>Particulary interesting is the T3 hands etc. is this a mistake in the indixes
>and not in the actual system if the indexes were correctly calculated?
That's a good question. To look at it, I ran three sims: Hi-Opt I, AGII and AGII with corrected indexes. The results show that AGII with repaired indexes received nearly identical results for 12vT, 13vT and 14vT as Hi-Opt I.
1. So, indeed the indexes were way off and fixing them corrected the problem.
2. But fixing them only gave the same results as Hi-Opt I with no indexes, one of the easiest TC systems existent. And these three plays are among those that require the multi-level, third side-count.
Then the question becomes how would the three plays perform by reverting back to the Intermediate Aguilar count (not requiring the second or third side count.) So, I ran a fourth sim.
3. 12vT performed better without use of the third side count. 13vT was a wash. 14vT performed worse without the third and second side count.
Thus the answer to your question is yes the errors in the indexes certainly causes the bulk of the problem. But, the strategy itself is also in error.
As you brought up these particular hands, I would also like to address the difficulty of AGII. Aguilar is again indicating that keeping the counts is the bulk of the difficulty. (In fact he is backing off his claim that you can easily keep a third, multi-level, balanced side count and saying you can just look at the cards on the table instead of keeping a RC for the third count. But, if you do this, all his correlatoin numbers and WAC calculations are wrong as they assume a correct RC.) He is ignoring the USE of the counts which is even more complex than keeping the counts. Let me provide an example of the difficulty of this strategy using the hands you brought up. I will not provide any indexes in the example as I do not violate copyrights.
Suppose you have somehow managed to calculate and keep in your head the three separate balanced counts. A very difficult task in itself particularly since this is a level III strategy. A new hand is dealt:
You receive 4 & 7 and the dealer has a 10. For the four, add 1 to count #1. For the Seven add 1 to count #1 and 2 to count#3. For the Ten, subtract 1 from count #1 and count #3. Do the same for the other player hands as they are played.
You have 11v10. Calculate the TC using count #1 and play according to the index.
You draw an Ace and now have 12v10. Subtract 1 from count #2. For 12v10, use count #3. Presumably you convert count #3 to TC and compare to the index, but the strategy doesn’t say whether you use TCs or RCs.
You draw a 2. For 14v10, you add counts #1, #2 and #3, and convert the sum to a TC and compare to the index.
You draw another 2. For 16v10, you SUBTRACT count #2 from count #1, convert it to the TC and compare.
At each decision point in the hand you need to use a different combination of the counts and recalc the TC and compare against the index while continuing to update the three counts. This is just one hand.
It should also be noted that in typical strategies, you rarely need to recalculate the TC after drawing a card. This because you have a sense of the proximity of the index and the TC from your previous calculation. But, this is not true in AGII because each decision can require a different set of counts.
I'll add one more point. If you take the above hand and play using Hi-Opt I with no side counts, the hand is over in two seconds. But, Aguilar is saying you should tell the dealer to wait you have to think it over while you make the calculations. Assuming you could somehow accurately keep three balanced counts and make all these calculation correctly, you would still slow the game dramatically and that greatly reduces your win rate.
The simple fact is that KO will beat the most advanced of the Aguilar counts played perfectly.
