I decided to do a little bit of math and find out what percentage of the time my pet would have to lose melee attacks to boss movement for me personally(in 578 average ilvl) to see a dps gain going from a hast>mast>crit build to a crit>haste>mast build. I don't have BiS gear so results may vary and all that. Here's what I got:
Simmed on a 5min patchwerk fight with 10% variance
So the difference in dps with my gear is 3,024.
3,024 dps represents roughly 907,200 damage done in a 300 second fight.
Then looking at the pet melee damage in the haste build:
(moved the pet melee damage line to the top of the Abilities chart for convenience)
With the pet swinging for an average of 51,412 it would need to lose ~17.6 hits durring the fight to account for that amount of lost damage. Which works out to losing ~5.6% of its attacks.
That's actually a lot lower than I was expecting it to be. I'm not really sure how many times the pet loses attacks on a movement fight. Maybe someone who's good with logs can check and see how many swings pets lose for the heavy movement fights.
Now to be fair any fight where a haste build pet is losing swings a crit build pet has to be losing swings as well right? I figure the best way to compare fairly is to just assume both builds lose the same percentage of attacks to boss movement. So then here's the crit build pet:
(again moved the pet melee)
The crit pet had 300 swings so assuming it lost 5.6% of them(16.8 swings) at 49,605 damage per swing its a total damage loss of 83,3364. That translates to 2,777 dps. That means that if I use a crit instead of a haste build I'd lose 247 less dps if the pet lost an average of 5.6% of its swings. That's not really what I was looking for, so its graph time to find out what percentage of the time my pet would have to lose attacks to make a crit build better than a haste build for me.
Here's the pretty graph:
And all the numbers for it:
So with my current gear my pet would have to lose over 50% of its melee swings before the crit>haste>mast build became better than haste>mast>crit. Of course if the pet is losing 50% of its melee swings all sorts of other things are going wrong as well, like not getting off basic attacks and probably missing kill commands as well. At that point all stats would become heavily devalued. So far as I can tell the idea that haste is worse for BM because of missed pet attacks just doesn't work. At least that can't be the only reason for haste to be worse than crit. Before this is mentioned I should point out that a haste>crit>mast build would likely require more lost pet melees before becoming inferior to the crit build as the average pet melee would be lower. That's the logic as I see it, combined with some very, very rough math. There are also a few different ways to work through this, I chose the one that resulted in the most visuals. If someone wants to work through a different way that would be great. If my logic was flawed point it out and I'll see if I can correct it.
MASSIVE DISCLAIMER: All of this is done with my gear and only accounts for the affects of losing pet melee attacks, in no way am I saying that a haste build will always be better than a crit build. You should always sim your own stat weights.