Here in Texas, the state lottery folks have added four numbers to the fifty already in play to make the game "more interesting," or some other similar spin. They say it'll make the game more exciting because there'll be bigger jackpots to play for, and more often, too.
Of course, nobody among them is willing to demonstrate the truth of this statement mathematically, because that might affect sales. It may sprout a lynch mob or two, as well.
To wit, in a game that consists of guessing six out of fifty numbers, where the order of the numbers doesn't count, there are
50 x 49 x 48 x 47 x 46 x 45 = 11,441,304,000different ways the game can turn out. That's over 11
billion possibilities.
Of these, there are 6 * 5 * 4 * 3 * 2 * 1 = 720 ways to arrange the winning numbers. (If the winning numbers were 1-2-3-4-5-6, you could win whether the numbers were picked in that order, or in any of the other 719 possible ways.)
Hold that thought while I digress:
There is a concept in probability theory called expected value. Basically, you multiply your payoff - what you can win - times the chances of your winning and that's your expected value. If the expected value is less than your bet, the game is - all things being equal - not worth playing because over the long run, you'll lose money. If the expected value is larger than your bet, the numbers say you should play the game, because in the long run you'll win more than you lose.
Example: You play a game where you bet $1 and roll one die. If you roll a 6, you win back your dollar and another dollar. Is this a good deal?
Answer: No. Your bet is $1. Your "jackpot" is $2. The chance of rolling a 6 - and we assume the die is honest - is 1/6. If we do the multiplication ($2 x 1/6), we find the expected value of any roll is $0.33. If you play the game a thousand times, you'll very likely lose around $667.
What fuels gamblers everywhere is the prospect of having the win occur early in play. If you bet on a Las Vegas roulette wheel and hit a single number on the first spin, you'll be ahead. If you keep playing, though, the percentages will catch up with you and you'll end up paying the difference between what the house offers and what the wheel says.
End of lecture. |
If you consider that the odds of winning are 720 divided by 11,441,304,000 (or one in 15,890,700, a Real Small Number), and if you assume that the lottery you're playing for is $4 million, then the expected value of any ticket is:
720 * $4,000,000 / 11,441,304,000 = $0.2517which is
just over a quarter. This means that, for every bet of a dollar, you can expect a "return" of twenty-five cents, assuming the payoff is $4 million. (This analysis, by the way, does not take into account any of the lesser prizes in a state lottery, for example, getting four out of six numbers. The existence of those prizes does not materially change this analysis.)
So how large should the jackpot be for the bet to be "reasonable" by statistical standards? Take a look at the odds: As soon as the jackpot rises above $15,890,700, the percentages turn in your favor
in the long run.
When you add four numbers to the pot, the number of possible outcomes increases to:
54 x 53 x 52 x 51 x 50 x 49 = 18,595,558,800.That's a whopping
62% increase in the number of possible outcomes. With this set of balls, the expected value of a ticket for a $4 million jackpot drops to under 16 cents. (And yes, the pun is intended!)
The size of the "break even" jackpot - as far as expected value is concerned - is now $25,827,165.
It's only natural that pots will get larger, and that those big pots will occur more often, since it's now 62% more difficult to hit the right numbers!
State lotteries have been described as "a tax on the stupid," since - as I've attempted to show - if you played a dollar a week, you'd be lots better off putting the dollar in a bank at zero percent interest. At least after 20 years, you'd have a
guaranteed thousand or so dollars in the bank. Playing the lottery, you'd all-but-certainly have a goose egg, and a lot of bookmarks.
Cheers...
Note: This post was revised on 9/20/00 to correct some flawed analysis.