One of the stats you will find on your persona window is the DPS modifier. This modifier has an intrinsically confusing name. When I’m talking about “DPS” am I talking about the modifier on my persona window, which is buffed by items, spells and combat abilities, as well as some AA lines? Or am I talking about my overall, parsed, Damage Per Second so let’s adopt some overall terminology.
I will write, at least in this post “DPSmod” when I mean the modifier, and just plain “DPS” or “damage/sec” when I mean overall, parsed or theoretical, damage-dealing performance, which can come from many sources, be it criticals, attack speed, +damage, good weapons, etc.
Ok, so what does the DPSmod do? First, we note that it is on a diminishing returns scale, just like Attack Speed is. So it has a hard cap of 200. The rollover says that this modifier “increases DPS by X%”. Anecdotally, I have heard it said that it makes autoattack hit harder.
So I made some runs with my level 33 Assasin. Assasins have a self-only buff that boosts DPSmod. To a value of 38, which the rollover says “boosts DPS by 41%”.
I was dual-wielding a piercing weapon in my primary hand, and a slashing weapon in my secondary. I fought the rust monsters (which were gray to me) on the shores near the Butcherblock docks, both with and without this buff on, and looked at an equal number of attack, with and without this buff. I sorted them by size, and produced this graph:
Doing the same thing for the secondary (slashing) weapon, we get this:
There’s some oddities in these data, but first let’s look at this chart of the ratios between DPSmod of 38 versus DPSmod of 0 for various statistics.
This dataset is consistent with the hypothesis that a DPSmod of 38 makes all autoattack melee hits 38 percent harder. Though the performance in this run was a bit lower than that, perhaps 35 percent. We also note that the slashing damage from the secondary weapon was helped more, and it was helped in a very odd way. The minimum damage produced by the primary (piercing) weapon was not very much larger under a DPSmod of 38 than under a DPSmod of 0. In fact, we note on that there is a fall off the normal line at the extreme left of the red line in the first chart above. Somehow, the minimum damage of the piercing weapon wasn’t increased by 41 percent.
By contrast, the secondary (slashing) weapon seems to have a flat spot at the extreme left end, the minimum damage seems to be a tad higher than you would expect. But while this second effect could be due to roundoff error and a small sample, the low minimum of the primary (piercing) weapon does not seem to me due to small sample size.
In any case, overall the hypothesis that DPSmod affects how hard your autoattack hits seems confirmed. As such, it seems every bit as important to overall damage-dealing as Attack Speed. And when you are trading off the two, remember that you will get more benefit out of boosting whichever modifier is smaller, both because of the diminishing returns scale, and because the most efficient allocation of some fixed number of points is to have both numbers equal.