The Dark 5 Hit (kinda)

Sword strap IS actually stronger than 3% haste if you have 0 haste.... but why would anyone NOT have any haste?

Making it harder than it needs to be...

But copy and paste is useful:

Exactly the same :x 100/97!

EDIT: Changing to 100/97 cause it's actually a lot neater (and strictly speaking more correct). 3/97 is just shorthand.

Finally some people w/ brains post :)

Sword strap would reduce a 528 delay weapon to 512.16 delay, where adding 3% haste would reduce it to 512.53.
but why would anyone wear 0 haste is beyond me

Am I being dumb or why are you getting 512.53? Both should be delay * 0.97.

Loving the spiteful karma response.

I couldn't remember the rest a few minutes ago, but this part didn't make the alla discussion:

The killshot you steal has to 1. Not be yours already 2. Be the person you overtake and 3. Is a function of your window versus their delay.

If you have three melees in your group, Its then a 66% chance of someone other than yourself having the killshot coming. Assuming all delays are equal to yours, there is a 42% chance of the killshotter's final swing falling into that 1.48 second window, but someone with a faster delay is going to be doing less per-hit damage and is thus balanced out by being less likely to be killshotting, so its more a factor of weapon DPS.

.66 * .50 * .42 = 13.86% chance, which I correlated to that python script from the SAM forum that estimated someone adding 1% haste for a 3% delay reduction at 66% total haste would steal a killshot about 14% of the time after enough simulations.

14% chance versus 24% chance, and the latter is still a full hit gained instead of some stolen parse-padder.

Aside to this, DA can proc on WS which increases the exercising of its probability by 7/6, while also providing an entire additional hit that can increase WS rate fractionally which Sword Strap will never do in real applications.

That's what FFXICalc shows <.<
then again i'm talking about 3% gear haste

I don't quite understand this, even if you DA in the middle of a fight, in the overall scheme of things you're still taking away a hit from someone else in the exact same way as your killshot scenario.

No idea what's going on there <_< unless I am being dumb.

1. Your newly DA'd hit has to be one you wouldn't have DA'd anyway too.

3. If you're using specific delays you're not using a general situation - specific situations can weaken DA as well. General situations tend to produce higher output numbers than in reality but are more reliable.

Make sure the killshot things also calculates in DoT damage taking the kill etc as well. There are far too many factors to even consider <_<

DA procs on WS are generally quite weak on highish fTP WSs.

Because a killshot can be mostly overkill, the mob may have just 1hp left.

A DA in the middle of a fight can get you a whole hit closer to another WS, which makes the mob deader for sure. This benefit also carries to the next mob instead of your delay resetting and a whole extra hit never actually happening.

People talk about 'diminishing returns' like having more than one source reduces all others to zero. Its only a reduction by your existing percentage, while a DRK will likely never have more than 10% DA to start with, even 1.8% (10% diminished, for 12% DA total) over 14 hits is still a 22.4% chance (simply 10% less likely for that 2% alone to give you a DA).

If you're getting Fighter's Roll, you can make the dimishing returns argument.
You forget this is the DRK forum.

At any rate, a DA is gonna be a 4.825/3.825 increase on Guillotine, or enjoy the 50% attack bonus of Spinning Slash (this is partly an assumption, but my 1491 TA proc SS last on a Ro'Maeve doll points to extra attacks indeed being affected by WS Ratio bonus, or it would have been more like 1.0-1.1k).

A DA proc on a WS is simply an extra hit, thats what a 1.0ftp is after all, but it also gets the WSC modifiers and (likely) ratio bonus which gives it a little more 'oomph' on top of more chances to DA.

4.825/3.825 * (chance of DA) is the increase from DA though making it very weak. DRK has among the highest fTP of common WSs with Guillotine anyway.

Ratio bonuses being applied to Double Attack hits is still up in the air, but if that were the case you would expect DA procs to add a lot more damage to SAM WSs than they do.

Speaking of SAM WS, I think pDIF's multiplier is calculated before the ratio bonus is applied. This makes up for that ~4% I've been missing for SAM WS on Kirin (because low cRatio = higher pDIF multiplier) by having a 1.25 max multiplier instead of 1.2.

Anyway, a DA proc on WS is just an extra attack with the WS's bonuses, think of it that way instead of relative to the WS itself. Indeed it gets overshadowed by the WS's damage, but its a whole hit worth of damage like any other DA and then some. The only issue then with WS DA procs is that they don't give that bonus hit's worth of TP.

Interesting. I would like if this was actually solved at some point :(

Well they do give TP :p just not like... full hit TP. I think you meant that anyway.

For the record, I don't disagree that Haste/-Delay's returns are overestimated with math, I just don't think it's by an amount enough to make it worse than other options in the majority of cases.

Its mostly that SAM *** that has certain people going 'BAWWWW NUUUU UR TOTALLY RONG!!'. Its only a matter of time, and seeing has Masa still hasn't delivered I think I have plenty of time anyway.As I put in plainer terms in an edit, over a ridiculously long fight (5min+) they come out pretty even. These fights don't happen though, and so Sword Strap is only ever Swinging Faster instead of Swinging More. Swinging more actually increases your damage and WS rate, while swinging faster just steals hits at the ends of fights and is generally just total parse wankery. Even a percentage chance of swinging more is better than swinging faster.

However, if you're swinging an OAT or OA3 weapon, its Sword Strap or GTFO, even moreso than getting Fighter's Roll.

I dunno, I'm still not sure I quite agree with your analysis of it. Trying to analyze it in terms of interaction with other party members is just... I dunno, just seems like there's too many variables.

A kinda simple example, you have your 3 meleee with your 50~ sec mobs hitting 14 times. If you add sword strap to all 3, mob dies 3% faster. Killshots haven't changed at all since they all are hitting at the new speed. DA unfortunately is much more difficult to model since it can cause additional WS's and what not. In theory though, you either have to have 3 DA's during a between the 3 melees or DA has to lead to an additional WS, or else the fight doesn't actually get any shorter. When the fight does get shorter it's in much larger chunks though, as opposed to the constant 3% of the sword strap. I'm possibly babbling a little bit here, kinda just emptying my head out here. Just doesn't seem as straightforward as you put it.

Also, semi related, but I'm pretty sure even when you start on a new mob, you retain your old delay. Probably a result of the whole engage/disengage bug waaaaaaaaaaaay back.

Sword Strap can cause additional WSs - land an extra hit relative to what you would have got and you have extra TP from that hit!

No, it doesn't. Thats the point I made about only swinging faster instead of actually swinging more. It takes at least 33 melee rounds of time without Sword strap for it to have gained you an extra 34th round with it. Even at 66% haste this is 98 seconds. Though its only a chance of happening, DA always gives you this extra round when it works.

Think of the 50 second fight example above: You can't land '0.42 of a hit'. All or nothing is what breaks Sword Strap in all but overly long fights, and even in those Pole is equivalent.

You delay resets when your target dies, but not when you switch targets. You want to autotarget the next mob though because it resets your delay.

Hmm, pretty sure even autotargetting doesn't cause delay reset but I'll have to wait til I get home to test it out.

Reset as in return delay to counting down from 513 or whatever it is. Changing target doesn't do that. That said sword strap would let you engage another mob slightly faster but no doubt less than half a second faster.

This is in a vacuum though - if you hit once more with Sword Strap when you wouldn't have, you gain an extra hit and extra TP - period. The method is somewhat irrelevant in this context.

You might not have been around when the disengage/reengage trick was running rampant. You could trigger the target-change-delay-reset by having a macro with:

/attack off
/attack off (forget why twice, probably just to take up room)
/wait 1
/attack on
/wait 1
/attack on

and pressing it just as you draw your weapon.

Delay resets to a basic 60 (one second) whenever you autotarget or freshly engage a mob. This macro wankery made you disengage and reengage before the 'timeout' before being able to engage again kicked in (thats why its so aggressive now, after it was patched), which made two-handers swing one second after each time they put their weapon away and pulled it out again.

I'm pretty familiar with it, I was using it to skill up with GS at one point, lol. The patch on this thing is why I think they made retargetting delays more correct. But maybe I'm crazy and confusing myself, gonna just check once I'm home to confirm it to myself.

You'll also have to consider that Engaged status is actually client side, so there is lag added in (likely twice, announcing your engaged status and then being notified of your first swing). You can use a memory hack to eliminate the reengage timeout and still do the disengage/reengage trick if you're so bold.

The best place to see it though is the Seed Mandragora fight. Turn /autotarget on, put your back to the wall, and watch as you swing to kill each mandy a little more than one (probably closer to two, see below) second apart. Switch to a faster weapon that still kills them in one hit and you still kill them at the same speed.

Thinking about that in particular, its probably two seconds now, 120 delay, which could be part of the reengage fix.

And? Kill happens sooner because you were using Sword Strap, boom, advantage gained. Didn't? No advantage. On average? Actual value approaches 3.09% over time.

Every single offensive stat in the game has some form of overflow or loss associated with it. Using that as an argument against any of them without some very precise math (read: not the above) to back it up is straight up *** HELP I AM TRAPPED IN 2006 PLEASE SEND A TIME MACHINE.

Or as Failure put it, stop arguing against marginal stat increases.

Do continue ignoring everything I just said.

Do continue acting like you were acknowledging the law of averages rather than acting like it was largely lossy in many situations.

The truth is, you can just about model a single fight as long as you know:

* DD gear/merits/stats
* Buffs on DD
* Food
* Debuffs on Mobs
* The exact stats of the mob (the hardest to get and one of the most important for an accurate model)

Given all that info it is possible to work out if X increase of Y stat will or won't increase kill speed.

But if anything in the above list changes then all the calculations need to be redone.

So the only way to work out things like this is using the basis of average improvements and projected improvements over time.

These averages/projections will inherently include a degree of error. As with this example, short kill times will probably favour Pole over Sword. Whereas in high DA situations (i.e. WAR main + merits + fighters roll) will probably favour Sword over Pole.

In a vacuum sword = 3.09% increase, and Pole is some where between 1.5 and 2.0% (depending on other gear).

Some people are happy to just use the average/projected values (above), but as this thread shows some people want look deeper. But when you are looking deeper you have too look both ways. Because in all honesty neither of the above items will give the rated percentage increase, as vacuum figures are the maximum an item can give. One stat may be 'devalued' more in a certain situation than an stat, and whether this is enough to make item X better than item Y (when the vacuum maths shows Y > X) is very hard to say.

I would say in a situation where you only do a few swings per mob then maybe pole would be better, but this is a pure guess as there is no conceivable way to build a general case model accurate enough to know for sure.

I don't follow how short kill times have an effect significant enough to favor Pole. If you really want to get into increases in damage (ie parse epeening) rather than overall performance (simply killing faster on average), bear in mind that DA procs have a greater chance to force you to hold TP unnecessarily on mobs with a small HP pool unless you're willing to overkill for the sake of the parse, at which point our interests diverge and this discussion is no longer worthwhile because we're looking at two different things. Otherwise, they have similarly lossy properties in low-HP situations.

Yeah, in a capped acc situation and even then not if you have dusk +1 hands. Otherwise Turban/E Body/Blitzer/Homam is 10STR/4DEX/10ATT verses 14ACC...

That's like saying Ares Body beats E Body. Yeah, in a capped acc situation.