I don't know what the correct value is, but you never re-time a mov.t based on the lift angle, which has no influence whatsoever on rate and beat error.
Only amplitude is influenced, by about 1.5 deg of instrument reading for each one of lift angle, but since amplitude can vary easily by 5 - 10 deg depending on power charge and position, that is also negligible.
Now, this discussion has been had already, complete with formulas and experimental confirmation, as soon I can find the original thread I'll merge this one into it.