'Don't be deterred by the hitfrequency in video poker'

December 05, 2000 5:57 AM


Ever have a casino visit when nothing goes right? When you wonder where you got the idea that gambling was fun? When you figured the law of averages said things had to improve, but they didn’t?

My chum, Charlie, was mewling about this very state of affairs a few weeks ago. I found him meandering aimlessly down the video poker aisles of the Beachcomber.

"Jefe," he whined, "they’re killin’ me. I put $20 into a quarter jacks-or-better machine and ran through the whole thing, five coins at a time, without one hit. The odds against that must be at least a million to one."

"Hmmm," I sympathized. "Muy mal, amigo. I don’t think I ever bombed entirely that badly. I wonder what the odds really are."

So naturally, when I got home I revved up my computer to do the ciphers. Charlie’s pretty good at video poker. So his chance of any return at all, based on expert strategy, should be 45.44 percent, give or take. What would be expected if a million players, each this proficient, fed $20 into quarter machines and bet five coins per round? It turns out that 62 out of the million could count on no hits whatsoever. Not quite the odds of a million to one that Charlie speculated. But not far off, either.

When the computer had finished chugging away on this calculation, I put the pedal to the metal and went a bit farther. A few related questions came to mind:

    1. How many of the million players could expect one, two, and so forth up to 16 hits in 16 tries?
    2. Recognizing that a high pair only returns what you bet and is therefore a push, how many of the million players would be forecast to get zero, one, two, et cetera actual wins -- two pair or better -- in 16 tries?
    3. And, what if you raise the bar and ask how many of the million solid citizens would be predicted to have zero up to 16 returns of three-for-one (triplets) or better in 16 rounds? The accompanying table gives the answers.

Number of players out of a million expecting indicated number of hits in 16 tries at jacks-or-better video poker

hits jacksor
or better
2 pair
or better
or better
0 62 12,415 153,580
1 822 62,692 305,261
2 5,132 148,399 284,413
3 19,947 218,574 164,882
4 53,990 224,203 66,569
5 107,918 169,829 19,847
6 164,777 98,268 4,520
7 196,049 44,307 802
8 183,688 15,732 112
9 135,985 4,414 12
10 79,278 975 --
11 36,014 168 --
12 12,498 22 --
13 3,203 2 --
14 572 -- --
15 63 -- --
16 3 -- --

As one example of interpreting this data, 62 out of the million players can expect to get zilch in 16 tries, as stated earlier. But 12,415 will experience no actual wins -- hands of two pair or better -- in this many rounds. And a substantial 153,580 will never see triplets or above in 16 shots.

At the happy end of the scale, three out of a million can expect a return of some kind on every round. Conversely, essentially none ought to hold their breath for 16 winning hands in a row.

The table also shows intermediate numbers of hits. For instance, 183,688 out of a million players can anticipate any return -- from a push on up -- exactly eight times in 16 tries. Only 15,732 out of a million are projected to get precisely eight actual wins, as opposed to returns comprising pushes, in 16 tries. And a mere 112 can expect eight hands ranked as triplets or better.

I discussed this with a blackjack buddy, who smugly said, "That’s what he deserves for playing the machines." So I ran blackjack numbers, too. With six tablemates in an eight-deck game, you’ll average about 16 rounds per shoe. Out of a million players, 35 can expect to lose and six to win all 16, respectively. Better than video poker. But hardly sufficient to merit bragging rights. And one video poker hit might pay 800 for 1, while blackjack by and large yields only even money. The poet, Sumner A. Ingmark, had these wise words about certain fine distinctions:

When a chance is small enough,
Twice the chance seems just as tough.