Did you run a simulation with no cone at all, but where you moved the cg to obtain the same stability margin as shown when the simulation has the cone?
That's what @neil_w and others are proposing as the correct method.
View attachment 546514
That adjustment made no difference in my case; the result is the same as with the cone removed and CG left as-is since there's enough margin. But, thank you for pointing it out since I did not do that initially. I suspect that since mine aren't going that high, the difference in delay is not statistically significant; but if I started pushing things higher, then the cone-based sim may make the delay too short. It also makes me wonder who else could be ending up with short delays due to Thrustcurve sims (or are they not?) I will definitely need a better altimeter and more testing. Who wants to donate a pile of 38/240-480 motors to my testing?