Hi-Lo index for 12v3

[UK-21]

In the I18 the Hi-Lo index for 12v3 is TC+2 (6 deck shoe).

Does anyone have the actual avg long term return numbers for standing, hitting and doubling this hand at this TC?

OTT at a neutral count (according to the chart I have that underpins BS)they are: -0.24719, -0.23303 and -0.46605 respectively. So there's 1.4p in the £ (or cents in the $) longer term between standing and hitting - little in it.

Thanks.

 

[Nynefingers]

Assuming 6 decks, S17, truncating...and assuming I didn't screw up, as this is my first time playing with the CVData index generator. These numbers were generated using "standard accuracy."

TC       EV stand       EV hit      Difference
+1       -.239289      -.234390      -.004899
+2       -.225750      -.235506       .009756
+3       -.214169      -.237621       .023451

Also, I disagree with your statement that 1.4% is insignificant. We are playing a game with roughly that much overall advantage, so if that difference seems insignificant to you, then there isn't much point in counting cards to begin with. True, you can beat the game without indices, but if you can remember the top x number of indices, the benefit is worthwhile.

 

[UK-21]

Thanks for doing that. I've always stood on a 12v3 when it was dealt but I'm interested in this hand from the perspective of a 9v3 being hit with a 3. Most people will double this hand, and so be stuck with a 12.

If you take the view that if all of the three additional cards per 52 are 10s, it means that, on average (of course) 19 cards out of 52 (10s, 16+3) would bust the hand, but 33 wouldn't - 21 cards would have little effect (2-6 - x17 - plus 4xAs), and the unknown 7,8,9 (average of 12 of them) would result in a strong hand. With the dealer having a three there's a strong chance of it turning into a made hand, which will stuff the 12.

If, at the other end of possibility, the three additional cards at TC+3 are all aces, then it would mean that 16/52 cards would bust the hand, but 36/52 wouldn't, which I would think would shift the odds of busting out a fair bit. But unless you're keeping a side count of aces you can't know.


Oh. . . and when I referred to "little in it" it was meant relatively speaking as a value of the benefit of just this deviation over the longer term. If you view this example relative to the far end of the scale, ie whether to double or hit an 11v6 (the longer term difference being 33.6p in the £ - so the loss in EV for hitting rather than doubling on that one is far, far greater) it's not a huge cost. In actual fact, the loss in EV in hitting a 12v3 at TC+3 rather than standing looks to be virtually the same as not doubling A2v6 at a neutral count but just hitting it.

Anyway, those figures have enlightened me. Thanks again.

 

[Nynefingers]

I'll admit that I didn't quite follow all that you wrote there, but it looks to me like you are overthinking this. Follow basic strategy unless an index tells you to deviate. Use the published risk averse indices, which you can find in plenty of places online and in books. They will generally be listed in order of overall value, at least for the first 18 or so indices. If you want to improve your EV further, you can find more extensive lists of indices or use CVData to generate your own.

 

[UK-21]

Overthinking this? Probably guilty considering the amount I play.

I have taken an interest in those index plays where there's little loss of EV between hitting, and the slightly higher (mathematically based) return of doubling. The maths purests will double them on the grounds that the maths show them to have a higher EV - the mathematically correct decision. But I've looked at it from the point of view of reducing variance, ie does the difference in EV between hitting and doubling justify the risk of losing, for example, 32 units instead of 16 at a high count?

A good example is playing A4v4. This is a marginal double at a neutral count. With one unit in the box, it's not a big issue to double it to two. But at TC+4+ with 16 or more units in the box? It's all relative to your bankroll of course, but as I don't have money set aside specifically to play BJ but hit the tables with 40-50 units in hand, for me it would be a very high risk play.

 

[sabre]

Yes, you are overthinking. Risk averse indices take into account the risk relative to the expectation. In most cases, the risk averse index isn't more than 1 TC higher. Not doubling A4v4 at TC +1 isn't too bad. Not doing so at TC +4 is terrible. True, you aren't dealt that particular hand very often ... but people who chicken out on the close doubles are probably chickening out on lots of not so close doubles as well.

 

[Automatic Monkey]

Overthinking this? Probably guilty considering the amount I play.

I have taken an interest in those index plays where there's little loss of EV between hitting, and the slightly higher (mathematically based) return of doubling. The maths purests will double them on the grounds that the maths show them to have a higher EV - the mathematically correct decision. But I've looked at it from the point of view of reducing variance, ie does the difference in EV between hitting and doubling justify the risk of losing, for example, 32 units instead of 16 at a high count?

A good example is playing A4v4. This is a marginal double at a neutral count. With one unit in the box, it's not a big issue to double it to two. But at TC+4+ with 16 or more units in the box? It's all relative to your bankroll of course, but as I don't have money set aside specifically to play BJ but hit the tables with 40-50 units in hand, for me it would be a very high risk play.

You don't have enough units in your hand to play a proper game.

It's a valid idea, but you've got it a bit backwards. Yes, at low or neutral counts hitting instead of doubling is a good idea, but at high counts those plays are too profitable to let slide.

There are a few exceptions- retrograde indices where after a certain high count you don't do the action. The most well-known is splitting 88 vs. 10. In a surrender game you surrender that in positive counts, and in a non-surrender game you would simply stand once the count gets up to around where a max bet is for most people. There are a few other retrograde split indices, and I recall there's at least one soft double where the EV goes down as the count increases (depending on your counting system of course.) It should be on one of QFIT's charts in Modern Blackjack.

 

[21gunsalute]

Thanks for doing that. I've always stood on a 12v3 when it was dealt but I'm interested in this hand from the perspective of a 9v3 being hit with a 3. Most people will double this hand, and so be stuck with a 12.

If you take the view that if all of the three additional cards per 52 are 10s, it means that, on average (of course) 19 cards out of 52 (10s, 16+3) would bust the hand, but 33 wouldn't - 21 cards would have little effect (2-6 - x17 - plus 4xAs), and the unknown 7,8,9 (average of 12 of them) would result in a strong hand. With the dealer having a three there's a strong chance of it turning into a made hand, which will stuff the 12.

If, at the other end of possibility, the three additional cards at TC+3 are all aces, then it would mean that 16/52 cards would bust the hand, but 36/52 wouldn't, which I would think would shift the odds of busting out a fair bit. But unless you're keeping a side count of aces you can't know.


Oh. . . and when I referred to "little in it" it was meant relatively speaking as a value of the benefit of just this deviation over the longer term. If you view this example relative to the far end of the scale, ie whether to double or hit an 11v6 (the longer term difference being 33.6p in the £ - so the loss in EV for hitting rather than doubling on that one is far, far greater) it's not a huge cost. In actual fact, the loss in EV in hitting a 12v3 at TC+3 rather than standing looks to be virtually the same as not doubling A2v6 at a neutral count but just hitting it.

Anyway, those figures have enlightened me. Thanks again.

I think you're overlooking the fact that the dealer could take 2 or more smaller cards and still bust.

 

[UK-21]

Thanks all for your responses. Before I respond further, can someone crunch the numbers on their sim, and post the EV values hitting and doubling on A,4v4 and A,2v5 at TC+3 and TC+4 (assume 6 deck shoe). OTT these are (on the table I have) 0.3p in the £ and 0.25p in the £ (0.0030 and 0.0025 - at 4dps) respectively.

Out of interest, I sorted my hand expectation chart by diff in value from hitting and doubling down on those plays where one or the other had a higher EV than standing. The two with the lowest were those hands above. The one that came in 3rd lowest was A,5v4 at 2p in the £ (0.0193) - over 6 times better than A4,v4. And, of course, these are all longer term EV stuff that will be swamped in variance.

So, would someone be so good as to publish the EVs for A2,v5, A4,v4, A5,v4at TCs +3 and +4, so we all know what loss in EV we're discussing?