Just for your information:
Comparing two kills is in no way a proper means for assessing which build does more damage.
That is the reason why theorycrafting relies on simulation tools to simulate tenthousands of fights in order to evaluate builds.
Of course, there are subjective aspects like the
"right" feeling, but you really should not base your decision on the comparison between numbers of this week and last week.
Nevertheless, what was said in the posts above is definitely true and different builds may deliver different numbers for certain fight situations.
Keep also in mind that, for single target fights, the dps distribution of a mastery build and of a haste build overlap and share quite a large portion of possible dps numbers with the haste build simply having higher numbers on average.