Patch 3.4 Stat-Weights

Updated 31/10/2016 18:30 GMT – Added the Dark Knight, shockingly bad Parry weight.

If you want Machinist Stat-weights done, I’ll need a community effort of people gathering data points for machinist turret damage at level 60. People will need to use random WD/DET/DEX values to figure out how they scale (including 0 WD/broken weapon), as well as identify the base CRT value of Turrets (if they crit). I do not have 60 Machinist and no longer have sub for FFXIV as I don’t have time to play.

However, if people get me the data I need, I can create a damage model for Turrets and with Japanese Formulas, create a very quick MCH spreadsheet (similar to the rest of my sheets) and calculate the weights in my limited free time.

  • STR: 1.000
  • VIT: 1.000
  • WD: 26.523
  • DET: 0.396
  • CRT: 0.541
  • Sw/Oath SS: 0.274
  • Sh/Oath SS: 0.330

Warrior w/ Deliverance  

  • STR: 1.000
  • VIT: 1.000
  • WD: 27.095
  • DET: 0.387
  • CRT: 0.546
  • SS (Once at 550+ for 9 GCD Berserk/ x3 Fell): 0.350

Warrior w/ Defiance removed, as there’s less than 0.05 variation between the two sets this tier, thus no notable DPS or BiS difference. 


Dark Knight

  • STR: 1.000
  • VIT: 1.000
  • WD: 26.547
  • DET: 0.380
  • CRT: 0.530
  • SS: 0.315
  • Parry: 0.011


  • WD: 14.956
  • STR: 1.000
  • DET: 0.207
  • CRT: 0.279
  • SS: 0.170


  • WD: 15.459
  • STR: 1.000
  • DET: 0.212
  • CRT: 0.297
  • SS: 0.173


  • WD: 15.195
  • DEX: 1.000
  • DET: 0.210
  • CRT: 0.282
  • SS: 0.141

Bard w/ Wanderer’s Minuet:

  • WD: 16.824
  • DEX: 1.000
  • DET: 0.215
  • CRT: 0.366 
  • SS: 0.233


** Black Mage **

  • WD: 11.884
  • INT: 1.000
  • DET: 0.206
  • CRT: 0.256
  • SS: 0.283

Yes. The SS weight is not 0.4. Both myself and a few others absolutely do not believe it at all; and the raw data of exponential scaling of SS does NOT scale twice as much as Crit. The *only* potential time where SS will be worth that much, is if you’re doing one of two things:

A) Comparing the base exponential increase of SS, vs the increase of 1 point of SS allowing you to get an additional cast within the same time frame,  which has many flaws. I can use this exact same method to skew the Dragoons SS weighting above Critical Hit Rating if I wished to.

B) Comparing the base exponential increase of SS, vs the next threshold of SS. In which case, 25/26 points of SS are worth exactly the same, disproportionately skewing the weighting of SS. Whilst this does reflect more real-world values, this would go completely against the methodology which myself, Sunny Hirose and Krindor use.

Summoner w/ Garuda – No book smacking:

  • WD: 11.835
  • INT: 1.000
  • DET: 0.192
  • CRT: 0.287
  • SS: 0.195

Summoner w/ Ifrit removed: 0.03 averaged variation between Ifrit/Garuda this tier, for some strange reason. Must be to do with the insane amounts of SS. 


Healer Weights are calculated PURELY for full time DPS. For science only; take these weights with a grain of salt. These weights do not take into account Mana consumption, or ANY form of healing/shielding. If you’re going to be hybrid healing/shielding, I’d recommend prioritising Crit over any of the other secondaries (with the exception of accuracy).


White Mage DPS ONLY: 

  • WD: 9.128
  • INT: 1.000
  • DET: 0.159
  • CRT: 0.182
  • SS: 0.135
  • Acc: Cap it


Scholar DPS ONLY:

  • WD: 9.128
  • INT: 1.000
  • DET: 0.190
  • CRT: 0.266
  • SS: 0.316 – 0.366
  • Acc: Cap it

The reason why the Scholars SS weighting is high, is that roughly 51% of its overall DPS is coming from DoTs. So not only is SS buffing your DPS exponentially due to how it affects your GCD/Casts, but its value is being amplified by also increasing directly affecting 50% of your total PPS via DoTs.

Also, the weighting scholar DPS varies a lot depending on the rotation you use for some reason. It could be an error with my model somewhere handling clipped dots though. Letting your DoTs completely drop, rather than clipping them, boosts the SS weight by 0.5. This is something I’ll be looking into and trying to figure out and does look like it to be a DPS gain, over clipping them.


Why are Tank Secondaries valued so much? Well Jeff, that’s very easy to answer. Apparently, the word on the street is that VIT/STR scale AP, or Attack Power, 0.45 per point.

In other words, 1 VIT/STR = 0.45 AP, or if we reverse this…. 2.222 VIT/STR = 1 AP, because 1/0.45 = 2.222.

This would literally make the AP: STR/VIT weights be:

  • AP: 1.000
  • VIT: 0.450
  • STR: 0.450

All I’ve done, is scaled up VIT/STR to equal 1 point of AP, or multiplied the VIT/STR weights by 2.222… But as I’ve scaled up VIT/STR, I also need to scale up all of the other secondaries.

If you wanted to see the weights in their “natural” format, you’d simply divide all the values by 2.222. As an example, the weights for a Warrior in Defiance, not scaling up to 1 AP, would be roughly:

  • AP: 1.000
  • STR: 0.450
  • VIT: 0.450
  • DET: 0.136
  • CRT: 0.170
  • SS: 0.125

As you can see, even in this format, the difference between Secondaries and VIT/STR are still small.

Another thing you can check, is the ratio between STR/VIT and CRT, so you can see that the scaled up values are similar.

0.450 / 0.170 = 2.647

1 / 0.378 = 2.646

Of course, all numbers you’re seeing here have been rounded up, so there will be a slight difference.


What’s up with SS and Bard?! Why did it Jump up so much?

The most important thing is the Function of a Bow Mage…

  • Bow Mage is basically a Black Mage
  • Is a Turret-DPS. Stand in one spot and PewPew
  • Has cast times, on top of a recast time. Quicker the cast time, quicker the Job can land hits.
  • Does not use Auto-Attacks

Okay, so let’s go into the Auto-Attacks, as this is the primary reason why Bards SS is valued highly.

With a Standard Job, or a Bard who’s not in Minuet, 1 point of AP/DET/CRT is scaling not only their Auto-Attacks, but their GCDs. The effectiveness of AP/DET/CRT, whilst a Job is using Auto-Attacks, will always be greater than SS, as all SS can affect are the Jobs GCDs.

When you remove the Auto-Attacks, your Stats are now only affecting one thing; your GCDs.

This relationship is reflected in the Stat-Weights.

Next, is how Determination functions. It’s a linear stat, meaning, whilst every point of DET you add on will be exactly the same as the previous, the percentage effectiveness will always be decreasing as your DET increases.

SS on the other hand, exponentially increases your DPS per point you gain. It’s effectiveness will always get better the more you stack, hence why Casters stack so much of it.

You can read more about SS exponential gains via this link:

So as our SS increases, the effectiveness on our DPS increases, but as our DET increases, it’s effectiveness decreases.

The one thing I’ll point out now, just so people aren’t confused as to why Stat-Weights increased across the board for all Jobs, is because they scale off AP. More AP, the more our secondaries are valued.

85 thoughts on “Patch 3.4 Stat-Weights

  1. Anonymous says:

    Thanks for the video and for this website ^^

    If i may do a little suggestion : Do NOT trust other’s results. Double check them before posting please. This suggestion is mainly for purostrider’s work. Yes he worked a lot for the blm community BUT this guys doesn’t know what methodology means. Because of that, he has wronged the community .

    ie : not so long time ago he tested Stat-Weights naked to only see what dmg he had while increasing ony 1 stat… When we linked him tests made by other ppl (a week-ish before) which all shown the same result… He said “because he made it (his datas) himself, others must be wrong”…

    I could give you a ton of example like that but it’s not the place to do it. Just keep in mind for future updates that you really should to check other’s work if you post it here.

    (and btw, I’m 110% sure that kenji’s formula matches puro’s datas, but since those datas can’t be trusted, the formula and the stat weight he found shouldn’t be posted here with a “it’s ok because it’s puro”)

    Liked by 1 person

    1. I remember all of the arguments on the Forums between Kenji, T0rin + others and PuroStrider very, very well. I actually had a conversation with Puro on Reddit at the time, and had to specifically explain in depth why they’re arguing his stuff is completely incorrect.

      It’s for this exact reason that I’ve absolutely refuted any of Purostriders work for my own data collection. I don’t have a working Simulator/Model for Black Mage, which is why I posted his stat-weights there, otherwise I’d 100% use my own work. It’s also the reason why I went and tested the Japanese Job Coefficients myself and there’s actually a few inconsistencies with it.

      Thank you very much for your comment and I shall definitely take it into consideration for any of my future work!

      A few things I’m doing are:

      1) Recalculate Job Coefficients
      2) Correctly identifying Linear Scaling of Determination from level 1
      3) Identify whether there’s a correlation between Determination and Attack Power
      4) New Damage Formula


  2. Anonymous says:

    I can’t wait to read your results ^^

    I hope things wont change too much in 3.0 so we wont have to learn everything again 🙂


  3. Jack Boulton says:

    You have no stat weights for MCH on here. I’d assume they’d be the same as BRD but a quick search says they’re different. Do you know the correct stat weights for MCH? I’ve also seen different stat weight (albeit marginally different) with and without Gauss Barrel / Wanderer’s Minuet.


    1. 1) I do not have them

      2) They are most definitely not the same. MCH doesn’t have any crit based procs AFAIK. Also, MCH Turrets scale DET a lot better than usual.

      3) All of those weights at garbage. SS is not valued at 0.02 for the MCH. Its literally impossible for it to be valued that much.


      1. This is a pretty hot topic for me right now, honestly, as with how low SkS is allegedly rated it means that it’s technically correct for MCH BiS to end up about 75 accuracy over cap because you’re taking a Crit/Acc piece over Det/SkS (Gordian Helm vs Eso). When you say “impossible for it to be valued that much” in regards to SkS, do you mean the estimated value is too high or too low?

        I would wonder if there is a breakpoint with MCH for SkS that would allow an additional GCD within Wildfire, but I wouldn’t be able to work out whether or not the loss of other stats would be worth it in the long run, as it is with WAR.

        Do you have any kind of estimates at all for MCH? Right now I’ve personally been assuming that the stat priority follows WD > Dex > Crit > Det > SkS, ignoring mandatory Acc, ofc.


      2. THe lowest I could see SS being valued, for a Machinist, is 0.09. Not any lower. SS, over a period of time, will always allow you to land additional GCDs.

        In the case of fights like last Phase A3S/A4S, which is a 6 minute Dummy fight, SS will definitely scale your DPS a lot better.

        MCH will take a while unfortunately. I need *a lot* of data it’s Turret, AAs and GCD skills.


    1. That’s something I’m still investigating.

      From small talk with a friend of mine, no. It hasn’t. The cast times seem to be exactly the same as they were before – just, it’s more fluid.

      You still suffer from the same old clipping issues, etc etc.

      Me and this mystery theorycrafter, who shall name himself soon enough, also discovered with his simulator that in longer “phases”, such as A3S final phase, the additional GCDs you gain from not clipping your GCDs and using Fuma over Raiton, makes it a lot more worthwhile. Also, using Fuma in this manner increases the SS weighting from whatever the heck it is that I mentioned, up to around 0.116.

      In other words, don’t read Shasta’s guide. It’s crap.


      1. Adam says:

        Interesting. One last question (Sorry!), but i was reading that bards want to have about 600 SS have you noticed anything supporting/disproving this?


  4. Anonymous says:

    Dragoon should ideally have at least 600 skill speed in order to for Geirskogul, Heavy Thrust (24s) and Blood of the Dragon to synergise well together. (Similar to the 9 GCDs Berserk thing for WAR)


      1. Based of something I Theorycrafted originally Day 4 of HW release*.

        I made the original DRG 3.0 Rotation, lol. That’s why I were curious as to what you were referring to, you’re speaking as if I don’t know about something that *I* Theorised.


  5. Hey Dervy, I’m compiling my own resources on stat weights across various pieces of gear and I’m wondering about the WD weight on the spellcasting classes. Are you still weighing the weapon’s physical damage or are you weighing it’s magic damage instead? It’s the different of about 400 or so of an effective INT weight using the magic damage.

    Ultimately it doesn’t matter as long as I’m consistent in which I use considering I’m just checking which is highest, but I’d still like to know for the sake of clarity.


  6. Anonymous says:

    Hello. Thanks for your amazing work. And i have few questions about scholar: Do you have any approximate info about scholar stat weights ? And when you’re planning to release healer stat weights data ?


  7. Anonymous says:

    Hey Dervy. Would it be right to assume that MCH Skill Speed would function similarly to Bard’s in regards to weighting? A lot of what applies to Bard/BLM appears to apply to Machinist as well in regards to speed. I know you previously said a few months ago in these comments that Turrets seems to scale well with determination, and I’m curious to know if you still think this after compiling more data. Anyway, thanks for the great work and hope to see you keep doing this as long as you can.


    1. BRD desires Skill Speed due to it reducing the cast time on Emperyal (sp?) Arrow, as well as minorly boosting the damage of the two DoT effects it has. MCH doesn’t benefit from those effects as it only has the 1 DoT, has no cooldown scaling with Skill Speed, and also doesn’t have the turret benefit from it.

      BRD also has a minimum Skill Speed cap of 600/700 (I forget) for an optimal opener allowing 2 Emp. Arrows, whereas MCH has no such cap unless you feel like getting 1000+ Skill Speed to allow an additional GCD in Wildfire which is by no means worth it. What MCH -does- have going for it in terms of Skill Speed is that the more Skill Speed you have, the less valuable stancedancing in/out of Gauss Barrel becomes, to a point where it’s no longer necessary for optimal DPS. However, as with the additional GCD, this cap is ludicrously high and you’ll burn through all your TP.

      At best, Skill Speed is a dump stat for MCH and unless there’s some incredibly obscure mechanism that we haven’t noted, Crit and Det far, far outweigh it.


  8. First of all Dervy Thanks for your hard work.

    The next thing is something that bothers me about the summoner, and which I couldnt found in any reddit discussion or the lodestone forum. The idea is to use piety materia (instead of our secondary stats) to gain more MP+MP Reg and therefore substitue Ruin I with Ruin III.

    The calculations i did bring the following: 1 Piety equals ~ 6MP. In addition we gain 2% MP every 3 seconds which makes 0.12 MP and additional 20% every minute through aetherflow (Energy Drain is sadly not effected) which gives a total of 0.18 MP per Tick per Piety
    I will go with a 7 minutes fight as example. This fight will have 140 Ticks and 1 Piety will regen about ~25MP. To exchange R3 with R1 we need additional 884 MP or to express in piety ~36 Piety. This exchange will bring us a benefit from 120 Potency or 3.33 Potency per Piety (or better 0.007 Potency per second per Piety).

    Every 148 Piety we gain enough MP to substitue another R1 to R3 bringing 0.81 Potency per Piety but will have a decreasing value of Potency per second. For the 7 min example it will be something of 0.002 making piety overall to 0.009 PpS

    My Problem is now to normalize this value and make it comparable to Crit/Det/SS. Maybe you can help me out.

    P.S. I also don´t know if the Reg from astro, bard or machinist is flat or a percentage of our value. This might further contribute. And of course there is a hardcap when every R1 was turned into R3. This might happen espacially with reg from other classes.


    1. That’s easy, divide the PPS by 100, +1 and you’ll get a multiplier. So that PPS value will be *1.00009

      So, Det vs PIE (granting an additional R3).

      140*0.000137 +1 = 1.020301783

      100 * 1.020301783 = 102.0301783

      So it’s a 2.0301783 PPS gain stacking 148 DET.

      Liked by 1 person

  9. Nicole says:

    im just curious if you know anything about healer stat weights? ive been looking everywhere because i heard for whm that det became a higher priority than crit but cant find anything. google brought me here but still cant see anything about it ; ; please help!


    1. I’m still working on them.

      DET is only a priority over CRT subjectively, purely from a progression and consistent healing point of view.

      If you’re wondering about DPS, it’s SS > CRT > DET.

      PIE will always, however, be the best stat to invest in as a WHM, up until a certain point.


  10. Anonymous says:

    Hello, u mentioned WHM’s dps stat priority above, what about SCH, is it same or crit would be better due to pet’s crit proc and energy drains ?


  11. Anonymous says:

    Hi Dervy, I have for a long time believed that Skill Speed is better than Determination for warrior. Does these weights take into consideration that Warrior has a net TP gain and fracture is an overall DPS loss?


    1. Fracture is a DPS gain. It’s a 300 potency skill, stronger than Butcher Block, without taking into account SS affecting DoTs. Having said that, getting multiple FCs in Berserk is better than fitting Fracture in there.


    1. Yeah. Me and Krindor have the same train of thought. When it comes to checking SS thresholds, SS and CRT are worth very similar. (I’d argue SS is better than CRT), but it’s extremely hard to value SS using thresholds, especially as for 25/26 points, SS is worth nearly 0 (except for affecting your DoTs).

      My weights do not account for thresholds what so ever. Essentially, use the weights I provide to build gear sets, calculate total weights, then check for thresholds. Which ever has the least “dead” SS Values, usually ends up having a greater overall DPS impact.

      If you don’t pay attention to thresholds, you’ll only lose, at max, 3 DPS.


  12. Anonymous says:

    Exactly what would one need to do to provide you with the information for machinist? Just parses with the specifiers you laid out and labeled as such?


  13. Spence Stragos says:

    Why the huge change across all stat weights on Blm? Crit went up so much but spellspeed went down compared to the weighting a this guy used back in 3.3:

    He mentions that he would get those huge leaps in stat weight so he capped it at .4. Idk, seems sketchy to me to have that huge of a decrease in SS weight from 3.2.

    All this aside, nothing can help a Blm as much as spell speed can in an actual encounter. Being able to finish a fire 4 cast with 1080 spellspeed is 504 extra potency over a Blm with 900 that had to dodge an aoe early and cut their cast time. On a dummy, sure, crit is valued very well relative to spell speed, but people don’t say Blm is the


    1. Because he uses methodology completely different from how I calculated weights.

      My weights are calculated using a base rotation and figuring out what the increase of the DPS will be over infinity, rather than what he did, which was calculate the weights within a time-frame, then checking the weight of SS when it allows you to get an additional GCD within that time frame, thus throwing the SS weight to insane levels. He then averaged out this weight.

      Just as evidence, here’s the DRG weight at 0.79 if I use similar methodology.

      So, 0.79 + 0.170 = 0.98/2 = 0.480 SS weighting for DRG.

      That’s roughly what he was doing.


  14. Ares L. says:

    I try to figured out, what will be better for a SMN with Ifrit. For me and my testing i can confirm that Det is higher than SS, pls don´t Forget, you get often SS when your EGI lands a crit. So somtimes i have SS and det up.
    When you put more in SS you can´t get the det. Bonus.
    Thats my opinion. Only for your info.


  15. _____________________________________________________
    Scholar DPS ONLY:

    WD: 9.128
    INT: 1.000
    DET: 0.190
    CRT: 0.266
    SS: 0.316 – 0.366
    Acc: Cap it

    Wow thank you for this! Just a question- Was this done WHILE Selene is spamming Embrace for a chance to proc her enhanced pet actions and was Fey Wind considered (I know thats allot of extra work, but I was just curious if it had already been done!) I feel like it would slightly increase the value of both Crit and SS that way.


  16. E.R. says:

    Dervy, I’ve stumbled across this set of research on my own, trying to work on my BLM BiS and such. But I notice you keep mention the 25/26 S.Speed thresholds. I have no idea where you’re getting the 25, as in my personal research of literally finding as many different combinations of gear as I can to get my sspd to different points, I’ve only seen 3 thresholds, 20 and 26 for the Fire4 cast timer, and 26 and 27 for the Fire1 cast timer and GCD. I’m curious if I’m somehow just completely wrong, and wasted hours of my life looking at when the cast/recast timers flipped over that .01, or maybe the the information out there isn’t as amazing as people still believe.

    My other thought also dealing with SSPD, is that, all these numbers are 100% dummy numbers, right? So while they are technically “perfect world” stat weights, they may not reflect BLM, SMN, and maybe even BRD’s actual value of sspd. Because being able to move and not break a cast an extra half second before you normally would have, could potentially save you that extra GCD vs having to break a cast just to survive.

    All of that said, I do honestly appreciate your work and that of your collaborators on this project, and I hope to get feedback on the conflicts in information in my little Lala brain.

    Thanks again,


    1. It’s 25/26 points of SS that are USELESS. GCD shifts 26/27/26/27/26 -> repeat.

      So of course, every 26/27 shift, one point is actually worth something (edging you over the threshold), but the other 25/26 do nothing to increase your DPS, other than affect your DoTs. You just confused the wording.

      Also, depending on the base cast time, the threshold intervals change. You can read it all up on this googledoc link:


      1. E.R. says:

        Thank you for responding, and forgive me for not replying sooner,

        You’re right, the wording apparently threw me off, must have been the lack of sleep for the week before. Going back and reading it again (after more than an hour and a half of sleep) I found that your wording is fine, I’m just an idiot.

        And yeah, I independently found the SSPD thresholds myself. But you’re also correct in the fact that the thresholds are based on the base cast time, such as Fire4 being a base of 3.00 seconds. And the pattern for that is 20, 20, 26, repeated. Meaning that 20 sspd could still very well affect your dps. Especially since blm rotation is so dependent on that 3.00 cast timer rather than the actual GCD.

        All of that aside, I still am a firm believer that SSPD, if not mathematically valued extremely high, is at least in practice, even more important than the .283 would lead one to believe. Based wholly on the necessity of movement.

        Thank you again for your work and your reply,


  17. Ergast says:

    I have a question. Where is Piety placed for healers? I’m asking because I have the 230 WHM anima, I’m thinking about finishing it once I get my 260 DRG weapon, and I’ve heard several times that extra Piety is a dps increase for WHM thanks to the extra amount of spells that can be casted before going OoM.


    1. It’s impossible to put Piety anywhere in my weights, as I focus purely on rotations which are calculated over infinity.

      If you’re concerned about Piety, you should focus on Piety that’s *already* on your gear, then meld Accuracy. Accuracy is the biggest DPS increase for healers. It’s not questionable.


  18. Suki says:

    Hmm.. interesting weights so far, thank you for everything you’ve done for the community! 😀

    I do have a few questions though, there’s 2 jobs missing on this list, while I know MCH’s priority is pretty simple, what do you think about AST for weights? I’d imagine their weights would be closer to a WHM’s than a SCH’s (that SS kinda blows my mind), AST has 2 potent DoTs, and a pretty strong spammable spell (Aero has a negligible gain over Malefic 2, primary usage of it is movement/off-GCD/card draw if any of those come up in the next 18s).. then there’s Diurnal vs Nocturnal for damage xD I’d imagine SS would be quite low for AST just due to the 5% AS on Diurnal (and not getting that extra SS scaling for DoTs).


    1. You could be on track. Indeed. Though, I still need to model it! I could easily make Diurnal vs Nocturnal models in one sheet very easily. It’s just allocating enough time.

      I have 2 weeks off University on the 18th November, so I’m hoping I can finish off Machinist/Astrololgian weights during that time off.

      Liked by 1 person

  19. Hi,
    Do you still need any data for the MCH information? I dont mind AFKing in front of a dummy for a while if all you need is raw data. Just let me know where i can send logs and if you need anything more specific than ‘random’ and Nil WD plot points, or if I should just craft a few random greens, break one of them and go to town.
    For now, pulling a random look at a dungeon I did with a MCH in it, it’s clear turrets do crit, and at a glance it seems it has similar crit values in both percentage and multiplier – maybe even a bit extra. Though this could just as easily be due to the restrictive and random nature of just one dungeon parse.


    1. All pets use the exact same crit multiplier and %chance rate as our characters. However, they have a different base crit rate. That’s what I want to identify.

      If you have any data-samples, feel free to record your data in a googledoc, and send it via or @itsdervy on twitter


      1. dac says:

        Having some MCH stat weights would be nice, and I could help out. I main MCH and run a parser a lot. I just don’t know what info you need and how to process/send you that info.


      2. GoogleDocs.

        I need Turret Data. Loads of NON-CRIT MIN/MAX damage parses using varying sets, as I need to identify the damage formula for pets. I wanted formatted similarly:


  20. Milkyray says:

    Hello Denvy.
    I want to thank you for all the work you are doing. You do such a great job.

    If its okay i want to ask you a question.
    In the Ninja 3.4 graphic we saw that critrate gets a higher stat weight if we have more of it. So i want to ask how many crit you used to calculate 0.282 for crit? And could it be a higher Weight if we habe for example 1300+?


  21. Anonymous says:

    Dervy, you mentioned at some point that your DRG rotation has changed slightly due to the new Stat-Weights. Any chance you could share your rotation please?


    1. Did I? Did you confuse something else? I mentioned before that due to animation locks, you can use a lower SS, thus allowing you to use your oCD in different ares for better optimisation, but the GCD rotation is and will always be the same up until 2.15 GCD.


  22. Thanks Dervy for the new Stat Weights. I have a question. The graphic you posted showed
    that the Stat Weight for Crit gets higher if we have more of it. Does it mean when we have for example 1366 Crit that the Stat Weight for Crit is higher as 0.282 ? Sorry for my bad english.


  23. Peri says:

    I’m curious with summoner, wouldn’t spell speed increase higher incrementally based if the spell speed buff they receive from their pets, as the more they have, the more they receive.


  24. Zak says:

    I’m wondering if you still need MCH turret data or it’s just that you don’t have enough free time to get the weights. If you still need data, I’ll be glad to try and gather some.


  25. Hey Dervy !

    I’ll be glad to know more about your methodology to calculate stats weight. In particular, some questions :

    1/ How do you choose base stats you’ll use to differentiate stats weights ? (for exemple real stats weight are very different if I have 600crit or 1200crit).

    2/ How do you know repartition of different damages sources ? Increase SS will increase DOT with a 1/7722 factor, GCD skills with an other factor given by GCD formula, but OGCD will not change at all. Do you modelize with maths a basic cycle for every class ? Do you use ACT data of a perfect cycle parse to give a weight to every damage sources ? Or something else ?

    3/ Once you have maths, how do you compute stats weights ? Is it “simply” by adding something like 50 to every stats then watching difference, or something harder like using a derivated function ? How do you deal with classes with a hard cap, like DRG, where skill speed drastically change DPS, not always in good ?

    Thanks in advance !


  26. 1) I set my original baseline as an average of the available stats from maximum ilvl of that tier. Once I’ve identified my baseline, I’ll calculate the weights, then use those weights to calculate a set of gear. I keep repeating this until the weights continuously match up with the highest DPS set. Of course, as I’m a Dragoon main, I can do a bit more vigorous testing with it as I know the Job and have multiple gearing/BiS spreadsheets to compare with (I never use ariyala)

    2) I work in Potency throughout the model, until the finall DPS calculation. My Model separates AC Damage, DoT Damage and oCD damage. The AC & DoT damage are in the same table, and divided over the duration of the rotation, where as oCDs are calculated as if they’re spammed 100% of the time with no downtime.

    As for rotations, it varies. There’s some Jobs like BLM and DRG which are basically the same 1-2-3-4s on repeat, so I use cycles until I’ve got at least 1 minute worth of actions on my timeline. In this case, their base rotation acts as a cycle. For Bards, I got rid of WB/VB in favour of a single IJ.

    One other thing, buffs, like BFB and Hawks Eye etc are not factored into the timeline, but averaged over a duration. Only exception is Berserk for WARs, as the 8th/9th hit factors into the rotation.

    3) I quite literally +5 divide the delta by the baseline DPS value.


    1. Thanks for answers !

      This methodology is pretty close to what I made for SCH :

      Your work is, I guess, more reliable than mine, because I assume it’s always possible to refresh DoT at right time, so I simulate too many Broil.

      I used DRG formula for SCH damages, what it work pretty well.

      I try to do same thing for SMN, so I tried your SMN formula, what it doesn’t match my data at all (formula prediction is ~10% above).

      Then, I tried DRG formula, and it’s match perfectly (excluding pet, pet is another story, damn pet…). I tried same thing by changing my stats (less DET, less INT), and data are still stuck to DRG formula prediction.

      Here are data :

      So what SMN formula mean ? Did I miss something ? It is include more than potency-to-damage conversion ?


      1. I guess I need to go back over the SMN data.

        You have to remember that majority of the time, I were working with data points which other people have sent me. I only have DRG at 60, sadly. So sometimes people mess up with their data collection.

        Also, according to the JP Theorycrafter, SMN/BLM/DRG/BRD output roughly similar levels of damage per potency, so maybe that’s why my DRG formula worked well.


      2. for 4.0, i’m planning on setting up a donation box, so people can donate for me to buy level 60/70 power jump pots. This way, I can ensure that all data collected will not me 3rd party and I can verify it myself.


  27. mauricedoison says:

    Maybe it could be usefull to have a clear protocol to exchange parse data ?

    something like :
    – Only as CSV output of an ACT parse, associated with full character stats.
    – 3min minimum duration.
    – On a dummy target only.
    – Repeat only a single skill, with no buff, no gcd, no party, no pet.

    I’ll be glad to help to collect data, If we can be sure there data are reliable and usable.


  28. Yup. That’s generally the format I want. But sometimes people don’t run their parser long enough. To identify min/max values for damage formulas, you need to leave your character parsing a single skill for a minimum of 10 minutes. In 4.0, I wouldn’t be surprised if we needed to do 20 minute parses per skill action to clearly identify min/max values.

    I base my formulas off a non-linear regression calculator (which was made in excel). The data I put in reflects on the outputted results.


    1. mauricedoison says:

      Sorry to bother you again !

      I’m still working on SMN damages modelisation, but all tests I made give me a Pet’s damage formula very different of SMN damages formula (damages average are really lower, but it don’t seems to be a linear factor, determination part seems to grow faster than SMN formula). Do you know a working pet formula ? What did you use for your own SMN modelisation ?


      1. Try these. I haven’t an idea why I never posted these on my blog:

        Ifrit Action Damage:

        Ifrit AA damage:

        I have have muddled those two formulas around.

        Garuda Action Damage:

        All formulas are divided by 3 to get the DPS.

        Also, base Crit Chancefor Formulas is 4.5%. The scalar is the same as players, so for Crit chance it’ll be ((CRT-354)/(858*5)+0.045

        And Crit damage os +0.45

        Tell me if there’s any issues. If there are, send me all of your summoner damage points, formatted in a neat table showing stats used as well as non-crit min/max damage values and I’ll have a go at recalculating.


  29. Cadu says:

    Hey dervy, i would like to know how i could help with astrologian stats weights.
    I’m main healer as diurnal sect and can do arround 900-1100 dps at a9s or a11s.
    I’m prioritising crit for healing and dpsing; thogh was looking to my fflogs and there’s saying that arround 53-54% from my dps is coming from DoTs; then i was guessing priorising Spell Speed than Determination and keeping stacking crit for healing.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s