ATO: very low proc rate in AoE a bug?

The formula uses recharge and number of targets. The problem is, like I stated before, that the AoE Immobilize recharges too fast and has a 16 target cap. Put the proc in a longer recharging AoE mez and it will have a dramatically higher proc rate.

It's not just the target cap, I've been told that the size of the AoE factors into it somehow.

I also stand by my belief that the formula is intended to normalize the rate for procs that are evaluated per-target. Historically, procs are always evaluated on a per-target basis, even ones that apply an effect to yourself. See Force Feedback for a good example.

The Brute and Defender ATOs work this way. However the Dominator and Tanker ATOs are implemented in a way that effectively makes them evaluate once per cast rather than once per target -- something that until I saw and tested them I didn't even think was possible. So they get the low low percent chance from the formula that assumes they'll be checked against each target, but instead it bails out after the first failure to proc. This is even more problematic for Tankers as most of their attacks are technically AoEs due to gauntlet.

All of this is my theory, and the devs are aware of it as well as the empirical evidence that was gathered in testing the Dominator proc. Even if they choose not to act on it, I still appreciate that they are willing to listen.

Uh, are you really really sure about this? Because I've done some testing on the Brute proc too. (Although not as much as on the Dom proc. The brute kills enemies quickly enough that it slows things down...)

The Obliteration proc has lower ppm than the Brute proc. But if the Brute proc was checking as often with even the lower rate I got for the Obliteration proc in Thunder Strike (51%), it should have had a near 100% rate when used in a group. But it didn't. It was more like 71%.

Now, the number of samples on that last test is pretty low. So it's not very confident. But it does make me pretty sure that when Thunder Strike with the Brute proc slotted hits 8 mobs, it isn't making 8 rolls at anything within the 95% confidence interval of my Obliteration test. Because those would all give a less than 1% chance of the Brute proc -not- firing. What I've seen fits a lot more closely if the Brute proc, like the Dom proc, is being checked once per activation only.

Of course, given that Beta just got another patch, I have to add to that they could have changed the way the proc works *again* since I did my testing.

Really, I have my doubts about the ppm procs normalizing things. In most cases a saturated AoE will still get more procs than a similar single-target. SBE procs in some AoEs will fire a lot more than the equivalent enhancement proc. And as far as I can tell from either reading discussions or my own testing, all the self-buff procs have a ridiculously low fire rate in AoEs compared to single-target powers, which is hardly normalized at all.

-Morgan.

I did some testing and it turns out that the Brute proc suffers from this, too. That's... surprising. It means that the issue isn't what I thought it was, which probably also means that that even if it were determined to not be working as intended, it's not an easy fix .

For reference, here's a baseline:

Will of the Controller: Chance for Psionic Damage

Fossilize
Times activated: 108
Enemies hit: 101
Procs: 63
Approx proc rate per target: 62%

Stone Cages vs. 1 target
Times activated: 117
Enemies hit: 101
Procs: 7
Approx proc rate per target: 7%
Approx proc rate per activation: 6%

Stone Cages vs. many targets (>16)
Time activated: 103
Enemies hit: 1561
Procs: 164
Approx proc rate per target: 10%
Number of activations in which the proc fired at least once: 86
Approx proc rate per activation: 83%

The last number is highlighted, because IMO that is the number we should see for self-buff procs that are not allowed to stack. You'll notice that it's slightly higher than when slotted in Fossilize because I was at the target cap most of the time. With an average number of targets, the effective proc rate would be close to the single target.

Notice how the proc rate per target is similar to when hitting a single target, but overall the results go up because I'm hitting more of them.

Here's the brute proc in Footstomp. I also had Force Feedback slotted and am including that for comparison, as it's also a self buff which is not allowed to fire multiple times from a single activation.

Foot stomp vs. 1 target
Times activated: 106
Enemies hit: 100
Procs: 45
Approx proc rate per target: 45%
Approx proc rate per activation: 42%
Force Feedback procs: 8
Approx FF proc rate per target: 8%
Approx FF proc rate per activation: 8%

Foot stomp vs. many targets (>10)
Times activated: 118
Enemies hit: 1180
Procs: 60
Approx proc rate per target: 5%
Approx proc rate per activation: 50%
Force Feedback procs: 76
Approx FF proc rate per target: 6%
Approx FF proc rate per activation: 64%

There's a bit of noise due to the lower sample size (I was testing a lot of powers). The key here is that the per-activation proc rate is very similar between hitting 1 target and 10, while the per-target proc rate drops drastically. This is in stark contrast to Force Feedback, which behaves just like the Controller damage proc.

Note that I'm fairly confident that the percent chance for it to fire isn't actually changing based on number of targets hit. That leave two possibilities:

(1) The proc is somehow working on a per-activation basis
(2) The proc is failing to check additional targets past the first

The difference is subtle, but can be tested. I can't at the moment, but if someone can slot the Brute ATO in a PBAoE power and stand around for a while with it on auto, it would be enlightening to see if it ever procs when hitting no targets.

I also want to test the Defender ATO, unfortunately my Defender didn't finish copying to beta and I ran out of time. I'll try to do that at some point.

There's a couple other data points that would be highly useful in determining exactly what is going on here. If someone gets a chance and could test the following, it would be immensely helpful:

• When Achilles' Heel: Chance for Resistance Debuff is slotted in an AoE power, can it ever proc on more than one target within a single activation of the power?
• When Siphon Insight: Chance for +To Hit or Decimation: Chance for Buildup is slotted in an AoE, how does the proc rate compare when only hitting a single target versus using it against a large group?

It wasn't changed, that was just never the whole story for AoE attacks.

There's a thread in the Issue 22 Beta forums where people are discussing and testing proc rates, and someone mentioned that the in-game descriptions have incorrect PPM for some procs. City of Data says that the Dom proc is 5 PPM, and I'm inclined to believe it based on my tests in single-target powers.

Wormhole will probably still get a near-100% proc rate due to it's long recharge. (The "near" being due to... well, see below.)

There's something I noticed that may support your second theory. When I did that last test run of the Dom proc in Roots, I was a bit puzzled as to why the test run with multiple enemies had a slightly lower proc rate than the single-enemy test. Neither has that narrow a confidence interval, but it still didn't feel right. Well, yesterday I was reminded of a post I'd read about another ATO proc (don't remember which one), and someone said it only had a chance to fire if the attack hit the -targeted enemy-. The multiple-enemy test doesn't exclude ones that missed the targeted enemy, and since my test dummies all have the same name, there's no way to tell which attacks did or didn't. I did watch while doing a little more messing around with the Brute proc in Thunder Strike, and I never saw that proc fire when it missed the targeted enemy. But I also didn't watch all that long, because it was really boring.

I've got a couple ideas on how to do a reasonably conclusive test of this, although proving a negative tends to be a problem. We'll see how it works out.

(Also, this is why I say "near-100%" about Wormhole - because it would still have to succeed in that one particular hit roll if this is correct.)

If this is the case, it makes one wonder what's up with these procs in damage auras. Arbiter Hawk said this:

Damage auras have no selected target, obviously, but there do seem to be accuracy checks involved. It's really hard to tell what's going on with the procs in auras though. Except that they're really, really bad in them.

I'm pretty sure it can proc on more than one target. People tend to put it in AoEs a lot and all. However, I'm not sure what this is supposed to tell you, since it's an enemy debuff and this issue is all centered on self-buff powers. (Well, the defender proc isn't a self-buff, but it seems to end up functioning the same way.)

The second's an odd one to bring up in this context, since that's a single-target ranged proc. However, there are at least two powers I know of that hit multiple enemies, but take those sets.

One is Electric Control's Jolting Chain. It's well established that self-buff procs in this power only affect the caster on the first target, not the jumps. On the jumps, the enemy being hurt is technically the caster, it gets the effect of the proc. (Jolting Chain with the Entropic Chaos proc has been observed to heal enemies.)

The other is Traps' Acid Mortar. In pet powers self-buff procs are known to get one chance to fire on the caster when the power is cast, and after that has chances to fire on the pet's attacks. (With caveats based on the nature of the proc - for instance, if a pet takes immobilize sets, an immobilize proc would only fire on attacks that can immobilize. Not an issue in this case.) I figured, I have a traps corrupter, it's worth a try? Well, the SBE at least isn't firing off even once in most volleys, let alone twice. The regular proc might be better than the SBE in this case, but I don't already have one of those sitting around. Not sure how much further pursuit this one is worth.

Siphon Insight? I've got something that can test that too. Scrapper, Darkness Mastery, Torrent. Never saw it proc twice on one cast, although my number of samples is not that high. Enough that getting 4% on procs per hit seems a bit too low for a proc with a stated rate of 10%. (13/308).

-Morgan.

Hmm, interesting. I'm not quite sure how that could happen, but if it is it's certainly a new wrinkle. It's fairly easy to disprove though (named boss and combat log), so that's one avenue toward testing its viability.

PBAoE powers have the same issue - no explicit target. I suspect that the targets are evaluated in the order of distance from the center of the AoE in all cases, which is why it's would be the targeted enemy for AoEs that use a target.

The reason I asked about this one is because Achilles' Heel is designed so that it can't stack from different casters, and is set up in a similar way to the ATO procs. City of Data has the details.

I almost didn't mention Decimation, but threw it in there because both use the same mechanics, and there are some odd slotting possibilities with pets occasionally (you'd have to have some way of observing the proc on the pet, but often the pet window and/or combat log can suffice).

It also might affect the splash knock on Propel when it gets added, but I'm not sure how the attribute checking works with ranged damage sets.

I should have been more clear, I was actually asking if anyone could test it with the crafted version of the enhancements, not the SBEs.

How fast does Torrent recharge for you? The Siphon Insight proc has a 10-second internal cooldown on it that might be skewing the average.

The real question I'm interested in seeing information on is how does the rate compare from using Torrent on one enemy to using it on a big group.

These procs aren't identical to the ATOs, but are similar in various ways, and I'm trying to piece together a picture of exactly how they function mechanically. I think I've managed to wrap my head around it and have a working theory of why it's behaving this way. Long story short it has to do with how the game handles periodic (ticking) effects, and its apparent inability to cancel them upon a failed tick and be replaced by another in the same processing interval. I was close before, but just looking at it from the wrong angle -- I thought the cancel was preventing it from checking other targets, but it was actually trying to do the opposite.

Based on this, I'm going to predict the following and we'll see if I'm anywhere close:

• Achilles Heel: Will apply to multiple targets in an AoE, and the proc rate for the Attuned version (don't remember if it's available or not) will be close to what it should be per target. Won't apply to a target that already has the debuff.
• Siphon Insight: Will act just like the ATOs -- only one target per cast will be checked and the proc rate for the attuned version will be exceedingly low as a result.
• Decimation: Tricky to test, but Acid Mortar should be able get the Boost Up buff on it. It should also behave like the ATOs and only check once per activation of its mortar attack, regardless of the number of targets hit.

As a tangent, the Brute proc is (incorrectly, IMO) using combat modifiers. I smacked a level 2 Hellion in AP and got a proc of 29.5 Fury...