say you think you can regrab 80% of the time on reaction

say on the other hand that you think you can stomp 50% of the time on a combination of reaction and prediction.

constant regrabbing gives like, I dunno, let's say 8% per dthrow (too lazy to look it up).

stomp combos into knee, which, ignoring the fact that it will just kill people a lot of the time, does like, I dunno, let's say 32%.

so like, you'd need to hit 4 consecutive grabs to be equivalent to a stomp (hypothetical, but w/e, I think it's pretty close to reality).

chance of hitting 4 re-grabs in a row = .8 * .8 * .8 * .8 = ~41%

chance of hitting a stomp = .5 = 50%

so like, stomping's better with these made up numbers, but I'm sure if you were to extend it to a more advanced model, you'd see an even bigger disparity between stomping and re-grabbing all the time.