2022年6月30日 星期四

Paradox of "Try One More Time"

Can we buy more tickets to get more chances of winning the lottery? 


The host can design a game with a probability of p and an award of R.
Assume that N players will come. The ticket for each is T. How much should T be?

There is a cost for each game, C. 


NT - NpR - C > 0


Thus, the host can expect the income N(T-pR)-C.

T must be larger than C/N+pR, T > C/N+pR, otherwise the host loses money anyway.

For a player, one will be attracted if R > T and be very attracted to buy a ticket if R is much larger than T, R >> T.

When N is large and very large, the outcome will be close to the expected value, N(T-pR)-C. The more N, the closer the outcome. The host can obtain the expected outcome when N is large. 


What is the expected outcome for the player?

Suppose that he buy n tickets. 

He spends nT and expects to gain npR. The net gain is expected to be npR-nT or n(pR-T).


Note that, T is larger than pR, that n(pR-T) is negative. The net expected gain is negative for the player, n(pR-T) < 0. In other words, the outcome is expected to close to this negative value when the player buys more tickets.

For the host, he tries all efforts to attract more players, makes N large. 

He knows that there must be some player who gets the award. However, he does not care who this lucky person is because his expected income is N(T-pR)-C.  


The lucky one might be he, she, or YOU. Yes, it might be YOU.

If you think this way, you are trapped in "over-weighting of the rare event".