Statistician S, engineer E and mathematician M were going to bet in a horce race, each of them in a jolly good mood.
S: I’ve figured it out. Usually the odds reflect the probability of winning fairly well, but my analysis shows that this time there’s a slight disparity. My scheme has a 95 % probablity to bring me 10 % profit.
E: My system is even better – I’m going for the winners. I have studied the individual characteristics of the horses, their performance in the past, the condition of the track, today’s weather and about a dozen other things. Computer simulations give me a 87 % probability to at least double my bets, and a 98 % for a 50 % profit.
M: Poor boys, I’m a positive winner. I have a scheme that builds progressively on itself, and the odds are less than one in a million. But in truth I’m betting against myself – I’m testing a hypothesis, and if I lose, I will get the Field’s medal. So that’s really the odds that I don’t get the medal.
In the evening they meet in a bar, each one in a bad mood.
S: I lost everything. Well, there was always the possibility.
E: I lost everything. I didn’t measure the initial conditions well enough and the chaotic system went heywire.
M: This round is on me. With all the work to show that at any given time during the race the number of horses on the track is not uncountably infinite, I didn’t prove the continuum hypothesis.
[For those not in the know: It is known that there are at least two infinities: countable, which is the biggest number, and uncountable, which is the amount of real numbers. It is known that uncountable infinity is bigger than countable. Continuum hypothesis states that there are other infinities between the two. The continuum hypotheses would have been proven, if at some instant there had been more than countably infinite horses on the track. He won the bet because there were at most countably infinite horses in the race.]