Originally Posted by
POOLSIDE
I believe this is only volume, correct? As in, 380 people chose blah blah, 354 chose blah blah. Have you ever considered looking at the percentages? I feel like there could be value here. Bear with me.
For example, at a glance, I think the number one fade this week should be Indy. Yes, 373 took Washington and only 352 took Indy, but look at the other side of those games. 123 people chose Philadelphia while only 96 chose Jacksonville. Broken down into percentages of the people who chose each game, those two games look like this:
Washington - 75.2%
Philadelphia - 24.8%
Indianapolis - 78.5%
Jacksonville - 21.5%
Clearly, OF THE PEOPLE WHO PICKED THESE GAMES, a slightly higher percentage felt comfortable betting on Indy.
This might not be the best example, as the two sets of numbers are close either way. However, the percentage can be relevant.
As an example, let's say 373 people chose Washington and they were the most popular pick. But let's also say that 338 people chose Philadelphia and they were the sixth most popular pick, which would not have made the top five list. According to this, we'd be fading Washington and selecting Philly. But these numbers are so close to 50/50 that I don't see any value either way.
On the other hand, let's pretend (to be clear, I'm making these numbers up) only 235 people chose Washington. This wouldn't make the list. For the other side of that game, let's say only 25 people chose Philly. That says less that 10% of the people that chose this game, chose Philly. That seems like a relevant number, but only time will tell.
I'll do a percentage breakdown for all the games in another post. It'll take a minute, but hopefully it'll be worth it. I'll also list them in order of largest percentile discrepancy to least, which would hypothetically be a fancy way of saying most relevant to least relevant.