I know there are one or two interested in this sort of thing on the Saints board,so I will share. In our weekly Premiership score prediction game two posters chose exactly the same ten scores as each other.We were trying to work out the likely odds.
Probably not as unlikely as it initially seems. With two players it's about 1 in 10,000,000+, but as more people play the number of possible comparisons goes up by n squared, and the odds of matching two drop a lot. Edit: done it. p = 1-(1-q[sup]10[/sup])[sup](n(n-1))/2[/sup] p = probability of two matching predictions q = probability of matching two individual results (debatable but something like 0.2?) n = number of players edit2: cocked up, fixed now
They have similar prediction methods. I don't know if they play it safe or predict a freak result, but if it's the former and the game is, say, Chelsea vs Fulham, then they're probably both going to have a comfortable win; I'd go for 2-0, personally. In relation to the matching all 10 scores: If two people each have a selection of three lights to randomly select*, then at first there's a 1/3 chance that they'd get the same one each. After correctly guessing, it becomes 1/9. After correctly guessing again, it's 1/27, and then 1/81, 1/243 etc., so to get 10 same scores the odds must be quite high, especially as someone could (in theory) predict a game to finish 38-14. *Saw this on Derren Brown and they kept choosing the same lights.
Right. The actual odds are lower than it appears if you just do math, because people aren't going to pick scores like 1,2756 to 452. So really you're looking at picking numbers between 0 and 4. And even then, there's a tendency to duplicate scores because most people will tend to agree on which team is the favorite and by how much. The amount of variance in projected scores between two reasonably astute football fans is probably fairly small.
