Roulette's worst bet

For many years, I've warned roulette players that there's one bet on a double-zero wheel that's more hazardous to the bankroll than the rest.

You can bet a single number and face a house edge of 5.26%. Make a two-number split: 5.26%. A four-number corner, a 12-number column or dozen, an 18-number red or black, or just about anything else, and the house edge is, yes, 5.26%.

There's just one wager where the house edge is different. Make the five-number bet on 0, 00, 1, 2 and 3, and the number soars to 7.89%. Instead of the house expecting to keep an average of $5.26 per $100 wagered, it expects to keep $7.89%.

Ouch.

I've cautioned roulette players against making that bet a number of times, but my e-mail recently brought a request for more explanation. The writer wanted to know, "Why is the house edge higher on this one bet?"

There's a question of practicality involved. Casinos can't pay off in whole numbers per dollar and get the same house edge as on other bets on a double-zero game.

Casinos pay 6-1 on the five-number bet. To get the same 5.26% edge as on any other bet, they'd have to pay 6.20-1. They're not going to set up payoffs in a way that would require them to keep small change. It'd be bad enough paying $6.20 for a $1 bet, but at a casino that allows you to break chips down to quarter denominations, a $1.55 payoff for a quarter bet just isn't going to happen.

The other way to take the question is a matter of how the house edge is different given that 6-1 payoff.

You can see it in an example. Let's say you bet $5 on the five-number bet, and let's say the 00 hits. You collect $30 for your 6-1 payoff, and you keep your $5 bet, leaving you with $35.

Now let's say that on the same spin, I bet $1 each on 0, 00, 1, 2 and 3. I lose my dollar each on 0, 1, 2 and 3, but on 00 I collect $35 on a 35-1 payoff, and I keep my $1 bet, leaving me with $36.

We've bet the same amount of money on the same numbers, but because I bet it on single numbers and you bet it as the five-number bet, I have $1 more than you do. That's the difference in the house edge.

You could also choose to bet the five numbers as a two-number split bet on 0 and 00, and a three-number street on 1, 2 and 3. Let's stay with a $5 total, but as a $2 bet on the 0-00 split and a $3 street on 1-2-3. If 0 or 00 comes up, I win at 17-1 odds. I collect $34 in winnings and keep my $2 bet for a total of $36. If 1, 2 or 3 comes up, I collect at 11-1 odds for a total of $33 in winnings, plus I keep my $3 wager for — you got it — $36. Either way, I have $1 more than you do for betting the same amount of money on the same numbers.

To take it a step further, there are 38 spaces on the wheel. If the wagers were even propositions, you'd have $38 after your winner, and I'd have $38 after mine. My $36 leaves a $2 shortfall. Divide that by $38, then multiply by 100 to convert to percent, and you get a 5.26% house edge. Do the math with your $3 shortfall instead, and the house edge is 7.89%.

Let the bettor beware.

** * ** * **

One thing to keep in mind is that the caution about the five-number bet does not apply if you're playing at a single-zero wheel. If there is no double-zero, the five-number bet does not exist.

The single-zero wheel is a better play than the double-zero version, with a house edge of 2.7% on each wager. Problem is, single-zero roulette is not common in the United States. Most of the single-zero wheels I know of are in high-limit rooms. I don't remember the last time I found a single-zero game that fits a low-roller's budget.

Casinos that cater to high-end roulette players sometimes tack on interesting rules that reduce the house edge even more. In Europe, you'll sometimes find a rule called "en prison." On an even-money bet such as red/black, even/odd or first 18/last 18, if the ball lands in zero you don't lose right away. Instead, your wager is held in prison for another spin. If your wager's a winner on the next spin, you don't get paid, but you get your bet back. That drops the house edge to 1.35%.

A similar play that has been offered in Atlantic City takes only half your even-money bet if the ball lands in zero. Again, the house edge drops to 1.35%.

As always, I caution you not to overbet your bankroll, but if you find such a game that fits your wagering comfort zone, you'll get a better deal than on a double-zero game.