I mean the tl:dr is that you should trust your sims over me. That said, whoever is telling you to run stat weights is probably using very old habits, because stat weights are a bad sim measure and are in constant flux. Running top gear or the gear compare tool will give you better ideas of how changes will affect your DPS.
With the changes to mastery last patch, I’d be interested if you want to push even remotely close to 100%. We’re obviously geared differently so this isn’t apples to apples, but mastery is my least valuable stat at 72% currently.
In terms of haste, I personally wouldn’t let mine drop below 18% even if it meant dropping a little DPS because the spec plays like sludge below that point imo.
At any rate, you get more overall stats for the combined gems so if weights are this close, I’d suspect a mastery/haste gem would still provide more DPS.
As for the boots, they aren’t bad per se, but keep in mind that simming a boot with an embellishment is always going to far outperform a regular boot. You’d have to do a side by side comparison with another setup that has two embellishments.
Last I checked, for a two DPS embellishment setup, adaptive bracers and a shadowflame on the chest were ideal (this is the worst statted tier piece). Part of the reason for this is that the damage from shadowflame patch will proc your bracers reliably. This may have changed… Again, you’d have to sim. I personally prefer adaptive plus sporecloak, it’s a game changer for survivability, especially since we’re doing mythic rashok prog.
I’m not at a computer that can easily do a sim setup, but I could play with it and see what it spits out when I am.
We’re on the same server and I can make any mail gear at 447 so if you find a better setup for yourself, feel free to send me a personal order, no need for a big commission.
My suspicion though is that there may be a gap in the rotation, since your gear as it stands wouldn’t make your DPS “bad” by any means. If you have a fight you were particularly frustrated with I could peep the log.