Analyzing casino games seems a lot like walking a tightrope
November 11, 2014 3:00 AM
by Elliot Frome
This past week Nik Wallenda walked across a tightrope between two Chicago skyscrapers blindfolded. I’m not all that impressed.
I think he should try analyzing casino games for a living. Most days it seems a lot like walking a tightrope blindfolded, too. Only, if I mess up, I don’t just hurt myself.
I might cause a casino to lose a lot of money if I allow a game to hit the floor with a Player advantage. After more than 10 years in the industry, I’ve never done it, but I have heard cases where it did happen due to a math mistake.
Before I analyzed casino games, I was a developer of large-scale computer systems for a company in the health care industry. While those systems were far more elaborate and complex than any game, they also had clear discrete results.
If the client sent an “M” for male and “F” for female, we knew our main database had to be updated to show the same results. When you go to an ATM and try to withdraw $20, there is a clear proper answer for how the machine should respond when you say “Withdraw” and “$20.”
If while testing it, the machine spits out $100, something is wrong. If you type in the right password/pin and it says you didn’t, the programmer knows there is a mistake and can correct for it.
When you analyze a game, you never know what the answer should be – unless you’re analyzing a known existing game. You might misunderstand a part of the game or make a simple programming mistake that spits out the wrong strategy. There really isn’t an absolute answer you can look for.
Sure, if the Player’s hand outranks the Dealer’s, then the Player should win. But, maybe the programming that handles the blind payout has a bug in it. It might not be nearly as easy to find as the issue when an ATM spits out $100 when it should spit out $20.
You can go through the entire analysis and think a game has a 98% payback when in reality it has 102%. What happens when the game hits the casino floor? For a while it might be okay, especially if strategy is involved. Human error will likely cause the game to stay positive for the house, but not nearly as much as it could be.
Eventually, however, somebody out there is going to figure out the mistake and do whatever he can to exploit it. That’s when it gets ugly. It is very scary to think that a game I analyzed will ever get to that point.
A few months ago, I began discussing a game called Free Bet Blackjack. It is a blackjack variant. The Player is allowed to Double Down on 9, 10 or 11 for free. He simply requests the Double Down. The Player does not need to make another wager, but if he wins, he will be paid as if he did.
The same goes for splits (except 10’s/faces). The Player can even re-split (to 4 hands) and Double Down after a split – all for “free.” The Dealer will place a plastic lammer as a place holder for where the Player would normally make another wager.
If the Player loses, the Dealer collects the lammer. If the Player wins, the Dealer takes the lammer but pays the Player for it as if it was a real wager. Of course, you don’t get something for nothing. The game uses the Push 22 rule whereby if the Player has not busted and the Dealer busts with a 22, the Player’s hand will be a push instead of a win.
When analyzing the game, I looked at hard 9, 10, 11 and all the Pairs. The goal was to determine when the Player should take the Free Double Down and/or Split. It didn’t take very long to do figure out the strategy. The next step would be to simulate the game and I’d be done.
Except, I missed something!
The good news is I was analyzing the game only for my column and I caught the mistake before anything was published.
Once the Player gets a Free Split, he is playing only for wins. You might think he is always doing this, but pushes happen a lot in blackjack and they most definitely impact the strategy.
In Free Bet, if you’ve split a hand and have a plastic lammer up as your wager, you only get paid if you win the hand. If you push, the lammer has no value and the result is no different than if you lost.
When a Player splits, the first hand keeps the original cash wager and the second (and subsequent hands) have only a lammer as a bet. So, if you split 8’s, and get a 10 on the first, you have an 18 with real cash on it and would play this as per normal Push 22 strategy.
You might draw a 9 on the second hand, for a total of 17. What if the Dealer has a 9 showing? Normally, you would push if the Dealer has an 8 underneath. But, since this is hand No. 2 with only a lammer as the wager, a push is the same as a loss.
If you push, the Dealer is going to take the lammer and you get paid nothing. As a result, our normal strategy goes out the window and we are supposed to hit. But, only if it is not the initial hand with the real wager on it. If the 17 was dealt to the first hand (or a non-split situation) and the Dealer has a 9, you would stick.
I was fortunate to catch the error and that no harm was done. I hope to keep that track record up.
Buy his book now!
Elliot Frome is a second generation gaming analyst and author. His math credits include Ultimate Texas Hold’em, Mississippi Stud, House Money and many other games. His website is www.gambatria.com. Contact Elliot at ElliotFrome@GamingToday.com.