DPS: Snipers have it better?

[JoyProtocol]

This is not just about operation bosses tho. I used the "what if a boss has 66% defense" argument just to bring the point across.

 

The disparity of your formula for offhand reduction will also impact everyday gameplay. For instance, take a common scenario of a sniper and a gunslinger taking shots at each other from behind their cover screens. Cover screen provides 20% ranged defense.

 

Sniper damage: 120% * 0.8 = 96%

GS damage: [100% * 0.8] + [30% * (0.8-0.33)] = 80% + 30% * 0.47 = 80% + 14.1% = 94.1%

 

This means IF you are correct that the offhand accuracy is always reduced by static 33% penalty (that does not scale with target's defense), a simple high-noon duel between a sniper and a gunslinger would result in an almost 2% damage difference in favor of the sniper.

 

This is why i suspect your assumption that offhand accuracy penalty is always a flat 33% reduction is incorrect. I just can't believe bioware would design the offhand mechanic in such a way that goes against everything they were saying about mirror classes being equal.

 

From what I understand (by comparing the numbers), you are missing an important balancing factor in GS versus Sniper - GS have a lot more variance in damage. While "Over 10,000 separate cases" the Sniper would probably end up with superior damage, a Gunslinger has a marginally higher chance of being noticeably superior.

 

What I mean is if both attacks hit, and they both hit at maximum damage, the Gunslinger deals superior damage.

 

Furthermore, Gunslinger benefit more from accuracy. As you are probably aware, Accuracy over 100% reduces target defenses.

 

So, let's say we have 105% accuracy

Sniper damage: 120% * 0.85 = 102%. Maximum hit is 120%.

GS damage: [100% * 0.85] + [30% * (0.85-0.33)] = 101%. Maximum hit is 130%.

 

 

So, let's say we have 110% accuracy

Sniper damage: 120% * 0.9 = 108%. Maximum hit is 120%.

GS damage: [100% * 0.9] + [30% * (0.9-0.33)] = 107%. Maximum hit is 130%.

 

 

So, let's say we have 115% accuracy

Sniper damage: 120% * 0.95 = 114%. Maximum hit is 120%.

GS damage: [100% * 0.95] + [30% * (0.95-0.33)] = 114%. Maximum hit is 130%.

 

Note that with just the accuracy, Gunslinger always has the best "max" damage, but more accuracy gaps the 2% difference.

[Meetstick]

Tibbel, i believe your calculation for gunslinger is incorrect. The yellow part, specifically:

 

Original Quote posted by Tibbel:

For gunslingers:

C_g = W * (a + .3 * (a - .33))

= W * (a + .3 * a - .099) = W * (1.3 * a - .099)

 

dC_g/da = 1.3 * W

 

 

If, in your calculation, a is a very small value, then you actually end up with negative accuracy for the offhand(because you applied the offhand accuracy penalty by subtracting a static 0.33 instead of multiplying by 2/3).

 

What i think it should read is:

 

Quote:

For gunslingers:

C_g = W * (a + .3 * (a * 2/3))

= W *a* (1 + .3 * 2/3) = W *a* 1.2

 

dC_g/da = 1.2 * W (equal to the sniper)

 

From the tooltip, we know that if the mainhand accuracy is 100%, then the offhand accuracy is 67%, a 33% accuracy penalty. This however does not suggest that the offhand penalty will be the same flat 33% even if the mainhand accuracy is not 100%(as your formula assumes).

 

I think it is reasonable to assume that the offhand penalty is not a flat 33%, but rather a 33% reduction relative to the mainhand accuracy(as my formula assumes). Meaning, if the mainhand accuracy is 60%, then the offhand accuracy will be 40%(a 1/3 reduction compared to the mainhand) and not 27%(a flat 33% reduction).

 

The reason i think this is the case is that if the character is attacking a target with very high ranged defense (like a boss) where the base chance to hit is below 33%, then using your formula the gunslinger offhand would not hit at all, giving the sniper a significant 20% DPS advantage. I consider it very unlikely that BW would design the mechanic with such a huge inbalance in mind.

 

This is 100% (That's times 1.0) correct.

 

If you're going to set up an equation and say there is a 2/3 modifier to a variable, you would NEVER simply subtract an approximated value for 1/3. You're subtracting 1/3 of 1! What if it isn't 1 that you need to modify? Plus, when you differentiate, that subtracted value GOES TO ZERO, while when you are multiplying by 2/3 the a goes to 1 and you are left with 2/3... That is simply gone done the first way. Also, it will affect the final answer as when you distribute your .3 you're not multiplying into a term that is then multiplied, you are multiplying into a term that is subtracted... and that disappears when it shouldn't in the derivative. This reduces your eventual final answer significantly... by a factor of .1, as it turns out.

 

Look, it's natural to feel the grass is greener on the other side. There are legitimate gripes GS can have about their Imp counterpart's abilities. Instant EP and a non-stop-us-in-our-tracks-while-the-kick-goes-off stun are real downers when you don't have them. Also, as there is greater spikiness to the GS damage (individual hits will be greater or lesser depending on off-hand hits/crits), it may seem unimportant to realize that they will average out over time. But, please, don't quote bad math and say "This proves it!" There may be a few of us on these forums who can set up a differentiation to determine rate of change with respect to accuracy. And as smart players choose GS...

 

=P

[Elear]

As for accuracy translation to damage, you're forgeting that only main hand weapon gets bonus damage, so in reality offhand damage is closer to 10% of total, or less. More power you have, less important offhand is.

 

I would say that slightly more important is that Quick Aim procs only from offhand crits, but I'm not so sure about my observations

[JoyProtocol]

As for accuracy translation to damage, you're forgeting that only main hand weapon gets bonus damage, so in reality offhand damage is closer to 10% of total, or less. More power you have, less important offhand is.

 

I would say that slightly more important is that Quick Aim procs only from offhand crits, but I'm not so sure about my observations

 

I'm 99% sure this is incorrect. Could someone who's not at work confirm?

[Tibbel]

This is 100% (That's times 1.0) correct.

 

If you're going to set up an equation and say there is a 2/3 modifier to a variable...

 

I know how the algebra and calculus works. (I'm specced full Engineering in RL.)

 

You're both absolutely correct that if offhand accuracy = mainhand accuracy * 2/3, then snipers and gunslingers do scale exactly the same with accuracy.

 

It's not, though. Offhand accuracy is simply main hand accuracy - 0.33.

Take base ranged accuracy. Mainhand accuracy = 90.00%. Offhand accuracy = 57.00%.

Note that 0.57 is not 0.90 * 2/3.

 

Alternatively, look at base special (tech) accuracy. Mainhand = 100.00%. Offhand = 67.00%.

That's not a precise 2/3rds ratio, either.

 

Next, put on an arbitrary amount of accuracy gear. Accuracy for mainhand and offhand attacks increase by the same amount. If offhand was scaling like mainhand * 2/3, then we would expect 1% of added accuracy to increase mainhand accuracy by 1.00% and offhand accuracy by 0.67%. Instead, both scale at the same rate.

I agree that it's a discrepancy between the classes, but it's one that exists. In most situations it's not a large difference (but it is in some edge cases, like Sharee has mentioned -- getting hit with Diversion, for example).

 

By the way, if anyone's interested in the formulas for accuracy (and other stats), this thread on SithWarrior.com lists them in a format that can be copied straight into a spreadsheet. The main-hand and off-hand formulas are identical except for the base values.

Edited March 15, 2012 by Tibbel

[Meetstick]

I know how the algebra and calculus works. (I'm specced full Engineering in RL.)

 

You're both absolutely correct that if offhand accuracy = mainhand accuracy * 2/3, then snipers and gunslingers do scale exactly the same with accuracy.

 

It's not, though. Offhand accuracy is simply main hand accuracy - 0.33.

Take base ranged accuracy. Mainhand accuracy = 90.00%. Offhand accuracy = 57.00%.

Note that 0.57 is not 0.90 * 2/3.

 

Alternatively, look at base special (tech) accuracy. Mainhand = 100.00%. Offhand = 67.00%.

That's not a precise 2/3rds ratio, either.

 

Next, put on an arbitrary amount of accuracy gear. Accuracy for mainhand and offhand attacks increase by the same amount. If offhand was scaling like mainhand * 2/3, then we would expect 1% of added accuracy to increase mainhand accuracy by 1.00% and offhand accuracy by 0.67%. Instead, both scale at the same rate.

I agree that it's a discrepancy between the classes, but it's one that exists. In most situations it's not a large difference (but it is in some edge cases, like Sharee has mentioned -- getting hit with Diversion, for example).

 

By the way, if anyone's interested in the formulas for accuracy (and other stats), this thread on SithWarrior.com lists them in a format that can be copied straight into a spreadsheet. The main-hand and off-hand formulas are identical except for the base values.

 

That would be great, except it's wrong.

 

When you state that offhand accuracy is main hand accuracy minus 33%, what you're saying is that offhand accuracy is 33% less than main hand accuracy. That is NOT the same as main hand accuracy minus .33. .33 is one third (rounded down) of 1. You are subtracting a fraction of one.

 

I'm sure I'd be real impressed that you were speccing engineer IRL, if I wasn't speccing physicist. What you're studying to be doesn't make the math right. If it's not your math, stop defending it. If it is your math, look at it and fix it. THAT'S what an engineer would do when presented with a problem in their calculations. You do know you'll be working in teams, right?

 

If offhand accuracy is 67% of mainhand accuracy, then it is mainhand accuracy times 2/3, ROUNDED UP (so it's not 66.666666666...). And if you're a STEM student and don't just intuitively see that... Well, wow.

 

Where you're going wrong is in this "at soft cap" thing. You seem to be interested in an equation where a=1.0, or 100%, and then subtracting .33 from that. The problem with that is twofold. First, while .33 of 1.0 may be a third of that value, it is NOT a third of any other value of a. Which brings us to the second problem. In that situation, you are treating a as a CONSTANT (1.0), NOT a variable. If you were treating it as a variable, you would multiply it by 2/3 so that it winds up being 67% (rounded up) of ANY value a. And the only way you're gonna differentiate in terms of a and get anything that means anything is if a is a variable. Do you see?

 

'Cuz if you don't, PLEASE consider liberal arts. =P

Edited March 16, 2012 by Meetstick

[Twenynge]

That would be great, except it's wrong.

 

When you state that offhand accuracy is main hand accuracy minus 33%, what you're saying is that offhand accuracy is 33% less than main hand accuracy. That is NOT the same as main hand accuracy minus .33. .33 is one third (rounded down) of 1. You are subtracting a fraction of one.

 

I'm sure I'd be real impressed that you were speccing engineer IRL, if I wasn't speccing physicist. What you're studying to be doesn't make the math right. If it's not your math, stop defending it. If it is your math, look at it and fix it. THAT'S what an engineer would do when presented with a problem in their calculations. You do know you'll be working in teams, right?

 

If offhand accuracy is 67% of mainhand accuracy, then it is mainhand accuracy times 2/3, ROUNDED UP (so it's not 66.666666666...). And if you're a STEM student and don't just intuitively see that... Well, wow.

 

Where you're going wrong is in this "at soft cap" thing. You seem to be interested in an equation where a=1.0, or 100%, and then subtracting .33 from that. The problem with that is twofold. First, while .33 of 1.0 may be a third of that value, it is NOT a third of any other value of a. Which brings us to the second problem. In that situation, you are treating a as a CONSTANT (1.0), NOT a variable. If you were treating it as a variable, you would multiply it by 2/3 so that it winds up being 67% (rounded up) of ANY value a. And the only way you're gonna differentiate in terms of a and get anything that means anything is if a is a variable. Do you see?

 

'Cuz if you don't, PLEASE consider liberal arts. =P

 

Tis you who need to consider liberal arts. He's right.

 

P.S. This is coming from another engineer IRL.

Edited March 16, 2012 by Twenynge

[Tibbel]

That would be great...

 

I don't understand which part you're referring to as incorrect.

 

I'm finding the rate of change in the expected total contribution from weapon damage ©, with respect to change in main-hand accuracy (a). In the equations below, W is defined as:

W = (1 + AmountModifierPercent) * (WeaponDamageMin + WeaponDamageMax)/2,

where AmountModifierPercent is a constant specific to each ability found in the game files, and WeaponDamageMin and WeaponDamageMax are stats from our weapon. (For the following calculations, the main-hand and off-hand pistols are assumed to be equivalent to each other.)

Note that W is independent of a.

 

From the SWTOR Formula List:

(I) Special/Force/Tech Attack Accuracy % = 100% + 30% * ( 1 - ( 1 - ( 0.01 / 0.3 ) )^( ( AccuracyRating / max(Level,20) ) / 0.55 ) )

and

(II) Special/Force/Tech Off-Hand Accuracy % = 67% + 30% * ( 1 - ( 1 - ( 0.01 / 0.3 ) )^( ( AccuracyRating / max(Level,20) ) / 0.55 ) )

Substituting and solving for Off-Hand Accuracy %:

SpecialAttackAccuracy% - 100% = SpecialOffHandAccuracy% - 67%

SpecialAttackAccuracy% - 33% = SpecialOffHandAccuracy%

Which is to say:

SpecialOffHandAccuracy% = a - 33%

 

From the formula list, expected ability damage is:

(III) ExpectedDamage = C(a) + (StdHealth * StdHealthPercent)(a) + (Coefficient * DamageBonus)(a)

Expanding the weapon damage contribution expression, C(a), as a function of main-hand accuracy:

C(a) = W a + 0.3 W (a - 0.33)

C(a) = W a + 0.3 W a - 0.3 W * 0.33

C(a) = 1.3 W a - .099 W

And differentiating:

C'(a) = 1.3 W

Which part of the above are you saying is incorrect?

 

While .33 of 1.0 may be a third of that value, it is NOT a third of any other value of a.

This is the underlying issue with the difference in accuracy scaling between gunslingers and snipers.

 

a * 2/3 is how it should be.

 

a - 0.33 is how it is.

Edited March 16, 2012 by Tibbel

[Threskov]

Sniper do have it better. Even tho the damage is supposed to be the same, the off hand damage of the gunslinger is weapon dmg, and the sniper dmg is added to weapon dmg which add dmg to skills, so more tech dmg = less mitigation.

 

Theory is wrong again. Lets wait for 1.2 and see if something change.

[Tibbel]

Sniper do have it better. Even tho the damage is supposed to be the same, the off hand damage of the gunslinger is weapon dmg, and the sniper dmg is added to weapon dmg which add dmg to skills, so more tech dmg = less mitigation.

 

Theory is wrong again. Lets wait for 1.2 and see if something change.

 

Total Ability Damage = MainHandWeaponDamage + OffHandWeaponDamage + BonusDamage + BaseDamage

 

For snipers, the off-hand weapon damage is 0.

Because a sniper rifle deals 20% more damage than an even-level pistol, the total is:

Total Ability Damage = {1.2 * WeaponDamage + BonusDamage + BaseDamage} * a

 

For gunslingers, OffHandWeaponDamage is multiplied by 0.3 * (a - 0.33).

The total is:

Total Ability Damage = {WeaponDamage + BonusDamage + BaseDamage} * a + {WeaponDamage * 0.3} * (a - 0.33)

 

When a = 1, the two totals are nearly identical.

 

When a >> 1 or a << 1, then that's where the differences are noticeable.

Edited March 16, 2012 by Tibbel

[Sharee]

[Edit] Nevermind.

 

Hopefully we'll be able to solve this issue after we get access to combat logs and target dummies sometime in the future. Until then, have a nice day

Edited March 16, 2012 by Sharee

[Mekias]

Gunslingers look way cooler.

 

End of thread.

[Bartes]

Gunslingers look way cooler.

 

End of thread.

 

tho all look the same...

 

rant on

[MasterApollo]

OK, so from pumping raw numbers into a spreadsheet I have (generated using the swtor formula page everyone links) I came up with some general pointers to others in my guild so I might as well post it here.

 

First, like most are saying here, generally speaking snipers have to "do" less to get their damage down. With the main and off-hand for gunslingers, we rely a whole lot more on accuracy vs anything to get our damage down. If you can get lucky with it, you'll do stupid levels of damage when you hit. This is where, by comparison, a sniper will hit more often yet ultimately has a lower damage ceiling.

 

In a practical sense, it's hard to give an exact on it. If you want you can try and push your accuracy up with either an outright +accuracy rating or +aim (aim tends to be used more on pvp endgame vs accuracy on pve endgame) yet bottom line is you simply won't hit as much as a sniper. I say the distinction on pve and pvp endgame gear as if you want to be fun about it and not max out the useful stat bonuses (essentially where you'll get diminishing returns for pushing up a single stat) then mixing and matching across gear can be the way to go, although this tends to punish your max damage output with removing things like expertise on non-pvp gear.

 

With that in mind, in practical terms, I've generally found that the "best" way to go about things to give yourself more of a push in battle is to focus more on cunning and power, so that at least when you are hitting you'll always hit hard. If you want to mix things a little bit, do it in things you can get 2 of, namely your pistols, implants and relics. You'll always want expertise for the max threshold on your attacks, yet from there it's all about cunning, followed by power, then aim and then anything else is a bonus

[JoyProtocol]

As for the calculations, the off-hand debuff is a -33% accuracy used as a static number, and *0.20 damage used as a multiplicative. It's really not that hard.

 

If you're still confused why 33% is not 1/3, it is because 1/3 would need to be ...of something. In this case, the number does NOT change with anything else, so it is a static 33%. Besides, 1/3 would be 33.333333...%

 

All in all, Smugglers have a marginally greater damage potential, but it is much more chance drive. I like it that way!

"I wasn't planning on living forever anyways"

 

OK, so from pumping raw numbers into a spreadsheet I have (generated using the swtor formula page everyone links) I came up with some general pointers to others in my guild so I might as well post it here.

 

First, like most are saying here, generally speaking snipers have to "do" less to get their damage down. With the main and off-hand for gunslingers, we rely a whole lot more on accuracy vs anything to get our damage down. If you can get lucky with it, you'll do stupid levels of damage when you hit. This is where, by comparison, a sniper will hit more often yet ultimately has a lower damage ceiling.

 

In a practical sense, it's hard to give an exact on it. If you want you can try and push your accuracy up with either an outright +accuracy rating or +aim (aim tends to be used more on pvp endgame vs accuracy on pve endgame) yet bottom line is you simply won't hit as much as a sniper. I say the distinction on pve and pvp endgame gear as if you want to be fun about it and not max out the useful stat bonuses (essentially where you'll get diminishing returns for pushing up a single stat) then mixing and matching across gear can be the way to go, although this tends to punish your max damage output with removing things like expertise on non-pvp gear.

 

With that in mind, in practical terms, I've generally found that the "best" way to go about things to give yourself more of a push in battle is to focus more on cunning and power, so that at least when you are hitting you'll always hit hard. If you want to mix things a little bit, do it in things you can get 2 of, namely your pistols, implants and relics. You'll always want expertise for the max threshold on your attacks, yet from there it's all about cunning, followed by power, then aim and then anything else is a bonus

 

+AIM does not increase your Accuracy. Let's not confuse people further.

Aim increases ranged weapon damage, and ranged weapon crit chance. It does more for Troopers(and BH's), but we're not troopers (or, god forbid, BHs).

 

Expertise is great - IF YOU PVP. Does nothing for any PVE content.

 

As a general stat, for Gunslingers Cunning > Power > Crit > Surge > everything else.

 

I've come to really like accuracy, too

For what it counts, at high 30's I came across a whole lot of accuracy enhancements. I bought them ALL! And stuck them into everything I could. I then had a noticeable increase in my damage dealt, and killing power in both WZs and heroics.

Edited March 16, 2012 by JoyProtocol