The toughest thing to figure out of all is what effect STR has on damage output. Yes, it makes it bigger, but how do you trade off an item with, say, 10 fewer STR but a +1 Double Attack? Will it increase or decrease your output DPS?

I won’t answer that definitively in this post, but in this post, I’ll look at the effect of STR on auto-attack melee damage.

Gathering actual data on melee hits is very difficult. In order to get estimates that are accurate to one part in one hundred, or one percent, it would be necessary to gather data on 10,000 melee hits. This would have to be done at each strength level sampled. I consider that infeasible, so I did something else.

Namely, I looked at the effect that increasing STR has on the minimum and maximum damage shown in the persona window. It used to be that when you examined a weapon, the effect of your STR and other modifiers on the output damage of that weapon was shown in the examine window. This is no longer so. The examine window will always show the same damage, regardless of your strength. Below, we will refer to that as the **nominal damage** or perhaps the **examine damage**. In the persona window, the minimum and maximum damage values of the weapon, as modified by STR, are shown. We will refer to that as **persona damage**.

Prior studies have led me to believe that damage is rolled as a uniformly random number that falls between minimum and maximum persona damage, with a small chance of exceeding maximum in an “exceptional hit” tail. I calculate average damage as simply the average of the minimum and maximum. I expect that this will give us a good idea of the STR diminishing returns curve, and the effect of this tail can be ignored for our purposes.

Ok, on to some data. First, I looked at the Bo of Flowing Blood. Here’s a chart of Persona Damage versus STR.

I also calculated the ratio of persona damage divided by examine damage for minimum, mmaximum and average damage.

The Bo of Flowing Blood is a 2-handed weapon. Just to make sure, I ran these numbers for a one-handed weapon, Dabner’s Chitin Shredder. The proc on this weapon was ignored.

It is entirely clear that STR acts as a multiplier to weapon damage. Which means that the effect of STR on autoattack damage is entirely independent of weapon speed or dual wielding. Which is a good thing, in my view. Additive damage favors fast hits, multiplicative damage is neutral.

Ok, so what are we seeing? At the top of the range shown, about 650 STR, adding 20 points of strength increases the ratio by about 0.025, which is an increase of about 0.7%. Presumably this translates to an increase in auto-attack DPS of about 0.7%. I estimate that at this level of STR, +30 STR is equal to about +1.0% in autoattack damage.

You can see that the curve is flattening out some, and it will very likely continue to flatten. I’ll be studying the effect of STR on Combat Arts soon, though I expect it to be multiplicative, with exactly the same ratios.

As a rough guide, 30 STR is roughly a one percent boost to

damage output. This will aid greatly in deciding gear tradeoffs, as long as you remember to favor parameters that are small, since the most efficient allocation of points will always be an equal distribution of them.