Flat Betting & Counting - Avoid Casino Detection

I wrote a blackjack simulator to test a card counting strategy that seems to yield an overall advantage while virtually eliminating detection by casinos. I'm curious if someone can double check the math, because if my simulator is returning correct results, it could help all players avoid getting backed off of blackjack games.

This uses a simple K-O count (2-6=+1, 7-9=0, 10 & A=-1). I simulated using my preferred game, which is 2 deck, dealer stands on soft 17, double after split, double on anything. You can find games like this at Aria in Vegas starting with $25 minimums, or Bellagio, Mirage, & Wynn/Encore with $100 minimums. The simulator shows returns of roughly +0.5% for these rules while applying this system.

You bet off the top of the deck, starting with a count of 0. If the count reaches +2 or higher during that first hand, you continue playing (flat bet) at that table until the count goes below +2. Whenever the count goes below +2 after the first hand off the top, you go to a freshly shuffled game or wait for them to shuffle at your table. Basically, follow basic strategy and flat bet until the count goes below +2. It's that simple.

Here's the thing: it's not really detectable by casinos. Yes, you will get in fewer hands per hour, but you are not varying your bet with the count, and you aren't entering in the middle of the game - you're playing off the top. Instead of flat betting, you could also randomly bet and achieve the same returns.

You could hit casinos really hard with this - no teams, just you. Anyway, someone please check my math. Using a single deck game with DOA (some are found in Reno) returns approach +1%, which is huge.

Please describe the random number generator you use and if the card distribution was as expected, the number of hands simulated and the langauage in which you wrote the simulator.

Thanks

It makes sense; however, you may end up only playing one hand per shoe -- fine if you are playing with a bunch of other players. But I wonder that with a count of +2 and 4-5 players playing, whether your advantage system would be subject to the change in count as cards are dealt to the other players at the shoe. I mean, sure, you can bet when the count is +2, but what if the count is -2 by the time the cards get to you?

Interesting system however.

This sounds like a variation of "wonging." Wonging, as a rule, is good for the player, because you only play when the odds are in your favor, but I think that eventually somebody's going to notice that you're betting "weird."

Quote: boymimbo

It makes sense; however, you may end up only playing one hand per shoe -- fine if you are playing with a bunch of other players. But I wonder that with a count of +2 and 4-5 players playing, whether your advantage system would be subject to the change in count as cards are dealt to the other players at the shoe. I mean, sure, you can bet when the count is +2, but what if the count is -2 by the time the cards get to you?

Interesting system however.

Actually, you wind up playing 1.4 hands per shoe on double deck - which in my opinion is even better. You look like a gambler just walking around, playing a few hands here and there off the top of the deck. This is so far from traditional counter behavior that casinos will never see you coming.

With regard to your concern about the count changing prior to getting to you, that's already been taken into account. It's simply based on what the count is before the cards are dealt for that particular hand - if the count is +2 before dealing, you stay, if not, you don't. Basically, you're just removing a large part of downward curve on blackjack.

Quote: hml48

Please describe the random number generator you use and if the card distribution was as expected, the number of hands simulated and the langauage in which you wrote the simulator.

Thanks

Sure. It's written in delphi (object pascal). Basically, the algorithm looks at the number of decks to test, and adds the exact composition of cards to a list (4 10's per deck, 1 A-9 per deck). That list is then randomized using delphi's internal random() function after being initialized with a random seed (a random number is selected for each card in the shoe, and each card is placed at the randomly selected position in the shoe). Cards are deleted from the list as they are dealt. Before each round is dealt (after the first round has been completed), the count is looked at, and if it is below the specified limit (+2 turned out to be the magic number), a brand new shoe is prepared and the whole randomization process is repeated. It should be noted that when I disable the counting (just let the simulator play through each shoe), the simulator returns a negative expectation that is in line with publicly available norms for basic strategy players with these game conditions.

I played with varying penetrations, ranging from 40-70%, and the results remained consistent. The number of hands simulated once I found the proper number (+2) was 10,000,000. +1 actually works as well but the increase in hands per shoe does not offset the decrease in positive return - you're better off with +2. I was unable to find any conditions under which this would achieve a positive expectation for 6 or 8 deck shoes.

Basically, this just eliminates a big part of the downward mathematical curve on blackjack. The fact that you are playing many hands at an advantage more than makes up for the fact that you are playing at a negative expectation off the top of the deck.

Quote: wildqat

This sounds like a variation of "wonging." Wonging, as a rule, is good for the player, because you only play when the odds are in your favor, but I think that eventually somebody's going to notice that you're betting "weird."

I know what wonging is, but the difference here is that you actually play off the top of the deck. Casinos know how to deal with wonging, and for that matter, team play, and know what to look for. We can thank MIT for that. If you are consistently walking into shoes in the middle of the game, you are sure to be watched and soon caught.

I'm not sure how this is betting "weird". You're just a gambler that doesn't like to stay at one table for very long. As I said before, if you don't want to flat bet, you can randomly bet (or play a normally losing "system" with this - martingale, etc.) and achieve the same returns. Another advantage is that your very short time at each table provides perfect cover for not tipping.

Playing only neutral and positive counts, I'm sure you will have a modest advantage. Personally, I don't like getting up and down all the time, but I have been caught lots of times as a result of my longer sittings. I'd encourage you to add some strategy deviations. I list them in order of strength in my page on the hi-low count.

Quote: Wizard

... I have been caught lots of times as a result of my longer sittings...

to what degree are you a recognizable person, surely this is a problem for you?

I'm curious if you've ever run your sim using a 2x or 3x bet after the first hand yields a +2. If you have, could you share those results? I realize this may increase the chance of being outted, but it still looks like "gambler" behavior, jumping in playing one hand, then ramping up on the second...

Quote: thefish2010

This uses a simple K-O count (2-6=+1, 7-9=0, 10 & A=-1).

Doesn't KO use 7 as a +1 also?