As far as i can tell, there is only two scenarios where the additional 200~ (after reforging) expertise/hit cannot be turned into haste or crit.
The first one is the case where you have an expertise/mastery item, where you currently have reforged mastery. If you need to reforge the expertise into crit haste, you can no longer reforge the mastery. In this case you will gain mastery. Since this is a measure of last resort however, it should be uncommon.
The 2nd case is the case where you are already reforging expertise into other stats. In this case you might end up overcapping expertise, and thus waste some of the stat gain from the trinket.

There is plenty of situations where it will only be possible to gain haste, not crit however, f.ex. Crit/expertise items, where we can reforge expertise into haste. Since we would prefer these items over haste/expertise, having an item like that is not unlikely. Another possible scenario is haste/crit or crit/mastery items reforged into expertise.

Both cases are more likely to happen in the case of expertise, since the hit cap is higher than than the expertise cap.

In general however calculating all stat gains in a range between haste/crit could be a good approximation.

I would only put TV ahead of exorcism when you have full holypower and then you get art of war proc, Thats only becouse you may waste one DP proc.

Otherwise exorcism goes allways before TV and CS.

I've just studied some logs regarding Art of War, both my own and those posted by HamSlammer.

My own logs show AoW between 14% and 24% on a per-fight basis and an average on bosses of 19%. Looking at Ham's logs I got 20% for him on bosses in one log and 21% in the other, 22% for Ithanii in Bastion, and 22% for Khaos in BWD. I figured this via AoW procs divided by total autoattacks on bosses (including blocks, glances, and absorbs).

Based on their haste and JotP they should be getting around 25% faster autoattack speed (not even counting Windfury effects). So if haste was impacting (beyond more eligible swings) we'd expect 20% AoW times 125%, or 25% AoW proc chance. We're seeing far closer to 20% than 25%, well within RNG fluctuation.

I really think the times where it is noticeably above 20% are just RNG. I'm going to table it unless/until someone uses an extreme haste set (reforge/gem solely for haste) on a dummy for a nice long test run and comes up with a significantly higher than expected value. Bosses/raids are too short for a log to show a statistically meaningful highpoint (not that any have done so even with small sample sets).

Haste does not make our rotation any more consistent. Even pushing CS down to lower values we still perform 2 filler attacks, and the "inconsistency" of our rotation is entirely regarding those filler attacks (especially AoW or HoL procs).

Yes, crit is at the whim of the RNG, but under that aegis it could be yet more beneficial than average (lucky RNG streak), as well as below average (bad RNG streak). Crit may also gain some comparative value due to time off target. Any movement greater than your swingtimer somewhat obsolesces haste (devaluing it compared to crit), whereas the first attack on return could still be a crit. I've seen this very argument against haste (time off target) in a post discussing PvP stat values.

That said, haste gains ground at the moment for one important reason - PvP gloves. As we've determined the PvP gloves beat out non-epic gloves, many of us may be using them for a bit (buying other Tier items first are likely larger individual DPS upgrades). Since the PvP gloves increase CS damage, haste gains a minor comparative boost since it reduces the CSCD.

Either way the difference between crit and haste is rather small. I am personally striving for a balanced approach, rather than stacking one exclusively over the other. With more haste there are more things to crit. They may not be soulmates, but they play well together.

Possible Interesting Changes posted by Ghostcrawler @ Upcoming Class Changes - World of Warcraft

would this make getting a 3s CS even more valuable or will it devalue haste?

Unless the new hand of light proc has a pretty high chance, I believe we'll need another source of Holy Power, cause a ~3.8s CD CS won't be enough I think and we might see a huge drop on DPS due to a lot less TV used.

I highly doubt that they would be making these changes if they were going to drop our dps lower than what it is currently. I assumed that Divine Purpose would remain a 40% chance for Hand of Light, however that seems kind of high. If you think about how often HoL procs now, multiply that times 5, and thats how often we would be getting HoL procs. I agree that we still need another source of guaranteed HP. The mastery change is good for more consistent damage, but until we see how Divine Purpose will function, I will refrain from making any judgements.

It is true, but reading the quote it seems like they are removing some of the randomness out of Divine Purpose, and this can only happen with a fairly high Hand of Light proc chance. The way that was phrased, I reckon we'll probably end up with more HP then we can handle at a time during HoL procs, and possibly less than we need at normal times. Anyway, it could still be better than what we have now.

I don't know how much would this make 3CS more valuable, but I do think it shouldn't devalue haste at all, especially if the new Hand of Light still procs out of autoattacks.

40% chance of gaining 3 HP would be insane, yes, but I don't think something along the lines of 20-30% is too far off.

The mastery change adding holy damage is what I have been wishing for since days now. Although I would have preferred something along the lines of good ol'SoC (applied to TV, possibly), it sounds like they are adding a fixed percentage of skill damage as Holy damage. This could be of major help with DPS or not, it depends on how high is that percentage. Still, it should be an improvement.

To be honest, it sounds like the "new Hand of Light" (which should replace DP) could be still proccing from autoattacks like HoL did; if this was the case, a quite high proc chance would be probably needed.

About the damage, we don't know yet how much of a percentage of the original skill damage they want to add as Holy damage, so I'm not sure how you are calculating that.

And Goatk needs a buff, yes: it should attack properly all targets that the Paladin is currently attacking, plus the Strength buff really should be based on total Strength instead of base only. Those are quite needed fixes.

We cannot effectively theorycraft the Mastery change until we know more information. The very little that has been previewed suggests HoL will cause TV, CS, and/or DS to deal part of its damage as Holy (bypassing Armour and also being boosted by Inq).

Unknown important details:

• This may or may not still require a GCD to invoke.
• The quantity/percent of damage converted to Holy is unknown.
• Stacking is unknown - do 2 HoL procs before a TV is used make twice as much Holy, or is one lost?
• What happens with HP generation if DivPurp is HoL - with zero haste (impossible) it would take 9 seconds to build 3 HP via CS alone, meaning TV at best 10 seconds apart, which would require massive re-balancing not to nerf the spec (they have not espoused a need to nerf us).

HP generation obviously will be rebalanced if DivPurp changes. Otherwise I can see a severe flaw if HoL causes the next TV/CS/DS to do x% Holy damage. Notedly x% of CS will always be smaller than x% of a 3HP TV, so if quantity of TV drop due to less HP generation we would have more CS eating the HoL bonus. Thus the new Mastery would produce less gain than first glance may suspect.

Blizzard has recognized the flaws in Ret Mastery (huzzah!) and intends to correct the issue. It will definitely require some changes to spreadsheets once we know all the details. It's a late Christmas present we can eagerly anticipate without even knowing what day Santa will come down the chimney.

Oddly, I expect this change to push Str even further ahead of other stats, while hopefully normalizing Mastery to Crit/Haste valuation. If part of attacks become Holy, then there is less mitigation from armour, thus Str causes us to hit that tiny bit harder. However, the change may be tiny.

On an entirely unrelated note:
Yesterday I realized something about gearing. With our stats closely proximate and the ease of reforging for Hit/Exp caps, which precise stats an item has are not of utmost importance (more so, if the Mastery fix aligns with Crit/Haste rather than being far above or still below). What is most important are quantity of stats.

Example:
[Geordan's Cloak] vs [Shadow of the Past]
Str and Sta are matching, naturally, based on ilevel. So we go to secondary stats.
Geordan's has 98 Hit + 120 Crit = 218 ratings
Shadow has 112 Crit + 112 Haste = 224 ratings

Thus, under the supposition we can still Hit and Exp cap via Reforging, Shadow becomes superior to Geordan's based on the 6 extra rating points. This is an interesting artifact of the Reforging system combined with diminishing returns. The higher an individual rating the more "item points" that rating costs. Items with both ratings in exact balance should always surpass items where one rating is higher, as the other rating would be more than commensurately lower. Again, the only caveat being Mastery at the moment (and the great unknown as to its future value).

I believe you're misreading GC's description of the change. I dont think the %holy damage added to CS/TV is linked to HoL procs at all, instead the ret mastery is now a permanent holy damage addition to CS/TV and HoL is no longer linked to mastery. Then since they like the HoL proc they're replacing the Divine Purpose extra holy power effect with it proccing HoL.

This is the interpretation most are coming up with. Our new mastery effect will be a constant, rather than a proc. As described it should operate the same way as the Wrath (pre 4.x) talent Unholy DK's had which added a percentage of shadow damage to certain strikes, but in our case adding a percentage of the damage done by certain abilities as Holy damage. That is, CS damage would become (1.15WD) + (1.15WD * M) where M is our base mastery plus any additional mastery from gear, enchants or procs, and so forth for Divine Storm and Templar's Verdict.