I try to do a split between games that I think will be competitive and games that have national appeal. My goal is to get good splits between the pickers so it actually makes it interesting. If everybody goes one way, it's not a lot of fun. But at the same time, I'm not going to ignore the biggest games of the week. Last week nobody happened to take Missouri or Texas, but I couldn't leave those games off. I use the ESPN weekly schedule, copy and paste all of the ones that I think look like games that might get a good split, and then I try to whittle it down to 20. Last week I was forced to cut games like UL-Monroe and Texas State, Rice and UTSA, and FIU and UAB. If it comes down to a big game or a competitive game between two mediocre teams, I usually side with the big games.
I did
some stat keeping a couple years ago out of curiosity and I found that the larger conferences tend to be more represented because I want to have everybody pick teams that they watch week-in and week-out or teams that they are more familiar with. Even so, I include every team at least once in a year (including ones that are transitioning, so that includes Old Dominion). I keep a running list of teams I haven't used yet and it's down to 10 left. Baylor and Louisville haven't made an appearance yet since they've been so heavily favored every game this year. Louisville has some of their bigger games on weekdays, so that hasn't helped.
Oh, right, that's my other rule: No weekday games. People have missed Thursday or Friday nighters in the past, so I want everybody to have ample time to get their picks in. And on the last week of the year, I use playoff games from the FCS to round it up to 20 since there's usually a slight shortage. I think one year I had to use a DII playoff game. I don't know why, I just like keeping it at 20!
So I guess to sum up, it's kind of subjective
. I go based on my gut feeling without using spreads. I usually get the next week's games posted late Saturday night after all the action is finished.