Three Questions

[Bushie]

Hi all, more questions.

1. What effect does wonging out at -2 when using a balanced level 2 count (FELT w/indices in this case) have upon SD and WR?

2. How much does wonging out at this point reduce RoR?

3. Regarding FELT, a few test runs in shoe games using full deck divisors felt...off. It didn't feel right somehow. Would it be better to use half-deck divisors and half or double (?) the indice values instead?

[Three]

1. What effect does wonging out at -2 when using a balanced level 2 count (FELT w/indices in this case) have upon SD and WR?

SD would go down slightly and win rate would go up slightly but more significantly. The result is c-SCORE would go up very significantly.

2. How much does wonging out at this point reduce RoR?

It would significantly reduce RoR.

3. Regarding FELT, a few test runs in shoe games using full deck divisors felt...off. It didn't feel right somehow. Would it be better to use half-deck divisors and half or double (?) the indice values instead?

No. It would have next to no effect. You are just changing granularity. All the bets and plays would be the same.

[Three]

When you say full deck divisor you aren't saying full deck accuracy are you? You can use any accuracy of deck estimates with full deck divisors. It just has to be expressed as the number of full decks to a fraction that fits the deck estimate. Like you estimate 1 and 3/4 decks left, then the full deck divisor would be 1.75 decks or 7/4 of a deck. That would be the same as multiplying by its inverse, 4/7.

[Bushie]

SD would go down slightly and win rate would go up slightly but more significantly. The result is c-SCORE would go up very significantly.

I've calculated an SD of 32.26 p/100 hands and a WR of 1.9 p/100 hands. This is for play-all as outlined in BJA3. How do you go about calculating how wonging out at all counts at/below -2 affects these?

It would significantly reduce RoR.

I guess once I know how much wonging out affects SD and WR that I can figure out the RoR reduction then

No. It would have next to no effect. You are just changing granularity. All the bets and plays would be the same.

But isn't increased granularity what you want? Because it feels odd in a shoe game, using a count (or a count derived from) that was originally designed for single and double deck games..

[Bushie]

When you say full deck divisor you aren't saying full deck accuracy are you? You can use any accuracy of deck estimates with full deck divisors. It just has to be expressed as the number of full decks to a fraction that fits the deck estimate. Like you estimate 1 and 3/4 decks left, then the full deck divisor would be 1.75 decks or 7/4 of a deck. That would be the same as multiplying by its inverse, 4/7.

No dividing by half decks remaining. Like RPC does for hand dealt games. The full deck improvisation feels...off. I don't know, maybe its because I haven't used full deck division in a long time.

[Three]

I've calculated an SD of 32.26 p/100 hands and a WR of 1.9 p/100 hands. This is for play-all as outlined in BJA3. How do you go about calculating how wonging out at all counts at/below -2 affects these?

Simulation

But isn't increased granularity what you want?

Increasing granularity allows you to spread the data out and make finer decision barriers while using integer TCs. If you simply divide by double decks remaining to double the range and then double indices so they are only even integers you haven't made anything more accurate. I guess betting could be made more accurate if you ran a sim that produced some odd integers for betting bin barriers.

I don't know if I explained that very well. How about an example. With less granularity you have a ramp that goes bets a unit for each TC. And you have 2 indices +3 and +2. If you divided by half as much to increase granularity and had your ramp bet 1 unit for each even integer that would bet exactly the same. If you doubled the indices to +6 and +4 every decision made would be the same. You are still using the same number of pixels but the pixels are twice the size. If your ramp went from betting bins being TC +1, +2, +3, +4 to betting bins of TC +2, +3, +4, +6, +8 you would bet differently. If your two indices went from TC +2 and +3 to TC +3 and +5 because the decimal indices were +1.7 and +2.8 but the integer indices were +2 and +3 the "doubled indices" would be making more accurate decisions. By bringing the odd integers into play with the "doubled system" rather than literally doubled system you are using the increased granularity (more pixels). If you don't get the odd integers into play you haven't really increased granularity.