Index Generation

[BJmath]

On the counts tab, set the TC calculation to whatever you wish.

Two other questions re: index generation using your CVData: 1) Where can I set the index range, which is dependent on the level of a user-defined counting strategy? 2) I had a little difficulty finding out how to specify composite dependent index generation settings? For example, what does each column mean in the following table?

Dlr Player Decision S ...

Thanks!

 

[QFIT]

Two other questions re: index generation using your CVData: 1) Where can I set the index range, which is dependent on the level of a user-defined counting strategy? 2) I had a little difficulty finding out how to specify composite dependent index generation settings? For example, what does each column mean in the following table?

Dlr Player Decision S ...

Thanks!

1. You can't set an index range when generating. It can't know if an index is in the range until it has generated it. You can set the range in CVData/CVCX/CVBJ when using the table.

2. From the Help:

Composition-Dependent Indexes

If the Comp-Dependent option is selected, a new table is displayed. Use the Number spinner at the bottom to indicate the number of CD indexes you wish to generate. For each index, enter the dealer card, player cards and decision. There can be two to five player cards separated by commas. Click on the decision cell to show a down arrow and then click on it to select a decision type. You may need to click a few times. Aces can be specified as 1 or A. Tens can be specified as 10, T or X. The last column is used for custom bonuses. Check the box if the two cards held are of the same suit. This affects the index if there is a bonus that requires same suits

 

[BJmath]

1. You can't set an index range when generating. It can't know if an index is in the range until it has generated it. You can set the range in CVData/CVCX/CVBJ when using the table.

2. From the Help:

Composition-Dependent Indexes

If the Comp-Dependent option is selected, a new table is displayed. Use the Number spinner at the bottom to indicate the number of CD indexes you wish to generate. For each index, enter the dealer card, player cards and decision. There can be two to five player cards separated by commas. Click on the decision cell to show a down arrow and then click on it to select a decision type. You may need to click a few times. Aces can be specified as 1 or A. Tens can be specified as 10, T or X. The last column is used for custom bonuses. Check the box if the two cards held are of the same suit. This affects the index if there is a bonus that requires same suits

Thanks! If I can't set the index range for generation, what is the default range then? Is the default the same regardless of the level of the counting system?

 

[QFIT]

A generator has no idea what the index is before generating it. The range is whatever the highest and lowest counts are for the count, decks, penetrations and calculation methods.

 

[BJmath]

A generator has no idea what the index is before generating it. The range is whatever the highest and lowest counts are for the count, decks, penetrations and calculation methods.

The index range was narrower than the possible one in my generation. But I realized that's not really an issue since I could always trick your program to generate the indexes using a proportional (fractional) definition of the values in my counting strategy.

 

[QFIT]

Ahh yes, I currently have an upper limit of +40. I'll change that next release.

 

[BJmath]

Ahh yes, I currently have an upper limit of +40. I'll change that next release.

I ran index generations using a counting strategy in which each value of card is specified as half of my original one. However, I got indexes that to me were pretty inconsistent with my previous results (I assume the new indexes should be roughly half of the orginal one, assuming true count divisor is exact number of decks remaining). Did I not think right here? Thanks.

 

[QFIT]

If unbalanced, you need to also adjust the IRC. If the base startegy has indexes, you need to also adjust them.

 

[BJmath]

If unbalanced, you need to also adjust the IRC. If the base startegy has indexes, you need to also adjust them.

My counting strategy is balanced.

Are you saying to generate consistent indexes, I should not use a playing strategy with indexes included? That's probably where my problem was because I was using a previously generated playing strategy with indexes in it.

 

[QFIT]

The best way to generate a new strategy is to base it on basic strategy with the tag values. No indexes. If you wish to generate an additional index, then it's best to base it on the strategy WITH indexes. Indexes affect each other.

 

[BJmath]

The best way to generate a new strategy is to base it on basic strategy with the tag values. No indexes. If you wish to generate an additional index, then it's best to base it on the strategy WITH indexes. Indexes affect each other.

This helps a lot! BTW, what's the interpretation of the indexes generated based on the strategy WITH indexes (the so-called additional indexes)?

 

[QFIT]

Not sure what you mean. If indexes are in the base strategy, they will be used in generating the new indexes.

 

[BJmath]

Not sure what you mean. If indexes are in the base strategy, they will be used in generating the new indexes.

Another question: In your CVData, there's an option to check namely "half point supported". Does this say that if I have a counting strategy that involves half points, such as 2.5, I should check that option in generating indexes? Thanks.

 

[QFIT]

Yes, and only check it if you have halves.

 

[zengrifter]

Fast yes. Accurate, not close. I've known Sam Case for numerous years and respect his work. But this is just an old program based on a poor estimation technique.

By "poor estimation method" I presume you mean the 'algebraic approximation' approach?
I always like algebraic approximation index, on general principal. zg

 

[QFIT]

It's fun to come up with such algorithms. But, this is called an approximation for good reason. I respect the effort. But, wouldn’t actually use the results.

 

[zengrifter]

It's fun to come up with such algorithms. But, this is called an approximation for good reason. I respect the effort. But, wouldn’t actually use the results.

The aproximated index generator in my old trusty NeUltra worked great.
Later when Richard Reid came out with his own trainer he told me that he "improved" the algorithm, but I thought it was of no consequence.

Remember when Sam Case had a side-business generating algebraic indices? zg

 

[QFIT]

Then again, he also lead discussions on magic tricks. Don't know if they're related.