Wow that sucked!

[Automatic Monkey]

I'm just finishing up a sequencing mission at a theoretically very beatable shuffle in Atlantic City, and oh was this trip a disaster.

First clue- the bus ride. I got a mighty comp at the store in question so being I wouldn't need to stay out at the roach motels on Rt 30, I wouldn't need a car, so I took the bus from CT to AC to save expenses. It was late every step of the way. The leg from CT to NYC was OK but the next one was full, cramped and dirty. The malt liquor and fried chicken was being passed around like an ethnic joke being acted out. Some of these people don't have the best hygiene either, and BO was everywhere. Couldn't wait to get off.

Then the real pain began. I decided to take my first run single-ace, single-key, and had good results. Then it just turned to garbage. And the amazing part was, I was catching my bullets! Just nothing to go with them. Typically I'd end up with a soft double, get stiffed and the dealer would catch a 5 card hand. Unbelievable. I switched to a couple of other simplified methods that don't require memorizing 24-32 key cards so I could both record and play back keys on the same shoe, with about the same results. Finally I played a carnival game where you have over a 5% advantage if the next card is not an ace, to let the variance associated with sequencing work against the house rather than against me, and I got absolutely hammered. Sure, I was dodging the dealer's aces quite well, but he just caught hands time after time after time. Walked away to reanalyze.

I came up with this- that the variance for this type of sequencing is probably a lot higher than for counting because you are predicting only one card, whereas counting provides a degree of prediction for all cards. In other words, if the deck is rich in 10's, I not only have a better chance of getting a natural and of a double down converting desirably, but of the dealer busting not only on his first draw but on any card he draws. Single-key single-ace sequencing helps me get an ace on my first card but does exactly nothing for my second card, nor for the dealer's draws. So even if I get my ace, there's no expectation that the rest of the cards won't be working against me, thus the effect of every card other than my first one is pure luck. So it would seem that I can't bet a hand with a 2% advantage due to sequencing as aggressively as I would a hand with a 2% advantage due to counting. Does this make sense? Thanks in advance.

[shadroch]

Sorry to hear that.Hopefully you have a pass for Diamond Jims to watch the Big Game tomorrow.
I've been on a few of those bus rides myself. Endless trees,trees ,trees,when will they end.Now,I go mid-afternoon on Mondays and there is rarely more than a dozen people on the bus.Hope you at least got some fried chicken out of the ordeal.

[EasyRhino]

A casino bus ride to Jersey sounds like a portal to an entirely new and horrible level of hell.

Originally Posted by Automatic Monkey

So it would seem that I can't bet a hand with a 2% advantage due to sequencing as aggressively as I would a hand with a 2% advantage due to counting. Does this make sense?

From a totally non-mathematical level, this doesn't seem right. First, when you are guaranteed at least one ace in your hand, aren't you supposed to have something like a 30% edge or something ridiculous on average across all possible player and dealer hands?

Second, I understood the idea is to bet really, really, big on those hands where you catch the ace. Instead of a 10x spread, it's more like a 50x or 100x. However, you're only doing it for one hand. The massive bet would be a huge source of variance.

Finally, when straight counting, you're going to be playing a lot more "max bet" hands than when sequencing, right? That's going to give you a greater number of hands to smooth results over.

[RJT]

Originally Posted by EasyRhino

Second, I understood the idea is to bet really, really, big on those hands where you catch the ace. Instead of a 10x spread, it's more like a 50x or 100x. However, you're only doing it for one hand. The massive bet would be a huge source of variance.

Finally, when straight counting, you're going to be playing a lot more "max bet" hands than when sequencing, right? That's going to give you a greater number of hands to smooth results over.

Bang on with those 2 ER. And you have somewhere between 49-52% advantage if i remember correctly. You then have to factor in several other things, like how often you are accurate with landing the A in your hand and what not.
Then you work out your average advantage over just those hands to calculate how much money you should be putting out any time you are targetting an ace to your hand. The justification for this is that saying you have an average 2% advantage counting doesn't mean that you bet 4 units on every hand, you vary it to what your precises advantage is at any point. So when ace sequencing you have a huge advantage any time you know you have an ace coming and no advantage any other time, so you bet once again in correlation with your advantage.
I might give an example of how i would calculate how much i'd put out when i get back from swimming if anybody's actually interested.

RJT.

[Automatic Monkey]

Sure. The advantage to sequencing is that 50% times the probability of getting an ace. So with single-key sequencing in an 8D game, you start out by dividing that 50% by 8. This shuffle has about a 1 in 3 chance of the target being where it's supposed to be relative to the key, that's where I get my 2% advantage. 'm going to get a big bet down about 1 out of 3 hands. This is significantly better than the 1.2% I can get from counting such a game. But the variance problem I'm seeing is that after I get my first card, all other cards are random and I may have bet into a very low count. So if I didn't get my ace and get a stiff I'm screwed even worse than if I got a stiff betting into a high count.

[ScottH]

Originally Posted by Automatic Monkey

So if I didn't get my ace and get a stiff I'm screwed even worse than if I got a stiff betting into a high count.

In a low count you're more likely to turn your stiff into a good hand. I guess the dealer has that same luxury, but it's not as bad since you will more often hit your stiff to a pat hand as well.

[EasyRhino]

So what you're referring to is having a bet proportional to a 2% advantage out, instead of a bet proportional to a 1.2% advantage?

What about the Semyon Dukach method of playing three (or all!) spots at the table, to maximize the chance of catching that ace? Sure, you'd have big bets out on at least two hands where you had no particular advantage, but you wouldn't have to divide the 50% single-hand advantage by as much, it might pay off. (well, there's the heat...)

[ScottH]

Originally Posted by EasyRhino

So what you're referring to is having a bet proportional to a 2% advantage out, instead of a bet proportional to a 1.2% advantage?

What about the Semyon Dukach method of playing three (or all!) spots at the table, to maximize the chance of catching that ace? Sure, you'd have big bets out on at least two hands where you had no particular advantage, but you wouldn't have to divide the 50% single-hand advantage by as much, it might pay off. (well, there's the heat...)

You could make your biggest bet on the one hand you expect the ace to come, and decrease your bet as you get farthur away from that spot, since the farthur from the predicted spot, the less chance of it happening.

Although it seems to me the highest EV, yet highest variance play is to bet big on only the one hand you expect the ace to land. You could still bet multiple hands to steer the ace if needed.

[RJT]

AM,
The first point i'd like to make is that your advantage is an average advantage across all counts. If you only tracked Aces that you knew were in packets of high cards your advantage could be a lot higher than the 50% due to the higher probability of getting blackjack, but in the long run this would be a less profitable option as it would occur a lot less often.
What you should ask yourself is how many sessions have you used sequencing, how many individual times have you tried to land the (aka how many times per session x how many sessions) and how much money have you made (or lost) overall? That way you can work out your SD and what not.
I'm going to PM you my thoughts on bet sizing as i'm not sure if i'm comfortable discussing this subject on a public forum in this detail.

RJT.

[mdlbj]

Originally Posted by Automatic Monkey

I'm just finishing up a sequencing mission at a theoretically very beatable shuffle in Atlantic City, and oh was this trip a disaster.

First clue- the bus ride. I got a mighty comp at the store in question so being I wouldn't need to stay out at the roach motels on Rt 30, I wouldn't need a car, so I took the bus from CT to AC to save expenses. It was late every step of the way. The leg from CT to NYC was OK but the next one was full, cramped and dirty. The malt liquor and fried chicken was being passed around like an ethnic joke being acted out. Some of these people don't have the best hygiene either, and BO was everywhere. Couldn't wait to get off.

Then the real pain began. I decided to take my first run single-ace, single-key, and had good results. Then it just turned to garbage. And the amazing part was, I was catching my bullets! Just nothing to go with them. Typically I'd end up with a soft double, get stiffed and the dealer would catch a 5 card hand. Unbelievable. I switched to a couple of other simplified methods that don't require memorizing 24-32 key cards so I could both record and play back keys on the same shoe, with about the same results. Finally I played a carnival game where you have over a 5% advantage if the next card is not an ace, to let the variance associated with sequencing work against the house rather than against me, and I got absolutely hammered. Sure, I was dodging the dealer's aces quite well, but he just caught hands time after time after time. Walked away to reanalyze.

I came up with this- that the variance for this type of sequencing is probably a lot higher than for counting because you are predicting only one card, whereas counting provides a degree of prediction for all cards. In other words, if the deck is rich in 10's, I not only have a better chance of getting a natural and of a double down converting desirably, but of the dealer busting not only on his first draw but on any card he draws. Single-key single-ace sequencing helps me get an ace on my first card but does exactly nothing for my second card, nor for the dealer's draws. So even if I get my ace, there's no expectation that the rest of the cards won't be working against me, thus the effect of every card other than my first one is pure luck. So it would seem that I can't bet a hand with a 2% advantage due to sequencing as aggressively as I would a hand with a 2% advantage due to counting. Does this make sense? Thanks in advance.

What was your count?