How can you conclude that?
Unhasted he got 394 procs in 59 min.
If hasted had no effect on PPM, we could extrapolate that in his 42 Min of hasted testing he should have gotten 280 procs.
Instead he reported 363 procs in this timeframe.

How on earth can you conclude that haste did not have an impact on PPM as shown by this test?

Deleted - need to check first

Haste wasn't the only variable factor in his tests. Number of hits also changed.

Would be surprising, as combat potency states a % chance and isn't ppm iirc. However, if it is affected other % chance procs like poisons should probably be tested as well.

Sorry, I can't follow you.
Sure the number of hits changed.
Nonhasted he hit 2253 times in 59 minutes. That would mean he would hit 1604 times in a timeframe of 42 minutes. Instead he hit 2057 times in that timeframe (which comes down to a haste of 28%), because he was hasted. This was the whole purpose of the test.

The screencap shows explicitely that the proc/hit average did not change from hasted to unhasted. This goes directly against the interpretation that this test somehow shows that PPM and the "flat proc chance on hit" derived from a PPM value is computed relative to current weapon speed.
In contrast this test indicates that the proc/hit average remained the same, and as a consequence the PPM went up in the hasted test almost in an exact 1:1 correlation to haste.

He hit 29% more, and he procced 29% more. I find this test very telling. (If indeed nothing else was change and only autoattack and SnD were used)

I was looking at LAST instead of TOTAL.

I did too sorry for the confusion suicuique

So let me get this straight.
The proc rate on the PTR and on live is almost the same.
So this would mean absolutely no change to the mechanics.

Is this just a case that for the very first time, someone has actually observed how a part of the proc mechanics works, rather than observing a new change to said mechanics?

The test done by Baldycenarius (on live servers) implies that PPM profits from haste in a direct way.
The test done by Lujaar tells that PPM (on hasted attacks) does behave on PTR the same as it does on Live.

From this we would conclude that PPM benefits on PTR the same from haste as it does on Live.
But this is in direct contradiction to the OP.

Confused?
I know, I am.

I think this might need some looking at. I've been busy lately, but I'll see what I can do.

If Baldycenarius could repeat his test on the PTR, that'd be a big help.

Yes, one should check for the same type of proc on live and on PTR.
So, with fiery weapon we have tests now on live, need one for the PTR as well to see

For mongoose, we already have both, and they show that the PPM mechanic hasn't changed on PTR. So maybe it was changed some patches ago, but only for haste procs (such as mongoose), while others remained unaffected and still work according to 'common' knowledge (i.e. the PPM goes up with more haste effects).

there was data surely, but is old data.
and i have no copy.
if there was a change in live it was an hidden change.

i'm sure that snd was increasing my crusader active time at 60.
i hope it was not only a feel.

i'm the one that started the isterical post about this on eu forum, but i think we need it.

I'm afraid that is already implemented in a different but more obvious way:
AP vs Haste. AP is a linear stat on it's own, yet the effect scales exponentially on haste.
Removing this exponential effect could only be achieved by removing either one of these, which would probably have to be the haste component anyway.

Yeah, the time I heard about this was before I even knew of the existence of the EJ forums, so it's a long way back. Read it in some of these Rogue FAQ posted in the official WoW forums, which meanwhile also have been restructured, so I don't think anyone could find those testings, if there were any.

Maybe when i posted i should have explained fully. Yes the LAST part shows that they are the same in a single 5min test, but you can get lucky or unlucky sometimes, as with all %based effects. I saw as high as 10 on PPM(9 being the common) for hasted Fiery, and never above 7.5 or so for non-hasted fiery. I think its pretty conclusive how it works on live for rogues now.

Nothing changed but the number of hits in my test. I used setup(combo points from dodging) as my source of slice and dice in the haste test. Sometimes i got unlucky(like the LAST part of hasted) that i didn't generate very many combo points, but i also wore crystalweave cape to guarantee i was always at least somewhat hasted.

As requested i will try to get on the PTR, but i deleted my Test folder because of the automatic test folder update like live patches when i log in now and getting booted during a raid then logging back in to a patch is stupid.

I'm really sorry if this is raining on everyone's parade...

...but "procs per minute" implies, well, a constant number of procs per minute. the fact that PPM was internally translated into a %chance per hit based on weapon speed, and that haste effects led to a greater number of PPM than originally intended, and that folks did incredible dps as a result of that..... certainly sounds like a bug to me. Regrettably, this was a bug that absolutely worked in our favor, but, nonetheless, it should be fixed.

Best case scenario: this "fix" means the PPM per item becomes more constant, and developers can buff the direct impact of procs, because they no longer have to worry about haste increasing proc rates beyond the intended PPM.

oh. then ignore my previous dribble.

The larger issue is the instant attacks. If the percent chance to proc on an instant attack is adjusted to the new weapon speed, then you will do the same number of instant attacks but each will have a lower percent to actually proc.

Let's say you auto-attack every 6 seconds with 1 PPM. That's a 10% chance to proc each hit. Let's also assume you get an instant attack every 6 seconds. On average, you'd expect to see 2 PPM.

However, let's say someone uses "Bloodlusting Heroism" that doubles your attack speed. This means you'll attack every 3 seconds, or 20 times a minute, and *possibly* for a 5% chance to proc each time. You'll get get 1 PPM, though (higher if it stays at 10%, but this thread is about it changing to 5%). The problem comes when you look at the specials: 10 attacks at 5% chance = 0.5 PPM, because it was adjusted to the new weapon speed.

So, you'd go from 2 PPM to 1.5 PPM (using the completely made up values I provided for this example) if this change is actually true. Hence the hype.

I finally got a chance to take some nice big samples with Don Santos' Hunting rifle (it has no internal cooldown). I set ProcWatch to look for "You gain Santos' Blessing" messages so that I would only track gains, and reprocs are not a concern for this test. This was done on LIVE servers at Dr. Boom. These were done with auto shots only, so yellow damage is not assessed.

Test 1 Unhasted: I unequipped my ammo pouch, and made sure Aspect of the Hawk was not up (I have 5/5 iAotH). My shot speed was 2.25s.




Test 2 Hasted: I equipped my ammo pouch for a shot time of 1.96s and activated iAotH to proc extra hastes in the back-ground, with Quick Shots up my shot time drops to 1.7s for its duration.




My results indicate a slight decrease in proc rate with haste. The results are close enough that it can be explained by standard deviation in the sample, but this directly counters the previous test showing an increase in proc rate.

That assumes the adjustment applies to yellow as well as white hits. Tyran's test on page 2 seems to indicate it doesn't.

I just realized your test might be tracking both gains and losses. If you set it to "Fiery" it will track every time that is mentioned in the combat log, both as a gain and a loss. You might have had more reprocs in one of your tests so you wouldn't see as many "losses" on fiery, if that makes sense, which could have muddied the sample.

Fiery isn't a buff - it's an instant damage on proc.