TWF vs. THF. Again.

[cforce]

Why does math show that TWF does more DPS than THF, but in-game experience doesn't?

Cop-out answer 1: The calculations are so complex that the math can't possibly be right.

Cop-out answer 2: In-game observations are always subjective, and people see what they want to see.

While both of these cop-out answers tend to have some kernel of truth, I suspect that the math isn't as far off as some people think, nor are game observations as colored as other people think. I think there is a better answer.

Glancing Blows hit multiple targets.

Well, duh! Tell us something we don't already know! But, the fact of the matter is, most modeling DPS calculations are only only calculating damage vs. a single target, but most empirical observations come from some combination of "how fast stuff dies" and "looking at the kill count", both of which give full weight to multiple-target damage (and tend to obscure points where there is only one "boss" remaining and the whole party is attacking it).

However, it was not until actually playing with some numbers that I realized how big this gap might be.

For the purposes of this thought experiment, I took two WF Frenzied Berserker 20's, a THF'er and a TWF'er, assuming the same equipment for both (TWF had Min II Khopeshes, THF had Min II Falchion), and damage-optimal feat and enhancement choices. Against a *single target*, the TWF Barb should be generating more DPS. However, against two or more targets, the THF Barb generates more DPS. That's right -- adding only a second target allows the heavy damage a FB III gets on glancing blows to catchup and pass the TWF.

Moreover, in cases with a good thick cluster of mobs, the THF Barb generates a *lot* more DPS. Against 5 targets, I came up with the THF Barb generating 62%(!) more DPS than the TWF barb. For the benefit of those who don't frequently run the numbers -- trust me, this is a *huge* gap. In fact, the THF Barb needs to only maintain an average of attacking 1.5 mobs to equal the TWF DPS output.

Now, to cut off some objections:

Q: But cforce, no-one knows the exact percentage of glancing blow effects. Doesn't this make your analysis invalid?
A: Nah, it doesn't turn out to make that big of a difference, since the base damage and vicious components are so much larger than the effects. I assumed a 30% chance on the base effect, but even going as high as a 60% chance only changes the glancing blow damage by about 5%. Qualitatively, the observations are still the same: less at one target, more at two targets, a *lot* more at five targets.

Q: Doesn't the THF-Twitch technique give you a lot more attacks?
A: I assumed a Madstoned THF Barbarian using twitch attacking can get 133-134 swings/minute, (or a hasted one got 126-127) but I don't have any video confirmation on these numbers yet. If someone can demonstrate a faster speed, by all means post a vid. For THF to be attacking fast enough to equal TWF on *one* target, it would need a madstoned-attack rate of about 144/minute.

Q: What if I don't use "twitch" THF technique?
A: Well, you're a lot further behind on a single target. I mean, a *lot* -- to the tune of 25%. But, here's the good news: you're still a little ahead of TWF on two targets, and you're *much* further ahead on 5 targets -- close to *double* the TWF rate. Regular THF fighting does get a lot more glancing blows than twitch THF in the new combat system, so the "twitch" method is only optimal against one or two targets, now.

Q: What if I'm not a FB?
A: Well, the same basic principle applies, although the upside is a lot lower. Still, multiple targets goes to the THF Kensai over a TWF Kensai as well -- just not by as much of a margin. This should come as a shock to no-one: THF FB is *way* ahead of the pack on multi-target damage.

OK, I've tried to lay this out as reasonably as possible, which can only mean: it's time for irrational the holy war to begin!!! Flame ON!

[Monkey_Archer]

Cop-out answer # 2....

The "big" lead that TWF has over THF in calculations vs in-game testimonial, IMO, is due to the "max gear/max buff" assumptions that we make.

In game, level 20 warchanters + double TOD sets, etc... are not very common, and benefit TWF more.... as do weapon effects (which many mobs are immune) and sneak attack damage (which is never 100% ingame)

When you lessen these factors THF and TWF single target dps is alot closer then most people think

[krud]

Just curious. Is there a twitch technique for the other combat styles, such as s&b? Although "shadowhumping' to get more attacks strikes me as lame combat mechanics, it's even more lame if it were limited to only one combat style.

[Borror0]

Question: Are you assuming or not assuming Inspire Courage?

One thing to keep in mind is how more important single-target DPS is than AoE DPS, typically, both in terms of resource conservation and content.

[cforce]

Just curious. Is there a twitch technique for the other combat styles, such as s&b? Although "shadowhumping' to get more attacks strikes me as lame combat mechanics, it's even more lame if it were limited to only one combat style.

Technically, yes -- but the gain for every other style is fairly marginal compared to the gain for THF. Since it gives so much less benefit, no-one generally bothers learning the timing pattern for anything else.

(Although I'd be curious to see if this is *still* true post-Mod-9; I don't think the above has been confirmed in the new system.)

[cforce]

Question: Are you assuming or not assuming Inspire Courage?

One thing to keep in mind is how more important single-target DPS is than AoE DPS, typically, both in terms of resource conservation and content.

Yes, assuming Inspire Courage. I essentially assumed every possible end-game DPS benefit *except* Yugoloth pots and sneak attack damage.

[Gunga]

Hey Cforce...been a long time. I hope you've been well.

I'll admit to getting a bit dizzy when going through some of your dps calculations, but do you add the glancing blows into the calculation for THF? EVERY OTHER hit I make has a glancing blow number under the main hit. Have you accounted for this in your THF calcs?

[Cedwin]

Is this assuming particular builds? GTWF vs GTHF? Capstones? PrEs?

[cforce]

Hey Cforce...been a long time. I hope you've been well.

I'll admit to getting a bit dizzy when going through some of your dps calculations, but do you add the glancing blows into the calculation for THF? EVERY OTHER hit I make has a glancing blow number under the main hit. Have you accounted for this in your THF calcs?

Yes -- I'm using an extension of Absolute_Omniscience's DPS calculator (to include some of my attack speed calcs as well). Glancing Blows on every other hit (or 3 out of 4 when not twitching) are in there.

[cforce]

Is this assuming particular builds? GTWF vs GTHF? Capstones? PrEs?

Yes, yes (Barb capstone), and yes (FB III).

[WolfSpirit]

I thought they changed the whole attack chain/speed just because of/to correct the twitch THF technique. Is that not so?

[BlackSteel]

what str did u you use? 64-66 is a fairly consistent maintainable number for end game geared

[cforce]

Cop-out answer # 2....

The "big" lead that TWF has over THF in calculations vs in-game testimonial, IMO, is due to the "max gear/max buff" assumptions that we make.

In game, level 20 warchanters + double TOD sets, etc... are not very common, and benefit TWF more.... as do weapon effects (which many mobs are immune) and sneak attack damage (which is never 100% ingame)

When you lessen these factors THF and TWF single target dps is alot closer then most people think

Also an excellent point. Taking away all but the most basic gear (+5 Holy/PG, nothing else that's more than 100k plat on the AH, no raid gear) and Bard help, I see that the THF-Twitch Barb 20 can outrun the TWF Barb 20 against a single target when stripped down to bare class/PrE damage.

[Monkey_Archer]

I thought they changed the whole attack chain/speed just because of/to correct the twitch THF technique. Is that not so?

They only altered moving attack speed to be less/equal with autoattack. As a result, twitch speeds have been slowed down slightly.

[Gunga]

what str did u you use? 64-66 is a fairly consistent maintainable number for end game geared

I'm going for 80. Need a couple more pieces of gear.

[cforce]

what str did u you use? 64-66 is a fairly consistent maintainable number for end game geared

I used "only" 62 for the Barbs -- no Yugoloth pot is the difference, presumably. (I also used only a single-madstoned, not double.) The difference between 62 and 66 doesn't significantly alter the outcomes, though -- the STR adds a tiny bit more to the THF case, but it's pretty marginal.

[cforce]

I thought they changed the whole attack chain/speed just because of/to correct the twitch THF technique. Is that not so?

It is not so. THF twitch is the only thing that didn't get its speed *raised*, but it is still relatively faster than THF straight-up. The patch notes made it sound like they'd "fixed" this, but they did not.

[Thorzian]

its even simpler than that. the 2wf math pulls ahead in dps onlt through multiple runs through the combat sequence due to extra attacks. that's why in practice it doesnt work. the mobs are dead long before that combat sequence catches up. the only exceptions are raid bosses. there the 2wf guys, according to the math, do in fact dish out more damage.

again, in practice though, my experience is that my greataxe users steal and hold aggro more often than my khopesh or dwarf axe users. i'll be able to comment on a falchion user soon. i believe that this is only due to the burst damage however. the 500 hp shot to the back gets their attention even though the other guys are doing more damage overall.

[cforce]

I'm going for 80. Need a couple more pieces of gear.

For completeness, I just re-ran with 80 STR assumed, and the basic relationship is the same -- uber-equipped TWF Barb ahead vs. single opponent, THF ahead at two opponents and above.

[cforce]

again, in practice though, my experience is that my greataxe users steal and hold aggro more often than my khopesh or dwarf axe users. i'll be able to comment on a falchion user soon. i believe that this is only due to the burst damage however. the 500 hp shot to the back gets their attention even though the other guys are doing more damage overall.

I'm intrigued by this, and would like to hear more. I've wondered for a while if aggro might be time-weighted -- it's not "what have you done to me", but "what have you done to me lately".

This could also be explained by M_A's "few TWF'ers have every single piece of raid loot and optimal buffs" -- are you one of these people?