|This is the talk page for discussing improvements to the Aim article.|
Aim Considered Harmful to DPS? Edit
The math-lite explanation is that Aim causes you to you hit half as often for less than twice as much. The details get a bit complex, but that's the short story that everyone can grasp in about 3s of reading and understand immediately.
The obvious way around this is to eliminate the reduced rate-of-fire penalty. There are only two ways available in game that reduce the penalty enough for Aim to pay: Rock Mastery (from Shale) or the "unfixed" version of Repeater Gloves. If you lack those, Aim will reduce your DPS.
Below is mechanical "proof" in case you want to see formulas and numbers.
Now that I finally figured out how to word the reason Aim reduces DPS in one short sentence instead of an ugly wall-of-math, I'll add a short caveat to the main page notes section.
I was just doing a few simple calculations to see if/when using Aim outperforms not using Aim, and I think Aim is always a DPS loss even at 100% (modified) critical rate and 3.5x critical multiplier. Here's the math behind it:
non-aim DPS = (Dmg(1- Crit.Rate) + Dmg*Crit.Rate*Crit.Multiplier) / Base.aimtime aimed DPS = ((D+3)(1-2C) + 2(D+3)CM + 5) / (B + Aim.penalty) dps = d(1+c(m-1))/b adps = (8 - 6c(m-1) + d(1+2c(m-1)))/(b+a)
Taking some liberties and assuming that the dmg+3 and penetration+5 are insignificant enough to ignore for purposes of this comparison makes this a lot easier.
adps = d(1+2c(m-1))/(b+a)
This drops damage out of the picture and let's us very easily solve for the inflection points for aim time, critical chance, and critical multiplier given that we know the other values; ie.
a < b / (1 + 1/(c(m-1))) a/((b-a)(m-1)) < c 1+a/(c(b-a)) < m
We know B=1.6, A=[0,1.8], C=[0.03,0.5], & M=[1.5,3.5]. So, where are our boundary values? When does Aim pay?
a < 1.6/(1+1/(0.5*2.5) = 0.71. 1.3 / (0.2*2.5) = 2.6 < c 1+1.3/(0.5*0.2) = 14 < m
This means Aim is never pays. With 50% base crit, working haste (-0.2s), & aimspeed bow (-0.3s), you still need a multiplier of 14 to win, which is impossible. With a maximum multiplier of 3.5, haste, aimspeed bow, you need a pre-aim critical chance of 260%, which is of course also impossible -- anything more than 50% means "can't be done", since values over 0.5 are lost.
The only way to make aim pay off is to significantly reduce the time penalty from aim. There are only two ways to do that: Repeater Gloves (the "unfixed" version) and Shale's Rock Mastery buff. Either of those methods will set your aim time to zero and thus Aim will win regardless of your ranged critical chance or critical multiplier.
One last calc. Let's suppose you don't want to use Shale and you like the patched gloves. Can you make aim pay off? Let's suppose 2x haste for -0.4s, a rapid aim bow, and one extra rapid aim item (hat or "fixed" gloves) for another 0.3s for an Aim penalty of 1.5-0.7=0.8. Ah, but we already know the answer from our boundary calc showing a<0.71 required with 50% chance and 200% in bonus damage. Even if you use two rapid aim items (hat + fixed Repeater Gloves), you still would need extremely high crit chance AND nearly maximal crit damage to make aim pay.
BTW, because we dropped the dmg+3/pen+5 out, the true picture depends on your current damage, too. The higher your base damage and/or lower your target's armor, the more your results will match the simplified calcs. But, the picture is so grim for required chance & multiplier anyway that a low damage character will never be able to come close to acquiring the gear to make Aim pay off.
- Sorry, your calculations are interesting but the conclusion that aim was "guaranteed to reduce DPS" without Rock Mastery or gloves is questionable given that you completely ignore most of the bonuses of aim. First you ignore the +8 damage against highly armored targets (with Master Archery) to simplify your calculations. At least you say that you do. Second, you silently ignore the massive +15 attack bonus which affects hit percentage. Ignoring that is only valid if the hit percentage is already 100%, which is the case if the archer's attack score is at least 46 points higher than the target's defense. If both are e.g. the same, aim gives a 69% hit percentage as opposed to unaimed 54% which on average is equivalent to a 28% damage advantage irrespective of other bonuses. Surely hit percentage should be taken into account when calculating DPS because a miss always inflicts zero damage.
- Most importantly, your argument that it's better to perform two unaimed attacks rather than one aimed attack is not at all applicable to special attacks. You don't get to choose between one aimed arrow of slaying or two unaimed ones because the talent has a 60s cooldown and eats a fair amount of stamina as well. Same goes for scattershot (40s cooldown) etc. That's why you may generally want to maximize hit percentage even at the cost of decreased attack speed for special attacks.
- I do take away from this that it may usually be better to disable aim for auto-attacks though the difference may not be as profound as your calculations suggest. Electrolisk (talk) 10:31, January 4, 2013 (UTC)
- Electrolisk's math is sound, but the problem is that in real game play, misses (melee and ranged) are very rare past L6 or so. Also, archery specials aren't big DPS in and of themselves; they're mostly useful for debuffs, CC, and shattering. Arrow of Slaying is a red herring; it does HUGE damage versus weak, cute, and cuddly critters, but only regular crit damage versus even slightly tougher targets. Using it (ever) is a waste of stamina and time (it has a long animation time).
- This isn't to say that your points aren't valid. Aim is incredible during the Origin stories for an archer when even the tiniest bump in any stat can translate into a very significant effect. It's just that I rarely use Aim much past the Origin stories. The hit numbers will be smaller, but the fights will be shorter and easier without Aim. If you don't have rock-mastery or unpatched repeaters, turn Aim off and see for yourself. Bjond (talk) 08:20, October 10, 2014 (UTC)