Lucky Lucky Combinatorial Analysis?

**paul5795** · 12-19-2019, 01:33 PM

Miplet has created some real nice excel spreadsheets that apply combinatorial analysis to certain blackjack side bets. http://miplet.net/blackjack/ One side bet that is missing from miplet’s webpage is the Lucky Lucky side bet. Does anyone know if there is a public domain spreadsheet for the Lucky Lucky side bet that is similar to what miplet has done for the other side bets? If not, is there anything out there in the private domain? I would be willing to pay something de minimus (e.g., ten or fifteen bucks) for a copy of such a Lucky Lucky spreadsheet that allows, as with Miplet’s spreadsheets, for manipulation of the shoe contents. The Lucky Lucky pay table I am looking at is suited 777 (payout 200-1), suited 678 (100-1), unsuited 777 (50-1), unsuited 678 (30-1), suited 21(10-1), unsuited 21(3-1), total of 20 (2-1), and total of 19 (2-1). For those of you unfamiliar with this side bet, all of the totals above are three card totals, that is, obtained by adding the player’s first two cards and the dealer’s upcard. So, for example, a player blackjack does not win the Lucky Lucky side bet unless the dealer also has a face card or a nine (e.g., 10+1+10 = 21; and 10+1+9=20).

**Norm** · 12-19-2019, 01:57 PM

What's the use if you don't have it by count? Side bets typically have awful odds without counting.

**paul5795** · 12-20-2019, 06:41 AM

I have a count. Actually, a couple of counts. Since the contents of the shoe can be adjusted in Miplet's spreadsheets I can sample different shoes set to specific counts (e.g., TC=8) to see if those shoes are always, or almost always, plus EV. Frankly, although I have CVdata, I am not comfortable that I am using that software correctly, and at least for certain side bets I can gain some assurance from sampling using the combinatorial analysis spreadsheet.

**Norm** · 12-20-2019, 08:26 AM

OK. But, that won't be accurate.

**paul5795** · 12-20-2019, 03:34 PM

Why wouldn't the combinatorial analyzer be accurate for calculating the probabilities of the side bet for the first hand out the modified shoe? Since the side bets do not depend on any player strategy, the side bet probabilities for the first hand off the given shoe should not only be accurate, they should be 100% accurate. What am I not understanding?

I fully understand that this only produces a narrow snapshot. But it is a snapshot that I understand exactly how it works, unlike the other software I have purchased. I can even replicate some of the Miplet analyzer results. For example, for the side bet "Bust It," which pays out if, and only if, the dealer busts with exactly three cards, I am able on my own to calculate that at the very beginning of a virgin six deck shoe the odds that the dealer gets a three card bust with the third card being a six are exactly 0.488299%, which is consistent with what Miplet's CA program reports for this sidebet. This gives me confidence that if I reduce the shoe down to two or three decks, and set up the cards in that shoe with a particular true count in various combinations, that the results for those shoes will also be accurate.

**Norm** · 12-20-2019, 03:39 PM

Who cares what happens in the first hand? There is no reason to bet at that point. Your last sentence is simply incorrect. Alas, the farther into the shoe you go, the less accurate the results. And, side bets like this are rarely useful unless well into the shoe. I tried experiments like that a couple decades ago. You cannot calculate correct indices in that manner.

**paul5795** · 12-20-2019, 04:15 PM

Who cares what happens in the first hand? There is no reason to bet at that point.

I did not say I would bet at that point. I was simply using that as a data point to give me confidence in the accuracy of the analyzer.

Your last sentence is simply incorrect. Alas, the farther into the shoe you go, the less accurate the results.

This I simply do not understand. If I set up a shoe with 104 specific cards (i.e., the equivalent of two decks), a combinatorial analyzer that is brute forcing the probable results of say, the next three cards, should be 100% accurate in those probabilities.

And, side bets like this are rarely useful unless well into the shoe.

I know this. This is why I used the example of TC=8 in my post above, and not TC=0.

I tried experiments like that a couple decades ago. You cannot calculate correct indices in that manner.

I dont understand what indices have to do with this. Also, I am not devising my count system this way, I am sampling the accuracy of a given system as one gets farther and farther into the shoe.

Anyways, I will stop responding here. But I am still interested in any CA spreadsheet that someone might have that covers Lucky Lucky.

Thanks.

**Norm** · 12-20-2019, 05:05 PM

Originally Posted by paul5795

If I set up a shoe with 104 specific cards (i.e., the equivalent of two decks), a combinatorial analyzer that is brute forcing the probable results of say, the next three cards, should be 100% accurate in those probabilities.

I realize that you do not understand the math behind this, or why the last two decks of a six deck shoe are not relevant to a two-deck game -- as has been explained by innumerable folks over decades. Do what you wish.

**Freightman** · 12-20-2019, 05:56 PM

Originally Posted by Norm

I realize that you do not understand the math behind this, or why the last two decks of a six deck shoe are not relevant to a two-deck game -- as has been explained by innumerable folks over decades. Do what you wish.

Actually, OP’s post 3 is kind da sorta on the right track. Essentially, this side bet with regular payout table has house edge of 2.6%. Notwithstanding specialty counts used to exploit to maximum benefit, there is in fact a study, tied to hi lo, in the BJ21 archives. In simplest terms, advantage gradually shifts to player advantage in extreme negative and positive counts.

**Norm** · 12-20-2019, 05:59 PM

Obviously true. But not at all relevant to what the OP is claiming.

**Freightman** · 12-20-2019, 06:06 PM

Originally Posted by Norm

Obviously true. But not at all relevant to what the OP is claiming.

Perhaps not, however, if he better defines his question, then he might suggest something like “is there a study that defines lucky lucky payouts by count”.

Then, it becomes highly relevant.

**Norm** · 12-20-2019, 06:09 PM

Well, such studies are simple and he has the software. But, his methodology and theorizations are incorrect.

**Freightman** · 12-20-2019, 06:14 PM

Just rereading opening post. Standard payout on suited 21 is 15-1, which I stated as 2.6% house edge. OP has stated 10-1, which essentially doubles house edge, rendering the game unbeatable. I know of a store or 2 that has reduced this portion as described. If accurate, don’t worry about the study, just stay away.