Using Wayne Zimmerle’s MSA program, I did a small analysis of players that were Experts, Class A, and Class B. (Yes, dear readers, I now have some of your tournament histories in convenient form.)
The surprising thing about the (few) Experts studied is that they have an absolutely horrible record against players rated at least 100 points higher. Something like a 9:1 loss/win ratio. Class A and B had far better success by this metric. But the Experts’ ability to win against players they “should” beat (at least 100 points lower) was much better, capturing about 85% of the points (class A was about 75%, class B about 65%).
There are, of course, multiple possible explanations. There’s probably a greater difference in average rating between the Expert and his sub-100 opponents and the class player’s sub-100 opponents. And could it be possible that 100 point difference means more at the 2000+ level than at the class level?
I don’t know. But based on this small sample size, it does seem the ability to “scalp” higher-rated opponents correlates poorly with overall skill. It’s more important to perform at a strong level consistently.