This is regarding key numbers in the CFL.
The first problem is that sample sizes are *really* small. I mean, it's small for the NFL with 256 games a year, but the CFL has less than a third this number of games per year.
I've got CFL data back to the 2001 season, and I've got push percentage charts for that margin. (I've also got push percentage numbers based on limited line ranges as well, but these are based on such small sample sizes I don't really have any faith in them.) Skeeter is basically right. The value of numbers is much flatter in the CFL than the NFL or even NCAA. The 3 is still a key number, but based on what numbers I do have, but much less so than in the NFL. The second most common
margin of victory is 2. Third is 7, although, statistically, 1,2,4-8,10 are all basically in a tie for second place after the 3. There are several curiosities in the data I assume are due to random chance (I find it hard to imagine that an 18 point final margin will be more than twice as likely as a 14 going forward.)
A move from -2 to +1.5 is worth closer to 20 cents. About the same percentage of games fall within a 10 point margin in the NFL as CFL, so, basically, you're just flattening out a similar looking distribution. It's easy to get carried away with this, though. Plus, as counter-intuitive as it is to realize that in the CFL the 5 and 2 aren't dead numbers, it's also entirely possible to overcompensate for this.
In any case, I expect serious bettors to have ready access to this sort of information.