Thanks for the feedback.
I can definitely see the value, both from a historical and a repeatability stand point to using my non-ideal 20 hill test * .95 (what is ideal?)
The reason why I asked is because I train to race, and if I'm getting higher numbers from a race event, then shouldn't I use those numbers instead?
It would seem that's what I need to replicate (or try to) in my training. Granted, it's only a 15 watt difference, but in the world of marginal gains this one stuck out....