Archeology rewards a very nice 2her at 3.8 swing speed. Zinrokh Destroyer of Worlds

This slow weapon combined with TV/seal dmg/CS is gonna be pretty insane. I don't see fast weapons out performing slow until high tiers with alot of mastery or proc based set bonuses.

Overall, it's probably better to have a relatively short duration on HoL, something in the 8-10 second range. That way we will have time enough to react to gaining the proc, assess the situation to determine what will be optimal and then use the proc, all while leaving a little wiggle room. With a long duration, or no duration, it becomes another gimmick as was discussed earlier regarding Holy Power with no cooldown. HoL would become another thing we'd need to save from trash or game from dueling and critters in preparation for a pull. In the instances of fears, phase changes and the like, we still have the option of refreshing Inc or tossing someone a WoG, so it's not like we have no options in that situation either, even if they aren't necessarily the best dps options possible.

Overall Ret has shaped up really nicely. About the only place we still need some work is in the "free choice" points of our talents. Currently there's only 1 optional point, but this point needs to be spent in the Ret tree. I suppose you could also consider Rebuke and Repentance as optional points, but I have a hard time doing so. Just dropping a couple points somewhere is about all that is needed. Preferably from the last 10 points so that we have the option to put them in one of the other two trees and not just the Ret tree. I'm looking at Blazing Light, since we already have Exo damage in a 3 point talent, though I imagine it might currently have some use to Holy for leveling.

May i ask why everyone talk about slow weapons and TV? Unless they severly changed philosophy, TV should be normalized like all other %wpndmg strikes except passive dmg like seals. Therefore there should be no difference in slow or fast weapon regarding TV as it will always calculate with 3.3 speed... Then the question should be if mastery outscales seals. In any case the gap should be quite marginal

This is a common misconception. Normalization doesn't remove the benefit of slower weapons, it simply prevents them from double dipping.

For example, it USED to be that ability damage was calculated (roughly, obviously multipliers were involved) with weapon damage x weapon speed. Weapon damage is higher on slower weapons because they swing slower, so a slower weapon would net you higher weapon damage AND weapons speed, hence "squaring" the benefit.

Now, abilities are calculated by weapon damage x fixed multiplier. This fixed multiplier, in the case of almost all 2H specials, is 3.3. But you STILL get the benefit of the slower weapon in the weapon damage component, it's just less than the benefit you would have seen before normalization. This is why old weapons, like the 3.8 speed Arcanite Reaperr, were for a long time vastly superior to 3.4 speed epic items that had more DPS.

Until abilities and talents are tuned (i.e. tweaking coefficients, accounting for the new Mastery, etc.): it's still far too early to determine whether or not 2HP or 3HP TV's are superior; faster vs slower weapons; etc. At present, it's still even pretty difficult to determine priority of abilities and fillers, and their situational uses. There needs to be a little more revision and an acknowledgment from GC or the like that we're near a final pass. We can build numbers and run sims off what's on the PTR, but it is speculation, at best, at this point in time.

EDIT: Found the post in which GC stated damage numbers aren't final, and the 2HP paradigm is against what they're shooting for:

All numbers aside for the moment -- I do feel the "rotation" is nearing the end of its development stage, even thought there haven't been blue posts specifically clarifying this. Their intent is to make TV the centerpiece of the rotation, and this is exactly what we're seeing right now. Unless there's a major overhaul, I don't see much happening with what we're already seeing outside of a few percentage bumps or damage tweaks on fillers. The latter is something I see as somewhat feasible for PvP reasons, but that's beside the fact.

I don't think it's entirely out of the question to start theorizing what would yield the highest DPS. The 2HP rotation(just like the 3HP rotation) isn't affected more or less than the other due to mastery, because if mastery procs you're going to hit it no matter what the case(unless you're running or whatever). The argument is whether or not loss from Exorcism damage in the 2HP rotation, due to lesser chances to use the proc, would outweigh the increased use of Templar's Verdict and consequential seal/SoC damage.

I agree with your first paragraph. I'm inclined to believe that our rotation is near to being what is the desired intent. However, I disagree in that I do not believe Blizzard would allow/want Ret constantly using a 2HP TV over the 3HP. As stated: TV will be the centerpiece of the rotation. If the current design allows for some "flaw" in that we never use it at it's maximum power, I'm skeptical Blizzard would let that stand -- given it's against their stated intent.

There is a problem with theorizing DPS yield, especially on the PTR, in that we cannot train Inquisition -- which acts as a damage coefficient. This trouble is compounded by the fact that we're not also not tuned. (coefficients, Mastery, et al.) There's also the nasty little trouble in that we have to sack a TV every 30 seconds to keep that buff refreshed: a fact that seemed to be skipped by the RetPaladin article you cited. If Inquisition makes or breaks Exorcism's place in our rotation, that's something we cannot account for based on PTR numbers. Additionally I stand by the fact that Seal damage is pretty paltry in it's current incarnation, and will likely be receiving some attention. These coefficients and missing buffs from our Beta counterparts can amount to very large sums of damage.

So in some regards it is a bit too early to determine the outcome of your cited theory. There's also the question of the data collection in the post. Was this person using a target dummy? Latency? How many swings did s/he do? Sample size? I can only see selected screenshots of Recount, which leaves many question about the tests, the parameters, and the setting. And the author's conclusion is that it "feels nicer," not that it does more damage; followed by more hypothesizing.

This is not to state that the theory is invalid: it may very well be. But it seems that there are still too many variables that could change between now and 4.0.1's Live launch that have yet to be accounted for, tweaked, fixed, and/or answered before this line of thought can be properly evaluated.

We will probably have an answer in the very near future.

Could the mana return have actually been 2.5% second, but the display rounded it up to 3%?

After reading that post about the 2HP rotation I went and tried it myself thinking there was no way it could out-dps the 3HP rotation. A 2HP TV will do a little less than half the physical damage of a 3HP TV. You do get an extra seal proc, but that's nothing compared to the damage you'll be giving up if you have to replace another filler to fit it in.

Much to my surprise, the 2HP rotation did indeed out-dps the 3HP rotation in my initial test (by a little over 100dps on 10 minute dummy sessions). This was well within the error range for dps numbers, so its certainly not conclusive, but even just being on par with the 3HP rotation made me really scratch my head.

I started trying to compare my numbers to that other set of tests and was re-reading the post when a lightbulb went off. The post talks about delaying abilities "in an attempt to reduce the number of times I wasted Holy Power (because the amount of RNG)". Initially I thought this had to do with having to watch several different parts of the screen to keep track of the current state of our procs if you aren't using clcret, but then it occured to me that what might really be going on is that the delay in DP-generated holy power showing up in the UI could be causing extra abilities to be used and the HP they generate to be wasted.

DP procs don't register instantly. There seems to be a 1-2 second delay between when you press the button and when the server tells you that you've gotten a proc (probably somewhat dependent on latency). If you're at 2HP and use a DP ability it's not uncommon for the proc to register close enough to the end of the GCD (or even after) that you've already hit the next button, wasting whatever HP that button might generate. In the 2HP rotation, however, you always have that built-in safety margin so that if you get a late proc your 2HP TV turns into a 3HP TV.

It also occurred to me that if this is indeed what's making the 2HP rotation better than expected, then it should be possible to improve further with a "smart" rotation. Ideally we want to not waste any HP procs, but also have as many full-power TVs as possible so, instead of just hitting TV at 2HP we can use one additional ability at 2HP and then hit TV regardless. If that ability happens to be a CS or a filler that procs HP we'll still end up with a full-power TV. If that ability is a filler that doesn't proc TV then we'll get a smaller TV, but at least we know we won't be wasting any HP.

I modified clcret to support this new type of rotation (when an ability is used it checks to see if I'm at 2 HP and if the ability that was used was a DP ability and sets a "might be 3 HP" flag if so; then TV is used if I'm at 3 HP or if the flag is set). I then ran 3 more tests of exactly 10 minutes each and compared the results. The 2HP and 3HP rotations were almost exactly the same dps (2HP won by about 10 dps this time) but the "smart" rotation was almost 150dps ahead of the others. This difference could still be due to RNG, but I don't think I ever saw either of the other rotations get within even 100dps of this rotation, so it may also end up being significant.

Number of TVs in each test and their average hit/crit damage:
3HP: 73 @ 6256 (hit) 13371 (crit)
2HP: 106 @ 4260 (hit) 8237 (crit)
"smart": 102 @ 5298 (hit) 11430 (crit)

It's hard to compare number of TVs directly due to not knowing how many were from the mastery. The tests had 207, 208, and 209 melee attacks respectively, so the chances should have been about the same, but RNG is RNG. Still, its certainly interesting to see the "smart" rotation with significantly more TVs than the 3HP one and about 4% fewer than that 2HP one with damage a little over the average of the other two rotations.

The "smart" rotation also had the pleasant side effect of changing the "next ability" button far less than under the other rotations. Basically it would only change when an FoL (and occasionally AoW) proc happened, and when these proc near the end of a GCD there isn't a huge penalty if you hit something else and get them on the next GCD. The other rotations have the annoying property that TV often pops up mid-to-late cooldown when the DP proc is finally registered and not making the quick adjustment can waste HP.

Any conclusions about rotation or DPS made on the regular PTR are pointless.

You don't have all the abilities (like Inq), you don't have mastery stats, you have MUCH higher haste and crit in your 80ICC gear at level 80 than you will have at 85 with 85 gear. ArPen on 80 epics got converted to haste or crit, so you have 'too much' of it in any case. The gear stats also change sharply, and cata weapons get relatively much more oomph than the wotlk versions. In a way, you're dps testing (even at 80) with too low relative weapon damage. With TV being weapon damage based, this means you're now putting too much of your dps in holy damage and base ability damage rather than the weapon damage component. THis can very much skew the numbers towards a 2HP rotation being favourable.

Even on the beta, many conclusions are pointless. I do more dps as holy spec and prot spec (and my ret gear is even 'better' than my holy/prot gear since I mainly took the ret quest rewards). I even do more dps as ret spec in my tanking gear, using a 1h+shield. It shouldn't be this way, and napkin math says it can't be this way, but there seem to be a number of bugs atm that make much of the theorycrafting moot.
- Seals are doing less damage than they're supposed to according to the tooltip. My best guess atm is that something got mixed up changing Exorcism and they now only benefit from AP and no longer from SP. Or some other magic damage suppression effect is happening.
- Melee white is hitting lower than it should. Either mobs have unusually high armor, or some armor calculation is off whack. Seeing how low damage some casters are taking,and how much damage they do I'm having this suspicion that armor calculations are reversed, so we're attacking a mob and it gets our armor value while we get theirs. It would explain why caster mobs tend to hit harder than melee mobs, and why cloth players seem to be taking the same kind of damage from being melee hit as plate users.

Didn't check if there's addon already for modifying overlays but the default HoL overlay can be barely seen sometimes.
If you rely on it and want to change the way it looks, put this code in some lua file that is loaded (just remember where so you can edit it later if needed)
Can change the values as you please, my code switches AoW with HoL.