Q. I was poking around on the Internet, looking for information I could use in blackjack, and I came across a piece Al Krigman had written couple of years ago about splitting 10s in blackjack. His conclusion was that you shouldn't split 10s, which is the same thing you've written, and I've seen from Henry Tamburin, Frank Scoblete, Stanford Wong and pretty much everyone I've ever read on blackjack.

He had a chart showing the average returns for standing on 20 and for splitting 10s against every dealer up card, and something bothered me. I wondered if you could explain.

It says that if you stand on 20 against an 8, your average result would be a profit of 79 cents per $1 wagered. But if the dealer's up card is 7, your average profit is only 77 cents per $1. Why is that? Every chart I've ever seen says the dealer makes more hands starting with an 8 than with a 7.

A. The target here is beating or pushing your 20, not merely making a standing hand of 17 or better. If the dealer starts with an 8, there are six card denominations - 9s, the four 10-value cards and Aces - that will result in standing hands, and all lose to a 20. If the dealer starts with a 7, only the four 10-values and Aces result in standing hands.

That means the dealer hits more often, and has more chances to get to 20 or 21, with 7 up than with an 8.

You can see the effect throughout Krigman's chart. The average return for standing on 20 vs. a dealer's 2 is 64 cents per $1 wagered, increasing to 65 cents vs. 3, 66 cents vs. 4, 67 cents vs. 5, 70 cents vs. 6, 77 cents vs.7 and 79 cents vs. 8. The return starts decreasing when the dealer has a 9 up. That's because a one-card draw - an Ace - results in a 20. That's enough to drop your return to 76 cents per $1 wagered.

To finish the list, the return is 56 cents per $1 when the dealer shows a 10 value, and 65 cents when a dealer shows an Ace. All are much higher returns than you'd get for splitting the pair. The highest return for splitting the 10s is 43 cents per $1 in the original wager when the dealer shows a 6. With a $10 bet, your average profit is $7 if you stand on 20, but only $4.30 if you split the 10s. You risk $10 more, but average a profit of $2.70 less. That's why you don't split 10s.

Q. In plain dollars and cents, how much should I care if a casino has 9-6, 9-5 or 8-5 Double Double Bonus Poker? I play dollars, and the casino my wife and I really like has 8-5 DDB, but if we go to one a little farther away with a buffet we don't like quite as much, I can play 9-6. You talk in percentages a lot, but what does that mean in dollars?

A. Let's talk percentages first, then convert to dollars. Assuming a standard pay table on other hands, 9-6 Double Double Bonus returns 98.98 percent with expert play. Drop the flush payoff to 5-for-1 for a 9-5 game, and the return drops to 97.87 percent. Reduce the full house payoff to 8-for-1, leaving 8-5 DDB, and the payback percentage drops another notch to 96.79 percent.

What does that mean in dollars? If you make the maximum bet of $5 a hand on a dollar machine and play 500 hands per hour - a steady but easy pace on video poker - your average risk per hour is $2,500. Your average loss per hour on a 9-6 Double Double Bonus machine is $25.50. That rises to $53.25 per hour on 9-5 DDB and $80.25 on the 8-5 game.

Session results will vary wildly. You can lose hundreds per hour on any pay table when the cards run bad, but a $4,000 payoff on a royal flush or $2,000 on four Aces with a low card kicker will make up for a whole bunch of losing sessions.

But your average losses will be roughly $55 higher per hour on 8-5 DDB than on the 9-6 version. It's for you to decide whether the better buffet and any other features at your favored casino make up for that.

Gambling author and columnist John Grochowski's weekly newspaper column began at the Chicago Sun-Times and is now syndicated nationally. He also regularly makes TV and radio appearances about gambling. His column appears weekly.