Martingale As "Cover"...

Martingale As "Cover"...

Has anyone ever analyzed using a negative progression "Martingale" as COVER play in a play-all card counting strategy?

What I mean is something like playing a $10-$1000 table with a bankroll adequate for $100 unit bet, but progressing 10-20-40-80 whenever the count is negative or zero, and then only ramping up to using a 1-10 spread in positive counts (i.e. at +1 betting "one unit" of $100 apparently "continuing the progression" in the eyes of the pit) and betting up more, 2 units $200 at +2, $400 at +4, etc... for an actual 1-10 spread, in reality, while still apparently playing the progression strategy in the pit's eyes.

Seems to me this would mean your average bet in negative or zero counts would therefore be less than 1 unit... in itself a benefit, no?

How would you go about modeling something like this for EV?

interesting

If they were to analyze your play during a negative or neutral count, it would look like a progression which could buy you some cover, but if they analized it just as the count went positive and you started to spread to $400 and beyond, it would look like you were counting but spreading $10-$400. What happens if early into your progression the count jumps. Are you going to go from $10 or $20 up to $400 in one jump? That could be something that would draw attention and cause them to take a look at your play and when they did, at this point you would be betting with the count, so rather than cover, this style actually draws attention just as you are spreading with the count.

Heh... well, I didn't say...

...it was "perfect".

Obviously if things don't "flow" exactly right for the timing on when you hit the maximum bet of the progression, and when you ramp up into the multiple bet units with high counts, sure, it might blow the illusion.

Jumping $10 to $400 obviously ain't a good idea in *any* scenario, LOL. Pit's always gonna notice THAT!

Unless you can really act like you just went crazy, you'd have to temper these jumps with however things are going, if you feel any heat, if you think you can pull off the old "stuck and steaming" act, etc etc.

Best situation would be when you just happened to lose 4 or 5 hands in a row, on all small cards coming out, just when you know the tens are coming, and you can then plausibly play the old chasing-my-losses act into a nice plus count. But again, obviously, it's not gonna be that smooth all the time.

I'm just trying to figure out how to model this, to put some hard numbers on it, rather than it just being a "how to behave" theory...

Modeling using CVData

Nestor,

Your strategy is within the simulation capabilities of Norm's excellent CVData program. You'll need to use an "Advanced" Betting Strategy for this.

Since your post did not describe the complete scheme, I made a few assumptions:

1. The "unit" used is $10.
2. For TC's of zero and below, you'll bet from $10 to $100 (1 to 10 units) using a Martingale approach: on any loss, you'll bet $10; after a win, you'll double your previous bet, capping at $100 (which you'll bet after winning an $80 bet in a non-positive TC).
3. If you bet $100 in a non-positive TC, your NEXT bet will be based on the TC, so if you win and the TC<+1, your next bet will be back to 1 unit. If the TC>0, then use rule 4.
4. For TC's of +1 and higher, you'll bet according to the count, with this schedule: $100 (10 units) at +1, $200 at +2, $400 at +3, $600 at +4, and $1000 (100 units) at +5 and higher.

Thus, overall you'll be spreading 1:100 in $10 units.

First, set up a simulation with the desired house rules, number of players, and playing strategy. Then go to the Betting tab and define a new strategy, basing it on the canned "Flat Bet" scheme. On the "Basic" screen uncheck the "Reset after shuffle" and "Running Count" boxes, and enter an Ending Index of 5. Then, enter the following schedule:

 

<=0 1 unit 

+1 10 units 

+2 20 units 

+3 40 units 

+4 60 units 

+5 100 units 

.

Now click the "Advanced" button. Locate the box at the intersection of the "Last Bet 1 unit/Won" row and the "<=0" column. Change the "1" there to a "2", and CVData will now create a new set of rows for a 2-unit bet. Now, locate the box at the intersection of the "Last Bet 2 units/Won" row and the "<=0" column, and enter a "4". Repeat to add an 8-unit row. Finally, locate the box at the intersection of the "Last Bet 8 units/Won" row and the "<=0" column and enter a "10".

Whew! Now you have a strategy that combines a partial Martingale at zero and negative TC's with a straight counting strategy at positive TC's.

I simmed a 6D, S17, DA2, DAS game for a heads-up HiLo counter with 75% pen, and got a win rate of $145.06 with a 1.513% Initial Bet Advantage.

Hope this helps!

Dog Hand

There is some benefit to this type of play

It increases your spread, which will make you more money, but will also increase the level of heat.

More generally, I like the idea of making your bet increases look like something else. This is advocated in a number of well respected books (only increase after a win, steaming act, etc), and is something that alot of good players incorporate into their game.

As you said, you need to decide when you abandon the cover betting strategy, and this will depend to a large degree on heat, and how long you are hoping to play at that casino.

The most frequent issue I run into is when I am "steaming" late into a SD or DD and my nice positive count drops to zero or negative while my 20 pushes with the dealers 20. It is very obvious when I reach out and pull most of my chips off my bet, but very costly if I let it ride.

I am trying to refrain from using the "P" word since it is very charged around here, but I agree that the family of bet systems you mentioned help with your cover if used judiciously.

Sub-unit betting.

I've seen this work (for awhile) as a cover play, but using random/non-progressive/sub-unit bets until one feels "lucky" or "Pissed" enough to throw caution to the wind.

One problem is that you must play at low min. tables that tend to be crowded and whose max. limits may be too low. DD seems to work best in a play all game strategy.

CVData, demo?

Hey Dog,

Wow, thanks for the detailed help. Much obliged.

I have a licensed copy of Casino Verite Blackjack but I just use it for play practice, I never had a need for the simulator up to now.

I assume CVDATA you mentioned is the separate product from QFIT priced at $160? Can I do this with just their demo version OR do I need to buy the full thing?

IAC, thanks for running it for me. Bottom line result seems to show it may be a worthwhile strategy to keep in the bag of tricks ;-)

Nestor

P.S. I can't recall ever reading this idea anywhere else, in any of the BJ books I've read or even online anywhere. Is it possible I've invented it? Can we call it the Nestor Method? LOL

Kidding, of course: I'm sure others have thought of this before but I really don't remember reading it... I guess maybe I "invented it just for myself" in any case, but just curious WHERE and BY WHO it may have been written about before, in the past, if anywhere...?

CVData is, indeed, a separate product

Nestor,

I don't know whether or not the demo version can provide these results, but I would think not.

The problem with your method is that you are, in fact, spreading from $10 to $1,000: a 1:100 spread. While I have no objection to spreading like that, I doubt you'll be able to play that way unmolested for any extended period. Consequently, if to avoid heat you confine your attempts to short sessions, you might just as well spread right with the count and plan to leave after the first hot shoe.

I ran another sim, where this time you bet $10 at all TC<=0, and use the $100-$200-$400-$600-$1,000 schedule for TC's of +1, +2, +3, +4, and +5 (& higher): in other words, skipping the Martingale part altogether. This gives a higher win rate ($153.10 vs. $145.06) and IBA (1.727% vs. 1.513%) than the "cover" approach.

Although $8/hr doesn't sound like much of an improvement, the real improvement is in the IBA: if your BR is sufficient to support $1,000 max bets with the "cover" approach, it should be able to support an even higher max bet with the "straight" approach, which would provide an even better win rate.

Hope this helps!

Dog Hand

I think you'd do it for...

...longevity, rather than absolute optimum win rate.

In other words, sure it may be true that skipping the Martingale part altogether yields a better game, better return, but what I think the player would be "buying" with this cover is a very solid ILLUSION of being a typical, out of control, crazy-ass chasing-his-losses kind of gambler (loser).

If the pit really believes that's what you are... IMO it's well worth the lowered EV! Buying their belief that you're a "crazy gambler loser" is the whole POINT of all cover plays and cover "act", isn't it?

Once they see you martingaling regularly, they laugh at you and leave you alone (never mind that correlation of plus count to those times when you just happen to hit the Martingale wall at table max bets, heh... that's just a co-incidence... nothing to see here folks... move along.)

;-)

Martingale cover (nt)

Martingale cover

This has been brought up on other boards. I use a variation on this in a store which I has up until now responded fast to APs.

There are a few ways to camo a ramp to make it look like you are a simple martingale player (I'm sure there are many variations limited only by your creativity....)

Problem is SCORE plummets, as EV falls while RoR rises (as you are often not betting in correlation to the count).

Can U Quantify...?

>>> Problem is SCORE plummets, as EV
>>> falls while RoR rises (as you are
>>> often not betting in correlation
>>> to the count).

Hi BJ, can you quantify this somewhat? I mean, I know there's a COST to it, but question is really "is it worth it" overall for the strong impression it gives to the pit that you are "just a gambler" and not an AP. How much exactly does SCORE decrease, and how much more RoR? And can you offset the higher RoR somehow to compensate (bigger BR, or ?)

TIA for any more info you care to post :)

Martingale cover costs

Well, it's sort of like answering "how long is a piece of string"..

1. Cover value - subjective. Depends on your level of play, the store, etc etc.
2. Offsetting RoR - bigger BR, of course... Or a smaller ramp...
3. As to the actual cost in figures, that's hard to say. It depends how you structure your martingale cover - what your betting pattern is in neg-neutral counts. what it is positive counts, etc.

I did a fair bit of sim work to try and find a progression cover that I was comfortable with. In my case, I kept RoR at the same level while my EV was roughly halved.

Was it worthwhile giving up 50% of EV to play this store ? For a few reasons, for me it was (still showing a cumulative loss there :-)

D.

SCORE falls from 36.29 (straight counting) to 32.30 (using cover). (nt)