Interesting ...
In charting it out... in my model the slope of the first number (a) is flat
and the slope of b is very negative crossing the x axis at about 8.
In your model the slopes are more evenly distributed and mirrored images of
one another. Even with all that, in a 5000 example sample this is the way
they differences between the numbers broke out
Difference meth 1 meth 2
0 494 526
1 1393 912
2 941 774
3 662 724
4 479 595
5 398 534
6 264 359
7 190 292
8 121 194
9 58 90
Total 5000 5000
Good call David, but I gotta get out of these stats meetings.
Bill D.