For The Shinies! A Guide For Thief

Please show me how to not accidentally overtping when max tp set has like 55% Triple Attack rate.

By not sucking. If you're going over 1500 religiously you're just bad, or doing it on purpose. (like, trying not to pull hate, at which point dps is completely irrelevant)

But really, most people really aren't using cento (yet). It's just... impractical. and THF is low on the r15 priority.

That was your worst trolling and bait I have seen so far. Good job.

With SAM roll you can easily make over 700TP per round (it will happen a lot with 55%TA), so you can instantly go from like 900 to 1600. That on top of how fast are attack rounds on daggers make it really hard to not overtp.

It's obviously easier when you solo without many buffs, but then Tauret is your best option anyway.

Does anybody have a current set of equipment for Mercy Stroke? Not interested in listening to a bunch of internet trolls have a debate about how worthwhile or worthless Mercy Stroke is. I have a Mandau, I like to use it on the rare occasion that I get to play around on THF for something and I think my most recent set is a bit dated.

Nothings changed in 2 years for mercy stroke except gere ring

ItemSet 368701
A Head
A Body (replace with a WSD herc DM body if you got one 6ish+)
D Feet

And oh yes, it's still firmly in the "very" bad category.

lol, that's okay. Do you think Lustratio Cap +1 (Path A) would be decent for non-SATA WS? Yeah, I know "decent" is very subjective here.

Yup, it's actually better, changed the set

Did not notice the 17 str difference

Crit damage on AF+3 hands doesn't make them best in slot for Evis over Meg or Mummu?

Mummu is quite a bit stronger. Artifact +3 hands have 4 % critical hit damage, 45 dex, 21 strength. Mummu have 6% critical hit rate, and with the set bonus they also add 61 dex and 24 str. Accuracy is pretty much the same between both pieces so it's a non factor. More critical hit rate is generally better than more critical hit damage, but the extra dex is also important. Evisceration's dex mod is 50%, so that's 8 more base damage on every hit regardless of crit or not. Mummu gloves +2 will outperform relic +3 in an evisceration build.

What are you getting the set bonus with?

Ring, possibly. But Mummu set is good for a few slots in Evisceration, at least it was.

I use Mummu hands + ring, but you can't really say that when using it like that you get +8dex on hands, then when compering rings, you also justify using it, because of +8dex from it on top of crit rate. It's the same +8dex.

The community THF guide has Lustratio+1 feet in capped attack, unstacked evis.

These are definitely better than herc with +6% crit dmg? Herc have DEX+24 vs 33 on Lust+1.

Im using dex/critDamage herc. Also you can only get 5% crit damage on Herc (unless with DM augment maybe).

Not 33 on Lust+1, 48 (Augmented path D), or 41 +3% crit rate (Augmented path B)

I'm in the process of doing some evisceration testing right now. I've been using evisceration a lot over the past few months and from my eyeballing of a few different pieces I'm fairly confident that the optimal build is slightly different from the one listed on the sticky. There are two pieces in particular I'm interested in getting data on. I'm also interested in comparing Tauret's evisceration with Twasher's evisceration and Twashter's Rudra's damage at 1000 TP. I'm doing the testing on Lady Lillith very easy for two reasons. First, I'm farming malignance so I'll be soloing her a lot anyway and her stats are consistent. And second, because she actually makes a surprisingly good target for this data collection. I'm just using 4 white mage trusts and koru and smorcing her down with weaponskill spam as close to 1000 TP as possible, so my numbers should be very consistent. And since she often utilizes her TP down move about 40% of my weaponskills are exactly at the 1000 TP mark.

I'm using Kparser for the data collection so I can get some reliable numbers. Weaponskill low/highs and averages will be included. I'll report once I have a sample size of 100 weaponskills for each setup, so that's 400 in total. I only started gathering data yesterday though so don't expect anything for about a week or so. I want a reliable sample size to back up whatever numbers the data shows. From my first setup last night I'm seeing the data stabilize at around 50 weaponskills. Numbers aren't changing by more than 100 in either direction after a fight. But I'm going to double the sample size for a good margin of confidence.

I'd be a bit interested to see how Adhemar Feet +1 stack up in an evisceration set path A or B. Assuming you're using head +1 only, Mummu hands, it would be -13 dex from Lust +1, but it would be crit rate +8%.

Edit: I'd see BiS being very close to this.

ItemSet 368761

Adhemar Path A or B
Lustratio Subligar Path B

Feet come down to this: (I'm tacking on set bonus for Adhemar onto feet)
lustratio leggings +1 Path D (+13 Dex over Adhemar, -8% crit rate)
Mummu gamashes +2 (+11 Dex over Adhemar, -3% crit rate)
herculean boots Aug: (Dex, acc/atk, Crit Damage)

My guess would be Adhemar would win Uncapped Attack/Capped accuracy pretty easily. Mummu would win Uncapped Accuracy. Capped accuracy/attack it would be pretty close on all those options. Herc is just a huge wild card for not that big of a payoff imo. You'd probably want some insane DM augment for it to even be worth the effort

The head and feet are the two pieces I'm most interested in comparing yes. Particularly this swap

Adhemar Bonnet +1 and Adhemar Feet +1

versus

Plunderer's Bonnet +3 and Mummu Feet +2


Originally I made the swap from adhemar to mummu and relic because of the accuracy. If I went path B on adhemar hat and feet my evisceration set was STILL 127 accuracy lower than my tp set, and that's pretty freakin huge, and I hated the thought of losing close to 170 accuracy going with path A. However, the relic head and mummu feet not only provide 59 accuracy more than adhemar path B, they also have 5 more attack (so essentially the same) and 18 dexterity, while only losing 3 crit rate and 6% crit damage. I originally swapped them in for consistency, but in my actual time using them I've been more than a little impressed with the damage output. I'm tempted to say that the 18 dex they add is enough to overcompensate the 3% crit rate at 6% crit damage. I want some hard numbers though, but I've swapped between both builds and I like what I see with the adhemar LESS than what I see with the relic and mummu. I know that eyeballing isn't reliable enough so that's why I'm firing up kparser and collecting some real data. But statistically the swap makes sense to me. Here's the combined stats in isolation

Adhemar bonnet +1 and Adhemar feet +1

34 strength, 68 dex, 40 accuracy, 70 attack, 6% critical hit damage, 8% critical hit rate (feet and set bonus combined)

-- 91 accuracy and 104 attack with dex and strength factored in

Relic hat +3 and mummu feet +2

47 strength, 86 dex, 86 accuracy, 62 attack, critical hit rate +5% (mummu set bonus to dex considered)

-- 150 accuracy and 109 attack with dex and strength factored in


They're actually a lot more comparable than you might think at first glance. Adhemar +1 focuses on the crits more than relic and mummu, but relic and mummu just offer a lot more raw stats. If strength isn't capped the fSTR is more base damage in addition to the 18 dex mod, and I'm finding from personal experience that it's significant enough to notice just eyeballing.

I'll report back once I have some hard data. Like I said, at the moment it's theorycrafting. Kparser will provide some definitive data.


EDIT:

This is what I'm currently using and what I've theorycrafted ItemSet 348286

I'll compare it against adhemar bonnet +1 and feet on the path B rout. I'm really loathe to want to go path A, because like I said, path A would be 170 accuracy lower than my TP set, and that's enough to start whiffing on swings.

EDIT: The mache earring +1 swap is another selling point I wanted to test out. I may actually do sample sizes closer to 70 apiece and throw in that as an added data point. The focus of this testing is meant to revolve around evisceration's dex modifier. Essentially I want to narrow down the point it becomes better to focus on dex and raw stats over crit rate and crit damage. Evisceration used to be 30% dex mod, but since the overghaul in 2014 it's using a 50% dex mod and from my experiences adding a lot of dex really feels impactful on the damage output.

If it depends what you're main handing then I do not have Twash btw. I use Aeneas/Tauret if mostly using Rudra's and vice versa if using Evis mostly.

I forgot about augments on the Lust gear though, derp.

That is going to be a really hard thing to quantify, since the window for DEX affecting crit rate is so narrow, we stack tons of DEX on our WS, and it would be different from mob to mob.

You're theory makes some interesting follow up questions.

Where would pillager's bonnet +3 fall compared to adhemar bonnet +1 and plunderer's bonnet +3? To me that was the logical replacement choice to adhemar, not plunderer's.

I personally feel that SE kind of screwed the pooch with Relic head. Skulker's +3 is going to look very very similar when it comes along.

For the time being I'm main handing tauret and offhanding my taming sari. I'm going to make twashter an entirely different data collection set and focus first on the armor only. Here's the list of data I want to test against.

1 -- The theorycrafted set I listed above.
2 -- The theorycrafted set I listed above altered to include adhemar bonnet and feet +1 instead of mummu and relic +3
3 -- The theorycrafted set I posted above altered to include two mache earrings +1 instead of sherida and moonshade
4. The theorycrafted set I posted above altered to include adhemar bonnet +1 instead of relic +3 (to test adhemar bonnet +1's crit damage versus relic +3's raw stats in isolation)
5. The theorycrafted set I posted above altered to include Pillager's bonnet +3 instead of relic +3 (to test artifact bonnet versus relic +3's raw stats in isolation)


I'll post Twashter mainhand evisceration and rudra's as a followup later.

When I look at the 18 dex I'm more concerned with the 50% secondary mod and how it affects wsd than crit rate, but that's a valid point as well. And I'll throw in some data for artifact hat over relic too.

I'm going to cut the sample size down from 100 to closer to 70 for each set then and just add more variety to the testing. Like I said, I've already got over 50 samples of my first data point and the numbers are only flucuating by low triple digits after each run now so the margin of error should be more than reasonable at that point.

18 Dex has the potential to add 0% Crit Rate, 1% Crit Rate, 2% Crit Rate or 2-11% crit rate based on what dDex is. Unfortunately, it's really hard to know if your are at the dDex where it's the 3-11% range.

I'm curious, are you seeing pillager's culottes +3 outperform lustratio subligar +1 Path B? 39 STR, 15 dex, 54 acc, 30 atk, 5% crit damage, 5% TA vs 43 dex, 20 acc, 38 atk, 3% crit rate

That's kind of the problem with evisceration. It's a multi hit critical weaponskill with a significant dex mod. That's going to make optimizing it inherently tricky. The reason I prefer to keep the relic legs +3 is because of the combined accuracy, dex and attack. Whereas lustratio offers more raw stats, it's going to lose consistency when accuracy matters. When swings start to miss in abundance, damage plummets. Again, I'm tempted to say that lustratio +1 has a higher damage potential and if accuracy doesn't matter it'll beat out artifact +3. But on the flipside, if the accuracy does matter then artifact +3 legs would win.

Since I'm testing on Lady Lillith very easy where accuracy shouldn't matter though, it'll at least be a good opportunity to try and determine "max damage" setups. Right now I don't know of much if any testing that's been done revolving around it's different properties. So I wanted to get some numbers together to at least be able to start looking at it from a few different angles.

Actually, I think the str on AF +3 generally mitigates the additional Dex on Lustratio. +~9-10 fSTR vs. +14 WSC. I'd be inclined to say that for the additional ~13 acc/5% TA/5% cDMG it's probably worth losing 4-5 base damage.

With at least 1 piece of AF, I'd guess you'll get better bang for your buck replacing Ilabrat ring with regal ring unless you need the STP for something.

You're probably right about that. Anyways I'll give you guys some hard numbers after I finish collecting some data with a few different setups. The average damage is literally fluctuating by only double digits now per run after surpassing 50 weaponskills, so I'm going to cut the testing at around 60 sample size per set. The margin of confidence should be reliable enough with that.

If you are using Tauret don't you have really high crit hit rate anyway? Are there not serious diminishing returns on crit hit rate?

If you mean the crit bonus on tauret thats white dmg only, doesnt work on ws

Yeah, Tauret is fine for evisceration testing. I figured since I'm apparently going to be spamming Lady lillith for a while anyway I may as well make my time a bit more productive. I'd done about a hundred VE runs over the past few weeks and got no armor, so I started collecting data yesterday. At the rate I'm going I have plenty of time to get some good sample sizes before I finish the set. I'll start around 70-80 weaponskills in my initial setups and increase sample sizes universally after should I still need to do more runs.

I'm going to keep an open mindset to this too. Up until now my experiences with evisceration have been all eyeballing, and we all know how personal bias towards eyeball numbers work. I'll be the first to admit I don't know for sure if my hunch is correct and I may be doing it flat out wrong. I'm interested in getting some real data though to find out. All weaponskill are firing off in the 1000-1300 range, and none of them are getting killshots to preserve integrity. Hopefully when all is said and done I'll have some good data on 5 or 6 different evisceration setups. We might be able to make some reasonable asumptions from that. In addition to weaponskill highs, lows and averages I'm also particularly interested in seeing how the bell curve lays itself out for each in the offense tab. Scoreboard is nice for general stuff, but for precise data collection Kparser is leagues ahead. I miss Motenten.