Need a math geek. (MNK)

[Powercow]

Running the numbers at base skill speed, the gap turns from ~27 pps to more like ~10 pps. It's still better to do it than not, but by a much slimmer margin. I believe that margin increases proportional to skill speed, because of the nature of how these numbers play together (skill speed doesn't affect AA frequency, and faster skill spins will get you better AA potency due to hitting more gcds between AA hits).

I did the math back when 3.0 was datamined and came to pretty much the same conclusion -- if utilized *perfectly* it's a single digit DPS gain. To use it perfectly the following must all be true:
1 - The combo you'd replace was a bootshine/true/snap combo. No buff/debuff/demolish refreshes.
2 - No Touch of Death, Howling Fist, Shoulder Tackle or Elixir Field uses in that time.
3 - There must not be a point in the next 3 minutes of the fight where PB is useful.
4 - There must not be a point in the next minute where Tornado Kick would be useful.
5 - Any applicable buffs/debuffs (Balance, Trick Attack, Blood for Blood, etc.) must be used with Tornado Kick but *immediately* wear off before they'd affect the next skill or autoattack.
6 - At no point can your rotation be interrupted for fight mechanics or the presence of adds that last more than a few seconds.

Failing any of these the incredibly tiny DPS gain is lost and you'll start to lose DPS, and the more of these you get wrong the more your DPS would be lower than if you just did the normal combo. Yeah the relative gain does increase if you stack Skillspeed but, well, if you're intentionally stacking skillspeed then I don't think you should look to Tornado Kicking to squeeze out that last theoretical 0.01% DPS when a switch to crit/det would net you a far, FAR larger gain.

tl;dr: Not worth it.

According to some of the people who got to play an early version of HW at a convention the potency used to be higher (like 500), but was reduced before launch. Given the incredibly close nature of the potential gain I think 330 is the highest it could possibly be without being worth it in a normal rotation. Either someone at SE got really lucky with a random guess or, maybe, somebody there did the math and figured it out too. My guess is the latter.