I apologize in advance for the rare non-gaming post and assure you it will not be a frequent occurence. But I have a question for any statisticians out there. Here goes.
Every week, I choose 10-15 professional football games to predict. What I do is select who I think will cover the established gambling point spread. So I don't pick winners or losers. I try to pick who will cover the spread. In theory, this should be a 50% proposition. So if I was a trained monkey randomly picking which team will cover the spread, I should get about 50% of the games right. In fact, Las Vegas sets the spreads so that approximately 50% of the money gets placed on each team.
Well, the season is over and my picking record is 119 wins and 83 losses over the course of 202 games. That is, I correctly predicted which team would cover the point spread 59% of the time. I was thrilled with the result and bragged to my brother about it. I told him that most NFL analysts on TV don't pick point spread winners correctly 59% of the time.
But Dominic (my brother) rained on my parade by saying his gut instinct told him it was not a statistically meaningful sample since I only predicted 202 games. I told him I thought that was enough games to draw some inferences about my handicapping ability! I said "Dom, if I was to flip a coin 202 times I doubt it would come up on a particular side 59% of the time."
So does anyone know if 202 games is enough to draw a meaningful conclusion about my ability to pick games? Or could it be pure luck? I promise my feelings won't be hurt!
PS: I didn't actually bet a single penny. I'm not really a gambler. This was all to see how well I could do at picking point spread winners.