Guilty as charged on focusing blindly on arithmetic rather than the conceptual underpinning. Ok, after that poor reasoning, here is a simpler way to show "it depends" (instead of the incorrect assertion that fixed WS delay alone is not enough to "invert").
Suppose the damage rate (also accounting for WS damage) for weapon A is a, and that for weapon B is b, and that a > b. Now, suppose that it takes time t1 for weapon A to WS, and time t2 for weapon B to WS, and t1 < t2.
After accounting for some fixed WS delay d (delay >= 0), we cannot say whether the adjusted damage rate for weapon A, a*t1/(t1 + d), is still greater than that for weapon B, b*t2/(t2 + d), because [t2/(t2 + d)]/[t1/(t1 + d)] > 1 as t1 < t2. It depends on the difference between weapon A and weapon B damage rates (a, b).
For Byrthnoth's example, t1 = 40/3 seconds for Soboro and t2 = 20 seconds for Hagun. With the implied 1-second WS delay (d = 1), [t2/(t2 + d)]/[t1/(t1 + d)] = 1.023810. For the 1-second WS delay to bring Soboro on par with Hagun, it follows that Soboro must be ~2.38% better than Hagun without the WS delay (based on ratio 1.023810).
Without WS delay, Soboro's WS frequency is 50% higher than Hagun's, and with the WS delay, 46.51%? It follows that the reduction in Soboro "efficiency" is 6.98% and the reduction in Hagun's "efficiency" is 4.76%.
Of course, human reaction time can also play a role.
XI Wiki



