Even skilled players lose sometimes

No doubt you've heard that certain skilled players — blackjack card counters and video poker experts — can swing the math and gain an advantage over the house.

This is true. Card counters with sufficient skill, discipline and bankroll can beat the house. Well-funded video poker players who stick to the best-paying games can be long-term winners.

But skilled play is no guarantee of winning at any individual session. The elements of chance are still too strong for that. Some days, it's just going to seem like every blackjack hand's a bust and the dealer pulls a 5 to turn every 16 into a 21. Even the famed MIT card counting team had some large losses before getting on a roll.

In video poker, losing sessions outnumber winners, no matter how good you are. Royal flushes account for about 2% of our long-term return — a bit more or a bit less depending on the specific game and pay table.

When we say 10/7/5 Double Bonus Poker returns 100.17% with expert play, we're including a normal frequency of royal flushes. Since royals come up an average of only about once per 48,048 hands and account for 1.67% of our payback in 10/7/5 Double Bonus, we're only playing about a 98.5% game between royals.

It's possible to win while playing a 98.5% games. Heck, there are a lot of games worse than 98.5% in casinos, and players do have winning sessions. All it takes are a few four-of-a-kind hands. If one can be four aces for an 800-coin payoff on Double Bonus, or four 2s, 3s or 4s for 400 coins, so much the better.

But the video poker expert trying to get an edge on the game isn't looking for an occasional winning session and a long-term bankroll drain. Winning sessions during non-royal periods are welcome, mind you, but the experts are counting on those rare royals, worth 4,000 credits for a five-credit bet, to overcome a slew of small losing sessions.

In the pro's world, there are more losing sessions than winners, but the big winners more than offset the losers. The long-term payback percentage cracks 100%.

To get to that profitable long run means being able to sustain the short-term losses. It takes money.

How much? I ran the numbers through Bob Dancer's Video Poker for Winners software. If you start with a $100 bankroll on a 25-cent game, the chance of losing it all on 10/7/5 Double Bonus comes up as 99%.

Hey, you might get extremely lucky, hit a quick royal and have a bigger bankroll to work with. But as a rule, if you're just starting with a few hundred dollars and play long enough, the house will take your money.

To reduce the risk of ruin to even 50% requires a starting bankroll of $6,974. In other words, if you're a pro playing with an edge on this game, you're betting five quarters at a time, you still have a 50-50 chance of going broke. For a 10% chance of going broke, increase the bankroll to $23,168. For 1%, the necessary bankroll is $43,336.

No matter how good you are, giving yourself a shot at long-term profitability means you need to be backed by plenty of cash. It takes money to make money.

** ** **

Every time I refer to video poker games by numbers such as "10/7/5 Double Bonus Poker," I get e-mail asking just what the digits mean.

The most common way for a casino and game manufacturer to change the payback percentage on a video poker game is to change the returns on full houses and flushes. That makes the one-coin payoff on those hands an easy identifier for players looking to differentiate between games.

Each unit change on those hands alters the payback percentage by about 1.1 percent. So we know at a glance that with the rest of the pay table being equal, 9/6 Jacks or Better, where full houses pay 9-for-1 and flushes 6-for-1, pays about 2.2% more than a common variant, 8/5 Jacks or Better.

Double Bonus Poker is a special case, because in the better versions, straights pay 5-for-1 instead of the 4-for-1 usually paid on non-wild card games. There are some Double Bonus machines that pay only 4-for-1 on straights, and those are games to avoid.

So on the full-pay version of Double Bonus, "10/7/5" means the game pays 10-for-1 on full houses, 7-for-1 on flushes and 5-for-1 on straights, and returns 100.17% with expert play. The most common version, 9/6/5, knocks paybacks down on full houses and flushes, leaves straight pays intact, and returns 97.9% with expert play. Watch out for 9/6/4 Double Bonus games, where losing a unit on the payoffs on straights knocks the return down to 96.4%.

Do watch out for changes elsewhere on the pay table. I once ran across a Bonus Poker machine that looked like it had a solid 8/5 pay table until I noticed three of a kind payed only 2-for-1 instead of the usual 3-for-1. The difference? A horrendous 91.7% game instead of 99.2.