If I'm reading the discussion below about the "R18", some people are suggesting that you stop doubling down with 10vs10. Is this correct? I understand the concept of "risk averse", IE you change a play in such a way that slighly decreasing win rate, but has a huge decrease in SD. What I've been doing is using higher numbers for 10vs10. Does anyone have any representive figures for how much not doubling 10vs10.
I've generally used the ultimate gambit so i'm used to always doubling that! Does anyone know if always doubling it would cost more then never doubling, IE can you "buy the right" to do it without suspicion by always doing it?
Also is the reason this play is so risky because it comes up more often then 10vsace or ace vs ace or are there other factors.

I have been looking at it this week, and doubling 10v10 is really awful. Especially if you already sue the Standing indexes for 16 and 15 v 10. There is very slight gain when the count is high, but it is offset by the Variance.

However, it probably is very effective as a cover play, because it attracts a low of attention. I remember the last time I made this play was over 150 hours ago. I was playing a double deck game and had a minimum bet out. Several people doubled and caught small cards, raising the count. I said "Somebody is due to catch a face card. If you get a small card I am going to double my 10." This was actually mathematically correct, as the next small card raised the count to the "low bet" Index. But it got big smile from the pit boss who was marking the rating card at the time.

One consideration about Ian's ultimate gambit. He seems to suggest that High Variance is itself a type of cover. He indicates that the wide swings in his bankroll obscure the slight positive trend, making him look much more like a gambler.

I am paraphrasing, and I would encourage you to read the original for yourself.

But if that is the case, then always doubling 10v10 is a fine way to increase your variance.

I've always used a hi-lo risk adjusted index of 7.5 for
10 v. 10. I don't understand why you wouldn't do it.

If it's good for cover as well as CE, why not?

Bill

I'm confused by what you mean when you say:

It is clear that never doubling makes more than always doubling, because you can increase the unit size when you never double

Are you saying that simply not doubling 10v10 gives such a decrease in SD that you can increase the size of your unit. If so how much?

Looking forward to your edition of you book don if it has info on this stuff! Keep up the good work, dude.

"Are you saying that simply not doubling 10v10 gives such a decrease in SD that you can increase the size of your unit.

Yes, that's precisely what r-a plays are all about, and 10 v. 10 is the granddaddy of them all!

"If so how much?"

Will be in the revised edition of BJA.

"Looking forward to your edition of you book don if it has info on this stuff!"

Working on it right now.

"Keep up the good work, dude."

Trying my best.

Don

SBA sims show the HiLo indices to be

                Rounded        Truncated

12v2              +4              +3

16v9              +5              +4

This is for 6 decks, S17, DAS, DOA, NS, SPA1, SPL3, 75% pen.

Notes:

A. The 12v2 play in Truncating mode kept fluctuating between +3 and +4 until it settled. Same with 16v9, in the same mode: between +4 and +5. This shows that for 12v2 the 'precise' number is very close but just below +4 while for 16v9 it's just below +5.

B. Truncating in positive TC territory is exactly the same as Flooring.

Here's the paste-up of all 4 sims, which took all of 30 minutes:

 
 

  * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 

  *          S B A   S T R A T E G Y   G E N E R A T O R        *

  *                     Version 4.01, 1995-97                   *

  * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 

***************************************************************************** 

Indices generation started  on (MM/DD/YY): 05-05-2000 at XX:XX:XX

Random seed = 1234567890


* * * * * * * * * * * * * * * * * * * * * * * * * * * * 

* R E L E V A N T   R U L E S  &  C O N D I T I O N S *

* * * * * * * * * * * * * * * * * * * * * * * * * * * * 

System used is HI-LO with playing table HI-LO
 
Remaining decks estimated exactly for true count conversion
6 decks
Shufflepoint 234 cards (75.00% penetration)

Starting at card number 1
American rules - hole card game
Surrender is not analyzed
Dealer stands on all 17
Doubling is allowed on any two cards
No resplit of aces
Only one card to split aces 

Precision: 4 sigma

Maximum number of pairs of hands at index: 300 million 

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * 

* S T R A T E G Y   G E N E R A T I O N   R E S U L T S *

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * 

SBA rounds when calculating the true count 

HARD STANDING 12 versus 2 up

Determined index(es): 4* 

Number of pairs of hands played: 6000696

SBA truncates when calculating the true count

HARD STANDING 12 versus 2 up

Determined index(es): 3* 

Number of pairs of hands played: 285033060


SBA rounds when calculating the true count

HARD STANDING 16 versus 9 up

Determined index(es): 5* 

Number of pairs of hands played: 69008004

SBA truncates when calculating the true count

HARD STANDING 16 versus 9 up

 TC         STAND                     HIT 

  4:  Adv.:-55.645, SEr 0.032  Adv.:-55.647, SEr 0.031  Dif.: 0.001, SEr 0.041

Determined index(es): 4
 

The asterisk (*) denotes statistically significant accuracy.