I happen to have a Torc on my barbarian so I went to test out its proc rate. Unfortunately, I don't have a conc opp since I don't play casters, so I'm not able to test it currently.
The testing location was the entrance of "The Tide Turns" on elite. After killing the caster and some of the melee trash, there are 7 archers left, ideal for testing proc rates of guard effects that proc from ranged. As usual, video was taken and then I analyzed the video's combat log. Unfortunately I probably won't be uploading the videos again because like before, they're very long and it takes a long time to get them uploaded. The videos totaled 3.8 hours.
At this point it might be worthwhile to differentiate between different types of guard effects. They are all when a monster tries to attack you, rather than when you attack a monster (some monsters also have guard effects, in which case it's when you attack them). Generally guard effects can be split into three types:
* Proc on attack (which I also call "hit attempt"): when an attack is done. Things like blurry are in this category, in fact usually these procs determine whether or not the attack goes through or is stopped via the proc.
* Proc on hit: when an attack successfully hits a target. This does not mean that it actually damaged a target, since the target may have DR, elemental resistance, etc., but it means that the combat roll was made and it was successful (did not miss).
* Proc on damage: When an attack hits a target and does damage to the target. This means that there was some amount of damage which the target didn't fully mitigate.
Torc falls into the third category, procs only when the victim actually takes damage past his DR. I verified it by looking through when it procs and seeing if the hit that caused it did damage. I also double-checked by using a shield (which was sufficient to block all damage), and got hit 30 times without having the torc proc, which has a 0.76% (1 in 131) chance of happening.
Summary of results:
Torc:
* Proc on damage: 14.9% +- 0.6% (95% confidence interval), so likely
15%
* Damage: Average 22.502, stdev 3.812, so likely
5d6+5
I should note that this is the proc rate for ranged attacks. There was some forum speculation about it being different for ranged vs melee attacks, which was not tested here.
The distribution of SP gain from torc was:
Code:
SPgain Count
11 2
12 3
13 6
14 18
15 30
16 46
17 63
18 116
19 141
20 191
21 188
22 197
23 178
24 191
25 150
26 122
27 103
28 84
29 50
30 34
31 20
32 9
33 7
Mabar cloak invis effect:
* Proc on damage: 10.2 +- 1.6% (95% confidence interval), so likely
10%
Levik's Defender heal effect:
* Proc on damage: 2.13 +- 0.25% (95% confidence interval), so likely
2% (it says so in the combat log, but worth verifying)
* Heals for 90 damage base, affected by healing amp
The dataset was:
* 69808 attacks
* 15493 hits (including those where 0 damage was dealt due to DR)
* 13120 damaging hits (where damage was taken)
* 1949 torc procs
* 279 Levik's Defender heal procs
The Mabar cloak was only used a portion of the time; it procced 146 times out of 1426 damaging hits while it was being used.
For the torc, in effect it means an average SP gain of 3.375 SP per damaging hit. This might be useful to know if you're trying to get SP back. For example, if you heal yourself for 200 HP and it costs 30 SP (just making up numbers because I'm not that familiar with healers), you had better be taking less than 3.375*200/30 = 22.5 damage per hit, because otherwise you're losing net HP/SP trying to use the torc to recharge.