I'm sure its useful, but weigh it against other classes.

Lets assume you have a 3.2 second swing timer (just pulling a random number assuming you have some haste gear). In the 20 seconds that buff is up you will get in 6 swings. Assuming a 20% proc on WF and we're looking at about 7 swings + 3 Crusader Strikes, or our 10 swings for maximum power. Assuming you get the first stack from your first autoswing and perfect results from there, you'll be looking at 16 seconds to get to the 440 attack power maximum. We get 4 seconds with 440 Attack Power.

Compare this to a Sword Rogue with a 1 second off-hand and a 1.5 second main hand (again, just spitballing numbers, but I'm assuming they have SnD and some haste). They will get a full 10 stacks in 6 seconds from only autoattacks, not even counting Sinister Strikes, Evis, etc. They will get the full 440 attack power for 14 seconds as opposed to our 4 seconds.

It is possible we will get procs from things like JoC (I have noticed it does proc "on swing" abilities like Executioner), but even so it is clear that Dual Wielding classes will be getting a ton more out of it than we will. I'm sure its a pretty nice trinket, but I doubt anyone will be in a position to wear one for a long while.

Haste is not better then Strength for me, although it is nearly as good, and much better then crit.

I think there is a bug with Rawr. The LotP buff does not seem to be added and does not rise the dps.
Has this bug already been reported ?

I'm not planning to go for one as I'm pretty happy with my current combo.

However, I think it's less important to compare how it affects different classes and more important to compare the new trinket to our current trinkets.

Is it better, or not? And if it is, by how much. That's all that should matter and leave each to reach their own decision based on that



I'd say the main problem is that we simply don't know what procs it. If it's just white and CS, yea it's probably not that great.

If it's White (+WF), CS, SoC and JoC, well then you can stack up 10 hits pretty damn fast.

In 9 seconds (assuming hasted 3.0 speed using drums and what not), you can have 3 white, 3 SoC, 1 CS, 3 WF, 1 JoC. That's 11 attacks in 9 seconds under optimal conditions. More realistically you'd have it all stacked in ~12 seconds (4 White, ~1 WF, ~2 SoC, 2 CS, 1 JoC) assuming 3.0 speed.


Just throwing numbers around, yea it doesn't look that good, but I'm still hoping for some more detailed modeling for it to be done (Rawr maybe?) to know for sure.

Thanks for the numbers. Your calculations look correct.

Just to better illustrate how the numbers are "off"
Normal Swings = 822
SoC proc = 360
WF procs = 214
raw WF proc rate = 26% (instead of 20%)

Normal + SoC = 1182
adjusted WF proc rate = 18.1%



I have a theory that WF procs off SoC, but only if there is no previous WF proc. (ie: cap on chain procs)

If that is the case, then for the 360 white hits that proc'd SOC, 20% (72) proc'd WF as well. That would leave only 80% (288) of the SoC procs "eligible" to proc WF.

Normal + SoC = 822 + 80% * 360 = 1110
adjusted WF proc rate = 19.3%


This data set is consistent with that theory, but we'd need some more data and better analysis to really prove it. Do we have any SoB trash/instance parses that we could use to compare WF proc rates with?

I posted a few WWS with SoB a page back. on boss fights WF is up ~100% of the time but on trash my shamman likes to screw with me and drop flametoung or nothing to try and beat me on the DPS charts so it's not always up on trash.

Edit: Post #3975 page 159

Its like Warglaives. They are technically best in slot for DPS Warriors, but most guilds will give them to Rogues first because they will get more out of it. No matter how good the trinket is, I'm nearly positive DW classes will get a ton more out of it than us, so chances are ret pallys won't be seeing a large amount of them.

That being said, I'm all for figuring out the actual value for the trinket in comparison to things like DST. It would be nice to at least have an option of not having to keep farming Gruul eventually.

After being caught fawning over the advantages of Seal of Blood, apparently Eyonix's official response to the significant disparities between the factions of ret paladins is...

AKA: "sucks to be you".

Seal of Command procs Windfury

Not only does seal of command proc windfury, but there is no internal cooldown from the windfury totem buff. While I was testing in blasted lands with a large club, I noticed my SCTD addon showed a double proc of windfury. I found it in the log file. The quote below is a copy of the non-verbose log listed on WWS.

Pardon my WWS showing me as a mob. These were my first uploads and I don't know why it did that.
WWS - No SoC
WWS - With SoC

My test consisted of using a large club for 20 minutes, once without seal of command and once with seal of command.

Without Seal of Command
449 hits
68 Windfury
17.8% proc rate -1.05 Standard deviation off

With Seal of Command
490 hits
104 windfury
164 SoC
27.0% proc rate 3.44 Standard deviation off
18.9% proc rate (SoC included) -0.62 Standard deviation off


These numbers show that I was a bit unlucky with proc rates, but it is well within the realm of possibility, with SoC included in the calculations. -1.05 off SD is about a 30% chance. 3.44 off is about 0.15% chance. However, I believe my combat log is the smoking gun, the proof that Seal of Command procs windfury.

What does this mean for us?
Alliance paladins are doing more damage than our models show us doing.
Any model which accounts for windfury needs to include a chance to proc on Seal of Command.
Unproven assumption - The disparity between strength and haste itemization increases more for Alliance paladins.

Overall, we're doing more damage than we think we're doing, and the haste itemization in Sunwell is even more "less good" for Alliance than Horde.

I think comparing Warglaives to a trinket is going a bit overboard :P And yea our Warglaives go to rogues.

What it boils down to for me is: Lets not have another DST.

At the start of TBC, DST came out and I'm pretty sure most paladins (at least alliance paladins) passed for it, assuming "it's not so good", "it's better for other classes", "haste isn't that good for SoC" yada yada, the same things you're saying right now.

I myself didn't grab a DST until like a month ago, that's over a year after it first dropped for us and we've had plenty.

Turns out, not only is it a good trinket. It's currently the best in slot with Shard of Contempt using Fel Mana Potions. And by a fairly large margin over most trinkets too.

I'm all for giving it to a DW class if we know for sure (and it does at first glance not look that good for us, I agree), but I'd rather not have history repeat itself with this new trinket, so I'm not going to form an opinion until it's known exactly what is what.

The argument of how much it boosts what class by how much is really irrelevant. Save things like Warglaives, if you go by that argument, classes like fury warriors would be naked until rogues are fully geared, classes like shadowpriests would still be wearing SSC gear until all warlocks have full T6.

So yea, find out first, then decide. Think we should keep the whole "loot for this class or that class" out of the discussion though.



Very interesting.

Can we let the latest RAWR take this into account (WF proccing from SoC)?

I'm not so sure about your last statement however (haste being better for alliance than horde due to increase in WF procs), plug it into RAWR and lets see what comes out.

Running this through Bellator's spreadsheet by changing how WF procs are calculated, given my current gear, but less raid buffs than usual (some things are missing which Rawr has now):

WF only off white swings: 1860 DPS
WF off white + SoC (which procced off white): 1874 DPS

So yea it is an increase, but unfortunately not as large of an increase as expected.



Trying to replicate what the spreadsheet spat out through theorycraft:

At the theoretical 7ppm SoC, and 20% chance to proc WF off those, you're looking at:

1.4 extra WF procs per minute

Nice, however the bonus is much less than you assumed I'm afraid.



Note: I'm getting some discrepancies for SoB calculation between Rawr and the Spreadsheet. While SoC calculation seems to be within ~60DPS given the same gear and buffs, there seems to be a huge ~90+ DPS discrepancy for SoB (Rawr values it much higher).

Zurm can you go over how Rawr calculates SoB damage and compare to the Spreadsheet?

I will make that change to Rawr, Avitus.

In addition, I'm going to need a consensus as to how exactly how SoC works with WF now. Pretend I'm retarded, and spell it out in such a way that no one could possibly confuse what you're saying with something else.

The consensus that needs to be established is the following:

1. Can SoC proc off WF?

Currently used assumption so far is YES, for both the current spreadsheet and Rawr.

Empirical evidence points towards this being true, with all the ~8.5ppm people have been reporting over the last 10 pages.

2. Can WF proc off SoC?

This is what's being discussed atm. Most recent results established by Fiola (+her secret gang of math monkeys) and Paragos point towards YES. More test results would further solidify this until we can call it conclusive.

(Note: Do read my reply about this before jubilating however, it's not as big a buff as expected).

3. Can there be chain procs?

Currently Unknown, mostly assumed to be not the case, however Paragos seems to have found a smoking gun in a recent WWS log (line 20:46'31.142 to 20:46'32.176) that shows it "can" double proc.

Why we haven't seen any WF proc off SoC proc off WF proc off SoC, I don't know, probably because it's extremely unlikely, maybe cause it's hard to tell what attack a proc belongs to during combat or maybe because there's some mechanic in place (which is what we're trying to find out here).


Anyone who has some input or would like to do some further tests or number crunching, please add your thoughts (if you have anything conclusive or at least with high indication towards something).


Results of this will be used for the next Rawr release (and possibly a small update to the spreadsheet to keep it current), so we can use any input we can get to make it accurate.

I don't think so.

The whole SoC + WF analysis shows that the % chance for SoC to proc off of a melee swing is only near the expected value if you remove the WF procs from the equation

For example, using the trash WWS paragos posted a few pages back:

Normal Swings = 822
SoC proc = 360
WF procs = 214


If we include WF procs, then SoC's proc chance is 360 / (822 + 214) = 34.7%
If we don't include WF, then SoC's proc chance is 360 / 822 = 43.8%

Going from memory (he has a large club on Armory atm), he used a Cat's Edge, which has a 3.5 AS.

3.5 * 7/60 = 40.8% << expected proc chance.


Using his other sample with a large club:
490 hits
104 windfury
164 SoC

w/ WF: SoC proc chance = 33.5%
w/o WF: SoC proc chance = 42.5%

expected proc chance: 3.2 * 7/60 = 37.3%


On the other hand, not including WF procs does not give us the expected proc chance, either.. so this could use some more investigation. I should apply the same sort of stat analysis here as I did for WF, but I don't feel like doing that right now. I'll try that later, and use some of the other WWS parses.

I don't think WWS from trash is a good measure. During trash you're not constantly hitting a mob, you're might lose SoC uptime, WF uptime etc.


The only results that should be considered are straight out Blasted Lands tests or boss results.

The issue I have with this is: There's been a handful of WWS logs in the last few pages (main brutallus where everyone is going all out) all showing ~8.5ppm for SoC.

What's causing this increase?

BTW, SoC proccing off WF is what Rawr and the spreadsheet currently use, so if it is wrong, we definitely need to change some things.


At some times I wish if we could have a definite response from a Blizzard dev about this. Certainly would solve all the dual proc uncertainty.

Well, the provider of the WWS personally guaranteed that he had 95% WF uptime for the trash WWS. We're using the same set of numbers to prove that WF procs off SoC, so I'm inclined to believe that the SoC uptime was sufficient to give us insight on whether SoC procs off WF.

(Some of the higher SoC DPS WWS for Brut, which should have 95+% SoC/WF uptimes, also show the same pattern)


All that said, it looks like the actual proc chance is more than 2 SDs from either model (WF procs SoC, WF does not proc SoC). I'll post some more math later, but this is gonna need a lot more analysis. = P

No. As has been said, it needs to be looked at through log analysis. Posting these incomplete combat log screenshots only serves to clutter up the thread.

You know, that's probably windfury proccing windfury. Remember, they accidentally reintroduced extra attacks proccing extra attacks in 2.3.

Warriors have been chaining sword spec again since then, too.


Edit: actually, I'm certain that screenshot is just windfury proccing windfury.

EditEdit: recheck your fancy pants mathing under the assumption that SoC can only proc once per combat round (IE has a very small internal cooldown) but windfury can chain proc.

Answers that question. Thank you.

I'm sorry if there was some confusion. I meant haste is still better for horde, but more so, due to strength giving that 1.4 extra windufry procs per minute an extra punch. Anecdotally, I took my average swing damage from my Sunwell parse (total swing damage/(hits+crits)and got 1.4 extra white attacks per minute being a 44.4 dps increase.

1. Can SoC proc off WF?

I've assumed yes, so long as it didn't proc off normal swing.

My own calculations put my proc rate on the Sunwell Trash parse at 37.9%. What is different in my calc over Fiola's is I took out the windfury attacks which were assumed to have already procced SoC from normal attacks.

360/(822 + (214(.592)) = 37.9% which is closer to the expect 40.8% from Cat's Edge, but still a bit low.

The same calculation on my 20 minute fight.
164/(386+(104(.627)) = 36.3% also closer to the expected 37.3%

I'm still under the assumption that 2 SoC procs cannot happen within one second of each other due to changes to the reckoning bomb pre-BC, for which I am basing my above calculations.

3. Can there be chain procs?

I believe there can be chain procs of windfury, but not for Seal of Command, so at most 4 attacks; normal swing, windfury, seal of command, windfury.

I looked at every windfury proc in my parse and it almost always lists the windfury proc before my normal swing, then Seal of Command, then the windfury attack. The only discrepancy is the double proc I show. I also see the windfury attacks have the same damage, 558. I'm actually a little suspect on this occurance because of that. The chance for a double windfury proc should not be that rare, .2 * .373 *.2 with a large club, or 1.5% chance per swing. I'd really like some corroborating evidence to prove this. I don't think the server would send me the combat log message twice, but I want to be sure.

I also agree that my trash pull I may not have SoC up at all times, and the 95% windfury is based on my shaman. He is VERY good about dropping totems before pulls and keeping them up, but I could be running out of range on the chaos of sunwell trash. The only guarantee is standing in Blasted Lands. Is a 20 minute parse enough for a ppm of soc test?

EDIT: Posted before reading Sapp's comment. I think that can account for my double proc. I still feel evidence is very highly pointing towards SoC proccing windfury.

Current summary of Theory

1. Autoattack procs Windfury 20%
2. Autoattack procs Seal of Command at a rate of 7ppm (Based on current weapon speed)
3. Seal of Command procs Windfury 20%
4. Windfury procs Seal of Command at a rate of 7ppm (Based on current weapon speed)
5. Seal of Command cannot proc more than once per cycle
6. Windfury cannot proc more than once per cycle *

* - Most unknown, disregard my own double proc as I feel it was explained by the bug(?) Sapp mentioned.

I'm not satisfied with that explanation.


If WF procs off of WF, then we'd see a 25% effective proc rate for all WF using classes.

ie: For every 100 swings, we'd expect 20 WF procs. From those 20 WF procs, we'd expect an additional 4 WF procs. From those 4 WF procs, we'd expect another .8 WF procs, from those .8 WF procs, .16 procs ....

Take the limit of that, you get 20 / .8 = 25 procs per 100 swings.