During your cooldowns, and especially with Bloodlust, you should always be prioritizing TV. So it would Inq>CS>TV> and then I'm not sure if it's HoW or Exo. It might be dependant on your gear, and you should probably sim it.

There's no need to ever push Exo unless your next white swing will occur in less than one GCD. Even then it would still have lower priority than CS (<3HP), TV and Inquisition refreshes at less than 3 seconds remaining.

A question - if I get back-to-back exo procs (use exo and get another proc), how far down the priority list does it go due to glyph clipping? E.g. do I have to consider using judgement instead and saving the proc for the next filler (since the chance that I'd get yet another proc within 6 seconds are rather low), or its negligible?

Playing with 4.3 simcraft, I see a very small benefit from prioritizing Exo over HoW (~100 dps), which is surprising given HoW nerf. Glyph of judgement is still showing a ~200 dps loss compared to glyph of exorcism, but that's using T12 gear (and same priorities). I'll try to add code for T13 and make a profile with recently added raid gear.

Exorcism's priority doesn't change when you get consecutive procs. The only case that really matters is getting a proc, using CS then Exo, and getting another proc before CS comes off cooldown. Even in this scenario, we'd still cast Exo again unless we got a DP proc. The lost damage is negligible since we're always guaranteed to get at least one tick and the vast majority of the time we'll get the full six seconds' worth.

The only time Exorcism's glyph becomes an issue is when chain-casting it, and only a "DPS" specced Holy has the resources to do that. Needless to say, something's wrong if a Holy paladin is trying to DPS.

The gear does look like its full of placeholders. Names and icons are obvious, trinkets most likely will be completely changed too. I figured its more fair to simulate the current bis gear with T13 set bonuses (which look fairly solid and may make it to live) so the numbers can actually be compared to something.
Ok - the simc version that is available for download doesn't have the JoT buff, even though its apparently included in the simulationcraft.org page. T13 is already implemented and can be activated with a couple lines in the profile. It seems to increase dps by another 1.5k compared to T12.

Here's the reason I don't like doing this, and it's at the core of how our secondaries work. As one increases, the relative value of both the others increase compared to the increasing stat. For example, as you stack mastery, haste and crit become more and more desirable at the expense of mastery. It's not always linear, but it's how our secondaries interact.

When I model things, I try to see how our native stats look on our gear, primary and secondary, and see where reforging nets us the highest gain. If we have a lot of native haste on gear, for example, and we can't ever reach the current 1.22 to 1 Mastery to haste/crit ratio, we need to look towards other options in secondary priority. So far I've done most of my expectational modeling based of all stats being relatively equal, then adjusting to see if anything's trending for a fundamental change. So far, nothing has, beyond what I've posted already. What a lot of readers don't understand from the OP is how our secondaries truly interact and change as we gain and lose secondary points. The way haste and mastery interact is very interesting, and it's extremely hard to predict what is worth reaching for, specifically the mythical 3 sec CS, and what we are going to have baseline regardless of gearing/reforging. I'm not sure how Exemplar feels specifically about the matter, but that's what I'm doing right now, going with the wait-and-see.

A bit of nitpicking, but that doesn't apply only to secondaries. As you get more strength, the relative value of other stats changes as well, because not all abilities scale the same with AP.

The fact that people assume that "oh, I read on EJ that mastery is better than crit, so it must be true" is something that's always been a problem and no amount of warnings in the OP will fix that, because things like "it depends on your current gear" will still just make people pick on stat then claim "my playstyle suits my choice". It's even harder to get people to trust some 'rule of thumb' or to get them to use a tool, when two of our tools provide different results.

Going off on a bit of a tangent (and knowing that Exemplar and simulationcraft's ret developer agreed to disagree), the complexities of interactions between stats, if anything, make me believe that a simulation will have a higher chance of producing accurate results (no offence to Exemplar's work). A formulaic 'averaged' solution to our rotation is just less likely to produce better (closer to 'reality) results than a simulation. Of course, a simulation will never be able to (quickly) provide you with the optimal way to reforge/gear, because you have to simulate every possibility, but you can infer which would be the a good reforge strategy for a given 'range' of stats, because barring haste caps, stats' values are continuous functions (at least for retribution paladins).

Something I hadn't noticed before, but now that I think about it since I was going to do some testing with the guild today:

Upon looking over the mechanics of Ultraxion, one of the encounters in Dragon Soul, you can see he's got a passive mechanic which removes his ability to dodge/parry.

Would it not be beneficial to use a different prime glyph other than GoSoT for this fight? And for that matter, to keep a set of gear specifically to disregard expertise for the sake of more raw stats?

Maybe I'm thinking too hard, but it doesn't seem like this is just a temporary fight mechanic unique to the PTR or anything. We'll very likely have to consider these as possibilities for at least this upcoming fight, it seems.

If you happen to carry dust on you, it would be worthwhile to switch out GoSoT for practically any DPS glyph on that boss, assuming nothing changes. A separate set of gear though? Undoubtedly, you could run a 0 (or as close to 0 as you can get) expertise set and do more damage. Would it be worthwhile? Depends on how difficult the boss is, and if whether or not your added DPS from that makes a difference. I suspect that the next raid will be a carbon copy of T12, and only the last guy is particularly hard.

Basically, dropping all your expertise will undoubtedly get you a DPS increase (Assuming he can't dodge/parry at all), but it's up to you as to whether or not it's an economic choice to make. For me, personally, unless there is a tight enrage timer and it's a messy DPS race, I wouldn't find it worthwhile to go through that much trouble.

Lord Rhyolith does not dodge or parry for 90%+ of the fight (phase 1). Are you presently swapping out glyph and gear? Yes? Great, feel free to do so next raid. No? Why start now?

Swapping Glyph is simple and would provide a bit of DPS, but not a huge swing (at best Glyphs are normally in the 200-300 DPS range, which is vaguely 1% total DPS). Since Glyphs are already situationally useful, I'd even encourage this (Firelands has good opportunities to swap between Divine Protection and Holy Wrath).

No-Expertise gearing tends to be counterproductive. Early on in a tier it winds up being a vast amount of previous tier's raid gear, rather than the upgrades that are dropping. Later on when you've farmed enough to have 2 sets of the present tier's gear you have to ask if it's worth a second set of gear for a farm boss you're demolishing weekly.

So, after a quick glance over Ultraxion parses from World Top 250s, there's still a significant difference (10k~) between Rets are parsing and where the top of the DPS are parsing. This is true even excluding feral druids.

My own parse is absolute dogshit, but even players (of the parses that I saw) seem to be well behind. This is really frustrating and irritating. If someone is seeing extremely different results, can you post please?