HoJ does not proc correctly for main-hand damage on rogues, making Rune of the Guard Captain give better DPS. This bug is not present on any other dual-wielding class.
Since I’ve had a lot of rogues claim this isn’t the case, or simply not believe me, Here is DPS data from 10,000 hits - 5,000 dual-wielding, 5,000 with only a main-hand equipped. Below is a table containing minimum, average, maximum damage range data for the HoJ procs. You can find the full graphs in the image linked on this post.
HOJ procs off both main and offhand and special attacks. It’s a good trinket but it’s just good, and it doesn’t track accurately, there is a good post in the big forum on this. The proc tracking of you DW it’s shown to not accurately track. You can discern the proc with math not with recount or logs saying “you proc = xyz”
The diremaul 2% hit trinket is a viable alternative if HOJ does not drop for you.
I’ve seen that post, and I went and looked at my combat logs accordingly. Recount is tracking procs accurately, that post is false.
Edit: that post and me wanting to prove it wrong is why I went and did all this testing in the first place.
Edit 2: Also, if you look at recount for a fury warrior dual-wielding, the recount data shows the procs accurately. It’s only for rogues that the recount data shows the low damage for - another reason that that post is false.
Edit 3 (lol): By that logic, if it procs off the off-hand, which it would over 1000s of hits, it would show the main-hand damage as well for those procs.
To confirm you need to know you your slice and dice casts exactly or flurry price exactly, ability missed, ability dodged, abilities parried, attacks you parried from npc attacks, your auto swing timers for main hand and offhand.
To include a non clickable link you can put it in code block. [code] at the start. [ /code] at the end. Do not include the space between [ and /code]
That was done to prevent it from making the non clickable.
However, it seems people don’t believe me that Recount is proccing correctly, so I am going to update that image within the next day with the Recount results from a fury warrior, and with links to the raw combat logs for all three situations.
When you are dual wield for some reason according to other testers it only tracks the offhand strikes and that tracking may not actually be accurate, the combat log could be reporting things funny. Remember this is not the same combat log was we had in vanilla, this one could work a little different. Also one of the observations shows that there is a delay between the proc itself and the combat log. Any of the following links can take you you the parent thread on this.
Probably has to do with HoJ just resetting your swing timer, instead of actually giving you an extra attack. Sometimes the proc is wasted, if it happens right as your attack would go off normally.
How you should test this thing is going to be tricky, based on what I have read and you have posted the testing quality needs to be improved to track this thing.
You will need to do the following to be 99% certain about HoJ.
use auto attack only.
you will need to time the event down to 1/100th of a second
you cannot use any combo points, abilities, or any gear that provides haste
you cannot be parried, from what I understand parry can advance your swing timer.
you need to track exactly the number times you miss, and some how make sure you know if it was main hand or offhand.
know your weapon swing timers, main hand speed vs offhand speed.
get 10,000 attacks, and even that is kinda a small sample because of how blizzard does RNG, because there really is no such thing as RNG, its a prescribed string of numbers designed to look like RNG, and the parent value is seeded from an unknown, so a larger sample is better. 10K is the bare minimum.
Kinda how I think it may work, and its possibly using the main hand swing timer given that offhand proc’s are said to be on a slight delay, and that would make sense.