If AoW procs every 15+s on average, wouldn't you use Exo at 21s (18s after the 1st Exo) instead of Cons? It would end with 2 Exo 2 HoW vs 1 Exo 3 HoW. Thus higher damage by prioritizing Exo?
Excellent question. If you could guarantee the average, then the answer would be 'yes'. Unfortunately you cannot guarantee the average.
0 - AoW procs. CS and Judge are off CD, so you are going to use both.
3 - AoW refreshes. Never had a chance to Exo, so this proc is lost.
30 - AoW procs.
You had an average proc every 15 seconds, huzzah! Oh... wait... it still didn't help. It's a gamble that it occurs. If you win the gamble, great. If you lose the gamble, too bad.
Spoiler to hide yet another table, since post is already long.
You have the possibility that you may do more damage with Exo>HoW as well as DP>Judge - but in the long term and on average you wind up with more damage the other direction.
Again - I do not blanket say HoW>Exo is for everyone. I continue to hold both in the OP as some modeling tools for some people show one, some show the other. I have no firm idea what may be the break point one direction or the other (see some more theorizing in my last post).
This is very much an end-of-expansion discussion, since comparing HoW>Exo and Exo>HoW (with my toon) found a variance of 0.27% (my spreadsheet, HoW wins) to 0.47% (SimC, Exo wins) total DPS change. The decimal is not misplaced - that is one quarter to one half of one percent. I actually get a larger variance in Simcraft (0.56%) between SimC's default setup and what I consider a more realistic setup (minor tweaks like making sure Inq is running with sufficient time before engaging Zealotry and delaying GoAK for the next Zealotry rather than possibly using in no-mans'-land - although based on those results maybe I can get half a percent more DPS by NOT watching timing as close as I do, this makes me sad). Heck, you even get a fluctuation of around 0.2% between individual runs of the same settings (this is not a knock against SimC - its a Sim, so it actually uses an RNG, it's doing precisely what it's supposed to do and doing it damn well - the above numbers used multiple runs of SimC with each configuration to more accurately compare averages).
Bottom line, since I like bottom lines - we're talking about fractions of a single percent of DPS between our various choices. As long as you spend your HP on TV and keep Inq running, it appears you could get all your choices 'wrong' and still do 90% of maximum DPS**. Consider that the RNG impacts Landslide, Gur, Trinkets, AoW, DP, Crit, Glance, and doubtless other things not immediately coming to mind. A good player having an unlucky day could lose to a mediocre player having a lucky day.
* - Ran DP > Judge in SimC and got results that overlapped the Judge > DP range. The low end of DP>J did appear lower than otherwise, so on average there does seem to be the loss, but it's something like 0.08% - less than 1/10th of 1 percent. This is why I didn't bother to create the distinction in my spreadsheet.
** - My spreadsheet showed HoW>Exo>Judge>CS was 90% of max DPS. SimC showed this as... wait for it... 95% of max DPS. Dropping Judge and CS behind HW and Cons still puts SimC at 91% of max DPS. As long as you can keep Inq up, use 3 HP on TVs, and spam an available button (you'll still suck if you don't press buttons or spam something on CD while something else is available) you're looking at 90% max DPS.