -15 -11 -9 -1
10 -27 -27 -30 -31 -39 -15 -10 -4 8 7
9 2 -1 -6 -12 -14 9 20
8 29 20 14 9 5 38
7 h 39 30 28 27
A9 24 21 16 14 12
A8 22 11 6 4 2
A7 * 0 -14 -18 -18
A6 * -6 -16 -31 -30 20
A5 h 9 -1 -12 -27
A4 24 9 -2 -13 -19
A3 22 11 -1 -11 -15
A2 20 11 3 -5 -5
AA $ $ $ $ $ $ $ $ $ -10
XX 24 19 14 11 12
99 -4 -5 -9 -12 -9 5 -22 -23 s 5
88 <24 <6 $
77 $ -31 $ $ $ $
66 1 -2 -9 $ $
33 36 9 -6 -14 -16 $
22 h * -4 -7 -14 $
Insurance 1 deck 6; 2 decks 6.5; shoes 7
The late surrender Uston APC indexes closely match (why are the others so off in that similar count?).
Dealer hits soft 17s late surrender
6 A 9 X A
s18 17 16 1 -2 -1
17 -62 -13 15 4 -1 2
16 -30 7 14 6 4 6
15 -27 12 77 3 0 2
14 -19 17 88 10 3
13 -13 27
12 -5 47
11 -37 -4
10 7 ..
9 -5 …
8 6 ..
7 26 …
A9 11
A8 0
A7 -24
AA -12
99 -14
88 -8
That completes the inclussion of these neglected counts.
But a word about risk adverse indexes.
Most methods of modifying indexes for only adding more money for double-downs and pair-splits when it is optimal for the long run growth of your bankroll are still compromises, which need not be. You can find out how much each true count point improves your return on ANY double or split by simply finding the inner product between your point count and the Griffin effects of removal for that decision, from the tables in his, Theory of Blackjack, and dividing by the sum of squares of your point count values. Instead of facing the possibility of numerous tables, for various starting edges, use this added number, to find out if you are beyond the usual index, to have the same edge improvement, that you had when deciding what fraction of your bankroll to put down for your initial bet.
It is an easy calculation, that you would use by thinking of each true count past the index number providing so much improvement, in percent. Then you would double or split if the improvement in making the play matched the edge you estimated, when you sized your bet, for your initial bet.
That should be your final exam!
Chapter 12, Moving on up…how high is the blackjack sky and how to safely climb on up