|
For the Shinies! A Guide for Thief
VIP
Server: Fenrir
Game: FFXI
Posts: 764
By Fenrir.Motenten 2014-08-22 13:42:42
Ok, Thf spreadsheet updated to give 2x attack on the first hit of Exenterator.
VIP
Server: Fenrir
Game: FFXI
Posts: 764
By Fenrir.Motenten 2014-08-22 14:53:24
All right, another addition to take a look at.
I'm trying to work out some way to give an idea on the variance of possible damage output on weaponskills, to indicate their general stability. Unfortunately, it's rather difficult to construct something like that in Excel. Still, I came up with something that should at least give some idea.
I added a dropdown list at the bottom of the setup page for 'WS Comparison'. It has values of Baseline, Crit, Multi, and Full. The purpose of this is to try to find a way to gauge how much certain types of spike damage influence the overall average damage for a weaponskill.
Full is the default, and everything applies normally.
Multi allows multi-attacks (DA/TA/QA) to count, but not crits.
Crit allows crits to count, but not multi-attacks.
Baseline doesn't allow crits or multi-attacks to count.
Note that guaranteed crits (SA/TA) and multi-attacks (Assassin's Charge) aren't blocked by the above filters.
By comparing these filters, you can see how much of the overall average is generated by the random events, and how much is coming from more consistent sources.
Example:
Evisceration:
Baseline = 2239
Crit = 3664
Multi = 2367
Full = 3874
In that, Multi has a small influence, while crits have a huge influence. As such, you would (naturally) expect to see pretty wide swings in damage output, from the low end to the high end.
Exenterator:
Baseline = 2792
Crit = 2792
Multi = 3009
Full = 3009
Obviously crits have no effect. Multi has a larger effect here than it does on Evisceration (since Evisceration will hit the 8-hit limit more easily), with +7.8% vs Evisceration's +5.7%.
Unfortunately, this doesn't distinguish between a frequent moderate difference, and an infrequent huge difference. The baseline on Evisceration is pretty low, but how often will it be down at that point? I haven't come up with a good way of quantifying such things yet.
[+]
Quetzalcoatl.Wakmidget
Server: Quetzalcoatl
Game: FFXI
Posts: 1388
By Quetzalcoatl.Wakmidget 2014-08-22 15:39:46
How does Dancing Edge compare to Exenerator on the spreadsheet now that the x2 attack on the first hit is implemented?
VIP
Server: Fenrir
Game: FFXI
Posts: 764
By Fenrir.Motenten 2014-08-22 15:57:21
Quetzalcoatl.Wakmidget said: »How does Dancing Edge compare to Exenerator on the spreadsheet now that the x2 attack on the first hit is implemented?
Rough guess, but it looks like Exenterator catches up with Dancing Edge below about 1.6 cRatio, and does better as cRatio continues to drop. (This is also at 5/5 merits; at 4/5 that drops to 1.4 cRatio, etc.)
However, since most of the time that you'd consider using either of those is going to be at higher cRatios, I expect DE to be the go-to weaponskill for consistent damage, and Exen just for the debuff. At capped attack, DE should be about 9% ahead.
VIP
Server: Fenrir
Game: FFXI
Posts: 764
By Fenrir.Motenten 2014-08-22 16:23:22
Also, found a mistake in the multi-attack filtering. New version corrects it.
VIP
Server: Fenrir
Game: FFXI
Posts: 764
By Fenrir.Motenten 2014-08-22 18:07:43
So, given the discussion, and the extra spreadsheet information, I looked into how I'd build a Dancing Edge weaponskill set. If using DE, I'm going for consistent damage over spikes, but will still take spikes if available. This led to me making a couple changes compared to what I'd normally use by default.
If attack isn't capped, this is what I'd aim for:
ItemSet 327519
And if attack is capped, this is what I'd aim for:
ItemSet 327520
Of note for the uncapped set: Cho'j Band instead of Epona's, and Vespid Mantle instead of Atheling.
For the capped attack set, Kayapa and Canny would both be preferred to Atheling.
Plunderer's Vest +1 is best just because of raw stats (even though crits don't matter for DE).
If attack is uncapped, Felistris's attack puts it on par with Uk'uxkaj even before the TA, so prefer that overall. If attack is capped, Uk'uxkaj seems a better choice (particularly if aug'd with dex).
Str-aug'd Quiahuiz Trousers are pretty much tied with r15 Manibozho Brais in uncapped attack scenarios.
Comparing with Wakmidget's:
ItemSet 327511
Love Torque vs Gorget is a wash if att is uncapped. Both are slightly behind Justiciar's in an uncapped situation, but the Gorget would be much better if attack is capped. Overall, I'd go with the Gorget.
Tati+Jupiter Pearl is a better config than Bladeborn/Steelflash if att is uncapped. If att is capped, Bladeborn/Steelflash has a notably higher average, and the amount of multi-attack is high enough to be worth using.
I'd go with Cho'j over Ramuh+1 if att is uncapped. Ramuh+1 wins if att is capped, but Epona's multi-attack rate is high enough that I'd lean towards that instead (and it has a fairly decent edge in average damage). Still, Ramuh+1 isn't a bad choice.
Wanion belt works if att is uncapped (though I'd take Caudata over Wanion), but elemental belt seems better with capped attack. Results are close enough that I'd probably consider the elemental belt the preferred choice here.
If att is capped, Plunderer+1 legs. Manibozho/Quiahuiz legs are a bit ahead of them if attack is uncapped, but it's close enough it should be fine to always use Plunderer+1 (plus the bonus of being the primary choice if you need acc). Pillager+1 also beats Mani if att is capped.
Re-reviewing the Exenterator set in more detail, I'm now able to build it up much closer to the Dancing Edge set. Compared to the set you posted, I'd change neck to a gorget, and head to Lithelimb Cap.
With near-optimal sets in a capped attack situation, DE is now only 5% ahead of Exenterator, and 5/5 Exenterator catches up at about 1.75 cRatio (instead of the 1.6 I found in my first estimate).
Anyway, that covers the 'consistent' weaponskills. I'll play around with comparing vs the others later.
Quetzalcoatl.Wakmidget
Server: Quetzalcoatl
Game: FFXI
Posts: 1388
By Quetzalcoatl.Wakmidget 2014-08-22 18:34:29
Dread jupon should beat pluderer for both de and exen I noticed when looking at bodies
Quetzalcoatl.Wakmidget
Server: Quetzalcoatl
Game: FFXI
Posts: 1388
By Quetzalcoatl.Wakmidget 2014-08-22 18:42:08
Also agi augmented legs from bee wkr should be better for exen as well for capped attack
VIP
Server: Fenrir
Game: FFXI
Posts: 764
By Fenrir.Motenten 2014-08-22 19:02:01
Now, how do we do these comparisons?
Well, I'm going to take a standard optimized Evisceration build. Putting this on Dynamis mob for capped attack, no buffs, and 0 over-TP rounds. I'll also use Mandau/Izhiikoh for the sake of comparison that's been brought up in the thread.
Interestingly, now that I optimized the DE build, Evisceration is only barely ahead of it (about 5%, 4882 to 4632).
Ramuh+1 comes out ahead of Epona's (4882 to 4876), so using that.
I'm using dex-aug'd Uk'uxkaj Caps on both builds.
OK, the basics are set. Now, first I'll look at the average number of hits landed with Evisceration -- 6.35
Now I want to look at the effect of crits on the final damage value. To do that, I'll adjust the crit rate in steps of 1/6.35 (15.75%), simulating 1 crit hit per weaponskill per step, and look at the average damage. Alongside that, I'll put the approximate percent chance of landing that many crits, given a 48.66% crit rate (base, gear, ws, and TP total for this test).
Code
Crits Dmg Chance
0 3718 2%
1 4095 9%
2 4472 23%
3 4848 31%
4 5225 23%
5 5602 9%
6 5979 2%
6.35 6111 ~
This is being compared with a Dancing Edge build that averages 4632 damage. Looking at the chart we can see that DE will average more damage if you only get 0, 1 or 2 crits, while Evisceration will average more damage if you get 3 or more crits. 0, 1 and 2 crits happens 34% of the time. So, 1/3 of your Eviscerations will do worse than DE, while 2/3 will do better.
The question is, what's your minimum weaponskill damage threshhold? If you don't specifically care how much damage is done (eg: fighting an NM where you're damaging the same mob over and over), go for the highest long-term average. However if you're doing something like Dynamis, where a particular minimum is useful (say 4500 damage, half the mob's health), but the high maximums aren't (you often won't weaponskill before the mob is half dead), then the consistency of the DE damage will be more beneficial. Plus, the higher peaks of Evisceration may never be reached, further reducing its average.
Let me know thoughts on this. I didn't spend as much time on it, so there may be considerations I've missed.
Quetzalcoatl.Valli
Server: Quetzalcoatl
Game: FFXI
Posts: 1420
By Quetzalcoatl.Valli 2014-08-22 20:36:46
Let me know thoughts on this. I didn't spend as much time on it, so there may be considerations I've missed.
God damn it, I had a post and lost it u_u;; I'll try and remember what I was thinking.
DC dynamis doesn't cap without food though?
Are we talking at 1000% specifically?
Whats the base crit rate on Evis?
Nefarious + rancor + plunderer+1 uk cap ddex cap?
Seems the crit rate is a little low?
Crit rate varies with tp 50% dex vs
Acc rate varies with tp 50% dex 50% chr (though zero gear intentionally added for chr+)
What set did you use to get 4632 DE? mithra dynamis DC no food no buffs.
Note here at the bottom, I set over-tp rounds to zero, it still says "average tp" I set it to .5 and still "average tp"
It's the amount of tp actually used on the ws, that's why there's ~ 250 extra for the evis, due to moonshade.
VIP
Server: Fenrir
Game: FFXI
Posts: 764
By Fenrir.Motenten 2014-08-23 02:45:53
I made a mistake and forgot to put the Moonshade Earring back on when reconfiguring the gearsets, so Evisceration's crit rate was a little low. It'll shift the distribution a bit (primarily the chance of a given number of crits happening; the damage at each step would only change a tiny bit, from the loss of 2% DA), but the overall idea remains the same.
DE set:
Mandau 119
Izhiikoh
Raider
Uk'uxkaj (dex)
Gorget
Bladeborn
Steelflash
Plunderer +1
Pillager +1
Rajas
Ramuh +1
Kayapa
Ele.Belt
Plunderer +1
Plunderer +1
And I typo'd. It's 4623, not 4632. Also, Taru thf/dnc with RCBs (rather than no food), capped attack (which I specifically noted, in order to use this gear config). Could also use Bison Steak (which is what I normally use for Dynamis; RCBs was from when I was looking at a Salvage idea), just losing a small bit from an fStr drop, giving 4590 on DE.
Target TP was 1000, with 0 over-TP rounds, but average TP is about 1083.
And yes, Dread Jupon would be another improvement over Plunderer's Vest +1.
Base crit rate on Evisceration is 10%, which is added to base, dDex and gear at 1000 TP. That goes up to 30% at 2000 TP, and 50% at 3000 TP. Base, merits and capped dDex add 25%. An additional 12% comes from gear (in this setup).
For Evis (proper earrings):
Mandau 119
Izhiikoh
Raider
Uk'uxkaj (dex)
Rancor Max
Brutal
Moonshade TP+
Plunderer +1
Pillager +1
Rajas
Ramuh +1
Kayapa
Ele.Belt
Pillager +1
Plunderer +1
Readjusting, using Bison Steak, and swapping DE to Dread Jupon body, and Evis to Artful+1 belt, the overall averages are 4992 Evis, 4638 DE (7.5% lead for Evis). 53.64% crit rate for Evisceration, 6.33 hits per weaponskill on average.
I also need to fix the probability distribution. Adjusting to the 53.64% crit rate.
Code
Crits Dmg Chance
0 3697 1%
1 4077 7%
2 4457 20%
3 4838 31%
4 5218 27%
5 5598 12%
6 5979 2%
6.35 6112 ~
Now DE averages better than Evis 28% of the time. Of course this is also skewed because you'll have more chances to crit on a multi-attack proc, where you have 7 or 8 swings, but the math gets more complicated and isn't really worth getting into.
Overall, very close to 1000 TP, DE will win about 1/4 to 1/3 of the time, and may be useful if you need a specific minimum damage but not a very high maximum.
By 1400 TP, overall crit rate reaches about 60% for Evisceration, and DE is only winning about 18% of the time (~1/6). But by that point you also want to start transitioning to Rudra's Storm. Will deal with that in the next post.
VIP
Server: Fenrir
Game: FFXI
Posts: 764
By Fenrir.Motenten 2014-08-23 03:39:54
Quick note: My assertion of Bison Steak also capping attack in the above post is slightly wrong. It caps attack for the main hand, but not the offhand. This leads to a very slight shift in the 'best' gear for a slot or two. I'm not going to bother fiddling with that right now, as I already dealt with the issue in an earlier post.
Considering Rudra's Storm in the consistency vs spike damage (for non-stacked weaponskills) is actually fairly simple.
5% of the time, damage will suck. No avoiding that. The other weaponskills only take a small hit from a missed swing, but for Rudra's, missing the first hit completely kills it.
Aside from that, your damage variance is going to be similar to Dancing Edge's. There are no crits to skew things, so it's just the standard pDif variance with a vaguely similar fTP total, though Dancing Edge will be more bell curve shaped (and thus more concentrated on its mean) over its range, while Rudra's will be more flat.
Now, Rudra's will have an overall average that matches Evisceration's at about 1300-1350 TP. However, like Dancing Edge, it's going to have a higher average than a good chunk of Evisceration's weaponskills at lower TP values.
The chart we used before still holds for damage per crit step. So we can use that to see at which TPs Rudra's will outdo N number of crits from Evisceration.
At 1090 TP (the lowest I can get the spreadsheet to go, and represents what you actually see when you hit your ws macro), it matches/beats 2 Evisceration crits. At 1225 TP, it edges ahead of 3 Evisceration crits. At 1400 TP it passes 4 Evisceration crits.
Also of note, at a mere 1150 TP it beats the average Dancing Edge weaponskill. If not for the risk of missing the first hit, it would be a perfectly adequate weaponskill to use in place of DE in Dynamis, given that you'll often be weaponskilling above that TP level.
Quetzalcoatl.Valli
Server: Quetzalcoatl
Game: FFXI
Posts: 1420
By Quetzalcoatl.Valli 2014-08-23 04:59:28
You know I'd never argue with you, thoroughly appreciate everything you do everything you do montenen~ just want to help out
Couple changes:
Taru RCB dynamis DC
Set 1 Evis @ 1000
Set 2 DE @ 1000 (aisance +1 is manually added in)
Set 3 Rudra @ 1000
Set 4 Rudra @ 1050
Being Mithra gets you 60 more damage on DEX ws, no effect on DE (CHR of a doorknob)~ MITHRA MASTER RACE!
WS Dmg: 5096
Quetzalcoatl.Valli
Server: Quetzalcoatl
Game: FFXI
Posts: 1420
By Quetzalcoatl.Valli 2014-08-23 05:45:18
Interesting side note, changing your tp gearset, will affect the avg TP for the ws, thereby (artificially?) changing it (VERY minor changes).
I made a few alterations to the TP set, and it changed the damage on evisceration
(zero over tp rounds)
So by using asperity necklace, i get a lower Melee dps, but a higher overall, as it assumes lower average tp at the time of ws.
By using nefarious collar, I get a high melee dps, but a lower overall, as it assumes a higher average tp at the time of ws.
((((mindblowing))))
The addition of that Average tp @ time of ws setting almost completely changes best sets, granted by a severely minor amount, only a bot could take advantage of, but yeah.
Quetzalcoatl.Wakmidget
Server: Quetzalcoatl
Game: FFXI
Posts: 1388
By Quetzalcoatl.Wakmidget 2014-08-23 10:47:18
Made a few tweaks to Dancing Edge and Exenerator sets:
ItemSet 327490
Didn't even notice the AGI on lithelimb cap. Thank you, I missed that completely. Swapped to gorget as well, was 50/50 on that one since AGI is an 85% mod. Other than dread jupon and kaabnax trousers with +8 AGI augment, I can't think of anything else off the top of my head for Exenerator atm.
ItemSet 327511
I was toying with gorget but since I hadn't used it in awhile I was just putting on what I had on-hand. I farmed a soil gorget and soil belt last night. Honestly, kind of surprised about Soil Belt on capped attack, 8 DEX it still a good chunk, but I guess 40% DEX is kind of meager.
Goona keep Tati Earring/Jupiter's Pearl just for ease of gearing for now and inventory space, plus it gives me a little more variance in attack if I use it in something a little more than Dynamis.
Forgot about Manibozho brais as well, which will work fine for now, and again gives me a bit of variance again.
EDIT:
Just tried a Dynamis-Tavnazia with these 2 sets, I was kinda using Dancing Edge and Exenerator interchangeably. Maybe it's because of Mote's new comparison that that aren't that far part anyway,and maybe I was just getting lucky on Exenerator and unlucky on Dancing Edge, but just my personal experience that run, Exenerator felt a bit more reliable than Dancing Edge did. It would still do what I needed it to whenever I used it. Dancing Edge would occasionally leave me having to melee down about 20% HP or so.
I'll play around with both of them a bit more the next few weeks though.
Quetzalcoatl.Wakmidget
Server: Quetzalcoatl
Game: FFXI
Posts: 1388
By Quetzalcoatl.Wakmidget 2014-08-24 08:39:43
Found one more option for Exenerator, Thurandaut gloves +1 or iuitl wristbands +1 with +2% Double Attack augment should beat pillager's armlets +1.
Haven't seen anything else for Dancing Edge other than what Mote already mentioned.
Ragnarok.Punisha
Server: Ragnarok
Game: FFXI
Posts: 34
By Ragnarok.Punisha 2014-08-26 14:25:36
Is Thaumas Coat better to TP in than Plunderer's Vest +1? What if I have 5/5 ambush?
Quetzalcoatl.Wakmidget
Server: Quetzalcoatl
Game: FFXI
Posts: 1388
By Quetzalcoatl.Wakmidget 2014-08-27 11:53:51
With the introduction of +9 stat rings, how does rajas ring compare to Ramuh ring +1 on Rudra's Storm assuming fSTR is uncapped? 4 DEX versus 5 STR.
SA Stacked?
TA Stacked?
Unstacked?
Leviathan.Xsoahc
Server: Leviathan
Game: FFXI
Posts: 107
By Leviathan.Xsoahc 2014-08-27 12:13:04
I +1 the feet for tp (over espial, although I have the chapters for plunderer's +1 now), but I've been contemplating whether or not to use my KI to +1 the hands now or wait until I get the head piece over getting chapter 6s.
The only reason I could see in doing so is the more AGI for exenerator. (Plunderer's armlets +1 vs iuitl wristbands +1). Even now I use the NQ for it.
Decisions.
Quetzalcoatl.Valli
Server: Quetzalcoatl
Game: FFXI
Posts: 1420
By Quetzalcoatl.Valli 2014-08-27 12:56:02
Eviscerate, don't exenterator. That'll save you all agi gear options.
(and be better. haven't used exenterator since the ws update other than to test that it's lacking now).
I always got ramuh being best for Rudra's SA/TA/unstacked. (40 damage roughly) ((consider buying them, if you are, mainly for the acc in addition to the dex, they help a lot on high level stuff)).
By Debodopimp 2014-08-27 13:56:31
Stop sucking at Thief.
Make Varja.
Win.
Quetzalcoatl.Wakmidget
Server: Quetzalcoatl
Game: FFXI
Posts: 1388
By Quetzalcoatl.Wakmidget 2014-08-27 14:03:13
Stop sucking at Thief.
Make Varja.
Win.
Vajra
Stop sucking at literacy.
Quetzalcoatl.Wakmidget
Server: Quetzalcoatl
Game: FFXI
Posts: 1388
By Quetzalcoatl.Wakmidget 2014-08-27 15:01:54
Now, how do we do these comparisons?
Well, I'm going to take a standard optimized Evisceration build. Putting this on Dynamis mob for capped attack, no buffs, and 0 over-TP rounds. I'll also use Mandau/Izhiikoh for the sake of comparison that's been brought up in the thread.
Interestingly, now that I optimized the DE build, Evisceration is only barely ahead of it (about 5%, 4882 to 4632).
Ramuh+1 comes out ahead of Epona's (4882 to 4876), so using that.
I'm using dex-aug'd Uk'uxkaj Caps on both builds.
OK, the basics are set. Now, first I'll look at the average number of hits landed with Evisceration -- 6.35
Now I want to look at the effect of crits on the final damage value. To do that, I'll adjust the crit rate in steps of 1/6.35 (15.75%), simulating 1 crit hit per weaponskill per step, and look at the average damage. Alongside that, I'll put the approximate percent chance of landing that many crits, given a 48.66% crit rate (base, gear, ws, and TP total for this test).
Code
Crits Dmg Chance
0 3718 2%
1 4095 9%
2 4472 23%
3 4848 31%
4 5225 23%
5 5602 9%
6 5979 2%
6.35 6111 ~
This is being compared with a Dancing Edge build that averages 4632 damage. Looking at the chart we can see that DE will average more damage if you only get 0, 1 or 2 crits, while Evisceration will average more damage if you get 3 or more crits. 0, 1 and 2 crits happens 34% of the time. So, 1/3 of your Eviscerations will do worse than DE, while 2/3 will do better.
The question is, what's your minimum weaponskill damage threshhold? If you don't specifically care how much damage is done (eg: fighting an NM where you're damaging the same mob over and over), go for the highest long-term average. However if you're doing something like Dynamis, where a particular minimum is useful (say 4500 damage, half the mob's health), but the high maximums aren't (you often won't weaponskill before the mob is half dead), then the consistency of the DE damage will be more beneficial. Plus, the higher peaks of Evisceration may never be reached, further reducing its average.
Let me know thoughts on this. I didn't spend as much time on it, so there may be considerations I've missed.
Did some more messing around with combinations of stats on Dancing Edge, currently what I'm using atm is:
ItemSet 327511
Upgrades I'm looking to make in the future would be changing out Jupiter's pearl with trux earring.
ItemSet 327490
For Exenerator the only change I can see would be changing pillager's armlets +1 to iuitl wristbands +1 with a +2% Double Attack augment.
In-game experience atm, Dancing Edge and Exenerator are are pretty close, but as Mote posted they are only about 5% apart, doesn't seem like there is much to eyeball.
I'm kind of curious though, assuming all the changes I mentioned in place of what I'm using now, how close are they on paper assuming capped attack(which is what I am building them for).
Still using unstacked Evisceration on things that are uncapped attack.
And thank you Mote for some very helpful suggestions as to gearing options I missed earlier.
Ragnarok.Punisha
Server: Ragnarok
Game: FFXI
Posts: 34
By Ragnarok.Punisha 2014-09-02 22:57:55
So what exactly are the 'best' setups for evisceration and exenterator? a lot of people have their own and they're different from the guide's links.
VIP
Server: Fenrir
Game: FFXI
Posts: 764
By Fenrir.Motenten 2014-09-03 00:02:45
So what exactly are the 'best' setups for evisceration and exenterator? a lot of people have their own and they're different from the guide's links.
For Exenterator, look at the post immediately before yours.
For Evisceration, take the one from the guide and sub:
Pillager+1 head -> Uk'uxkaj (dex) cap
Pillager+1 body -> Plunderer+! body
Rajas -> Ramuh+1
Iuitl feet -> Plunderer+1 feet
Also, Brutal basically ties Tati (nq) when att is uncapped (and obviously is much better when att is capped). Tati/+1 is only a notable advantage when acc is uncapped.
Asura.Dakenx
Server: Asura
Game: FFXI
Posts: 92
By Asura.Dakenx 2014-09-03 02:55:22
Stop sucking at Thief.
Make Varja.
Win.
Never thought I would see the day that this was said lol. Glad I made one but I did it for just the love of THF and I lucked into it being amazing now. But it is crazy to see all of this love for Vajra after years of the weapon being disregarded as trash. Really shows how much 1 update can change so much in this game.
Ragnarok.Luloo
Server: Ragnarok
Game: FFXI
Posts: 121
By Ragnarok.Luloo 2014-09-03 03:09:42
Quetzalcoatl.Wakmidget said: »Now, how do we do these comparisons?
Well, I'm going to take a standard optimized Evisceration build. Putting this on Dynamis mob for capped attack, no buffs, and 0 over-TP rounds. I'll also use Mandau/Izhiikoh for the sake of comparison that's been brought up in the thread.
Interestingly, now that I optimized the DE build, Evisceration is only barely ahead of it (about 5%, 4882 to 4632).
Ramuh+1 comes out ahead of Epona's (4882 to 4876), so using that.
I'm using dex-aug'd Uk'uxkaj Caps on both builds.
OK, the basics are set. Now, first I'll look at the average number of hits landed with Evisceration -- 6.35
Now I want to look at the effect of crits on the final damage value. To do that, I'll adjust the crit rate in steps of 1/6.35 (15.75%), simulating 1 crit hit per weaponskill per step, and look at the average damage. Alongside that, I'll put the approximate percent chance of landing that many crits, given a 48.66% crit rate (base, gear, ws, and TP total for this test).
Code
Crits Dmg Chance
0 3718 2%
1 4095 9%
2 4472 23%
3 4848 31%
4 5225 23%
5 5602 9%
6 5979 2%
6.35 6111 ~
This is being compared with a Dancing Edge build that averages 4632 damage. Looking at the chart we can see that DE will average more damage if you only get 0, 1 or 2 crits, while Evisceration will average more damage if you get 3 or more crits. 0, 1 and 2 crits happens 34% of the time. So, 1/3 of your Eviscerations will do worse than DE, while 2/3 will do better.
The question is, what's your minimum weaponskill damage threshhold? If you don't specifically care how much damage is done (eg: fighting an NM where you're damaging the same mob over and over), go for the highest long-term average. However if you're doing something like Dynamis, where a particular minimum is useful (say 4500 damage, half the mob's health), but the high maximums aren't (you often won't weaponskill before the mob is half dead), then the consistency of the DE damage will be more beneficial. Plus, the higher peaks of Evisceration may never be reached, further reducing its average.
Let me know thoughts on this. I didn't spend as much time on it, so there may be considerations I've missed.
Did some more messing around with combinations of stats on Dancing Edge, currently what I'm using atm is:
ItemSet 327511
Upgrades I'm looking to make in the future would be changing out Jupiter's pearl with trux earring.
ItemSet 327490
For Exenerator the only change I can see would be changing pillager's armlets +1 to iuitl wristbands +1 with a +2% Double Attack augment.
In-game experience atm, Dancing Edge and Exenerator are are pretty close, but as Mote posted they are only about 5% apart, doesn't seem like there is much to eyeball.
I'm kind of curious though, assuming all the changes I mentioned in place of what I'm using now, how close are they on paper assuming capped attack(which is what I am building them for).
Still using unstacked Evisceration on things that are uncapped attack.
And thank you Mote for some very helpful suggestions as to gearing options I missed earlier.
Are you sure DA+2 and 1 more agi wins over a crit rate+4 and crit dmg+3 from plunderer's body?
Quetzalcoatl.Wakmidget
Server: Quetzalcoatl
Game: FFXI
Posts: 1388
By Quetzalcoatl.Wakmidget 2014-09-03 08:10:50
Quetzalcoatl.Wakmidget said: »Now, how do we do these comparisons?
Well, I'm going to take a standard optimized Evisceration build. Putting this on Dynamis mob for capped attack, no buffs, and 0 over-TP rounds. I'll also use Mandau/Izhiikoh for the sake of comparison that's been brought up in the thread.
Interestingly, now that I optimized the DE build, Evisceration is only barely ahead of it (about 5%, 4882 to 4632).
Ramuh+1 comes out ahead of Epona's (4882 to 4876), so using that.
I'm using dex-aug'd Uk'uxkaj Caps on both builds.
OK, the basics are set. Now, first I'll look at the average number of hits landed with Evisceration -- 6.35
Now I want to look at the effect of crits on the final damage value. To do that, I'll adjust the crit rate in steps of 1/6.35 (15.75%), simulating 1 crit hit per weaponskill per step, and look at the average damage. Alongside that, I'll put the approximate percent chance of landing that many crits, given a 48.66% crit rate (base, gear, ws, and TP total for this test).
Code
Crits Dmg Chance
0 3718 2%
1 4095 9%
2 4472 23%
3 4848 31%
4 5225 23%
5 5602 9%
6 5979 2%
6.35 6111 ~
This is being compared with a Dancing Edge build that averages 4632 damage. Looking at the chart we can see that DE will average more damage if you only get 0, 1 or 2 crits, while Evisceration will average more damage if you get 3 or more crits. 0, 1 and 2 crits happens 34% of the time. So, 1/3 of your Eviscerations will do worse than DE, while 2/3 will do better.
The question is, what's your minimum weaponskill damage threshhold? If you don't specifically care how much damage is done (eg: fighting an NM where you're damaging the same mob over and over), go for the highest long-term average. However if you're doing something like Dynamis, where a particular minimum is useful (say 4500 damage, half the mob's health), but the high maximums aren't (you often won't weaponskill before the mob is half dead), then the consistency of the DE damage will be more beneficial. Plus, the higher peaks of Evisceration may never be reached, further reducing its average.
Let me know thoughts on this. I didn't spend as much time on it, so there may be considerations I've missed.
Did some more messing around with combinations of stats on Dancing Edge, currently what I'm using atm is:
ItemSet 327511
Upgrades I'm looking to make in the future would be changing out Jupiter's pearl with trux earring.
ItemSet 327490
For Exenerator the only change I can see would be changing pillager's armlets +1 to iuitl wristbands +1 with a +2% Double Attack augment.
In-game experience atm, Dancing Edge and Exenerator are are pretty close, but as Mote posted they are only about 5% apart, doesn't seem like there is much to eyeball.
I'm kind of curious though, assuming all the changes I mentioned in place of what I'm using now, how close are they on paper assuming capped attack(which is what I am building them for).
Still using unstacked Evisceration on things that are uncapped attack.
And thank you Mote for some very helpful suggestions as to gearing options I missed earlier.
Are you sure DA+2 and 1 more agi wins over a crit rate+4 and crit dmg+3 from plunderer's body?
Perfectly sure, because Exenerator can't crit.
Quetzalcoatl.Wakmidget
Server: Quetzalcoatl
Game: FFXI
Posts: 1388
By Quetzalcoatl.Wakmidget 2014-09-08 13:11:05
A quick question about Dual Wield and capping haste at 80%.
http://www.bg-wiki.com/bg/Dual_Wield
I have been assuming it counts as JA haste for THF, now according to BGwiki THF get 25% Dual Wield at 99, so, assuming capped equipment haste and no Dual Wield gear THF could cap 80% with just 30% magic haste, correct? Or does Dual Wield go into the formula a bit differently?
Bismarck.Marmite
Server: Bismarck
Game: FFXI
Posts: 176
By Bismarck.Marmite 2014-09-08 13:37:48
I think the dual wield % is divided by 2, meaning 25% dual wield would be 12.5% delay reduction for an individual weapon.
I think you need Suppanomimi on thief with capped gear and magic haste to cap the delay reduction.
Apologies if I'm wrong, it's been while.
|
|