RA indices and flooring

[pmc]

I read about risk averse indices in BJA3. Skimming the mathematical details that are mostly over my head, in a practical sense, my observation was that RA indices tended to be larger than EV maximising indices ie. an EV+ index of +1 would likely be an RA index of +2. I guess the point being that being risk averse means only betting on the outcome of a play decision when the larger associated index indicates a larger player advantage of that outcome occurring.

Assuming that's correct I then started thinking: say my count system's indices, when generated, used rounding. Now what if, while playing, I use flooring. If I required an index of +2 for a particular playing decision then using flooring would surely mean that my RC would have to be more positive before the floored RC conversion would result in that +2 TC.

Is my thinking correct and further is it reasonably (even if not strictly mathematically) correct to say that by flooring your play decisions become more risk averse?

[DSchles]

By flooring, your decisions become more risk averse only if the original indices were generated by rounding; i.e., if you don't use the original indices in the manner in which they were intended to be used.

And so, yours is not the best idea not only for that reason but also because, in the vast majority of cases, you'd be using the wrong alleged r-a index, since most plays don't have/require a r-a index associated with them.

Don

[pmc]

OK so if I have an existing index originally defined using rounding that is +2 then, while playing, I still make the system indicted play deviation but at a floored rather than rounded +2 count then that decision is still the correct deviation but applied in a more risk averse way?

Or have I misunderstood?

[Cacarulo]

OK so if I have an existing index originally defined using rounding that is +2 then, while playing, I still make the system indicted play deviation but at a floored rather than rounded +2 count then that decision is still the correct deviation but applied in a more risk averse way?

Or have I misunderstood?

As Don told you but in other words, if your indices are floored, you should use floored indices, and if your indices are rounded, you should use rounded indices.
On the other hand, if your indices are floored, the RA indices should also be floored.

Sincerely,
Cac

[pmc]

Sounds like I did misunderstand. Thank you for the clarification.

[Freightman]

As Don told you but in other words, if your indices are floored, you should use floored indices, and if your indices are rounded, you should use rounded indices.
On the other hand, if your indices are floored, the RA indices should also be floored.

Sincerely,
Cac

An interesting deviation. First, I agree with floored vs rounded or truncating indices. Also, understand that EV maximizing numbers are based on break even. So, for example, you have 8v6 with EV maximizing TC 1.0, an actual floored TC of 2.0, you are still under the floored RA index of TC 3.0. (Wong’s Professional Blackjack). However, the actual RC well exceeds TC 2.0 and for arguments sake when calculating an exact TC is 2.9.

Now by definition, everything between 2.0 and 2.99 is floored to 2. Just call it TC 3 and make the double.

To look at it from a different perspective - playing halves has a 6 deck insurance strike point of 3.4. By flooring to TC 3.0, you’re actually losing money long term.

On CVCX, you can get more accurate results of ramping by using 1/2 true counts. This also has the added benefit of making you appear to be sub optimal on bets.

Far more to that last comment, but that’s a story for a different day.

[DSchles]

To look at it from a different perspective - playing halves has a 6 deck insurance strike point of 3.4. By flooring to TC 3.0, you’re actually losing money long term.

I'm sorry, but this simply isn't true. First, the index is 3.3, not 3.4. More importantly, when you use indices, they always represent an interval (in this case 3.0 to 3.999) and not a single point. So, while, technically, not insuring from 3.0 to 3.2999 might be the superior play, insuring from 3.3 to 3.999 IS the correct play, and the positive EV from that larger interval exceeds the negative EV from the smaller interval, such that the EV for the entire interval (3.0 to 3.9999) is positive. If it weren't, the index would be +4 and not +3. But, it is correctly +3, and using it will make, not lose, money in the long run.

Do you understand?

Don

[Freightman]

I'm sorry, but this simply isn't true. First, the index is 3.3, not 3.4. More importantly, when you use indices, they always represent an interval (in this case 3.0 to 3.999) and not a single point. So, while, technically, not insuring from 3.0 to 3.2999 might be the superior play, insuring from 3.3 to 3.999 IS the correct play, and the positive EV from that larger interval exceeds the negative EV from the smaller interval, such that the EV for the entire interval (3.0 to 3.9999) is positive. If it weren't, the index would be +4 and not +3. But, it is correctly +3, and using it will make, not lose, money in the long run.

Do you understand?

Don

I do understand HOWEVER, why not make more money by not insuring under 3.3. With that out of the way, I don’t view it as an issue to insure hands under 3.3 that are good hands - such as 19 or 20, or not insuring 15 or 16 at strike point - and for that matter using additional information available to aid in those insure or not insure decisions.

Also, interested in your thoughts of flooring to full integer values when playing 1/2 true counts. I view 1/2 true counts as increasing EV while distorting competency to the house.

Recall a post of yours some years back (wish I had bookmarked it) saying something to the effect of - had 20 v dealer ace with TC 2.86 (somewhere around there), so I didn’t insure - it really was quite funny

[DSchles]

I do understand HOWEVER, why not make more money by not insuring under 3.3.

Yes, I can see that as a possibility. Just not worth a whole lot, but technically correct nonetheless.

Also, interested in your thoughts of flooring to full integer values when playing 1/2 true counts.

Not sure I'm following. If you use indices such as 2.5, 3.5, etc., why would you then floor them to integers? What good would the half values be if you don't use them?

Recall a post of yours some years back (wish I had bookmarked it) saying something to the effect of - had 20 v dealer ace with TC 2.86 (somewhere around there), so I didn’t insure - it really was quite funny

Don't recall it, but I can see that I could do this in my head. 8-deck game, RC = +20, 7 decks remaining. TC = +20/7 = 2 6/7 = 2.857 = 2.86. Don't insure. (You would never, EVER want to get inside my head. Scary things go on there. )

Don

[Cacarulo]

An interesting deviation. First, I agree with floored vs rounded or truncating indices. Also, understand that EV maximizing numbers are based on break even. So, for example, you have 8v6 with EV maximizing TC 1.0, an actual floored TC of 2.0, you are still under the floored RA index of TC 3.0. (Wong’s Professional Blackjack). However, the actual RC well exceeds TC 2.0 and for arguments sake when calculating an exact TC is 2.9.

I didn’t remember that PBJ had RA indices.

To look at it from a different perspective - playing halves has a 6 deck insurance strike point of 3.4. By flooring to TC 3.0, you’re actually losing money long term.

Actually, +3.4 means that the division you're making between the RC and the remaining decks must be exact. You can use that index to gain a bit more, but that doesn't mean
you're at a disadvantage at a floored TC of +3. When evaluating the entire bin between 3.0 and 3.99, the advantage for insurance is positive, and your bet is related to the whole bin.

Sincerely,
Cac

[Freightman]

Not sure I'm following. If you use indices such as 2.5, 3.5, etc., why would you then floor them to integers? What good would the half values be if you don't use them?

I wouldn’t - simply interested in your perspective - which you answered. Don’t think I articulated well.

Don't recall it, but I can see that I could do this in my head. 8-deck game, RC = +20, 7 decks remaining. TC = +20/7 = 2 6/7 = 2.857 = 2.86.

Depends on the hand - influenced of course by QTC

You would never, EVER want to get inside my head. Scary things go on there. )

I guarantee you don’t have exclusive license on that point