how much spellspeed do i want for blm?

im at 2370. i feel like i should drop that for more crit cuz i dont see that big f4 crit that often. my crit is curently at 2262

Quote:

---

Originally Posted by ErzaScarlet77

im at 2370. i feel like i should drop that for more crit cuz i dont see that big f4 crit that often. my crit is curently at 2262

---

It’s based on your rotation and preferences for mobility but 2370 is definitely very high. I run at 1650 with no issues.

As much or as little as you like. Black Mages always get the full benefit of Spellspeed, allowing it to be a matter of taste.

If you wanna go fast, go fast. If you wanna go slow, go slow.

I am at about 2700 myself, it's all personal preference.

Quote:

---

Originally Posted by Kabooa

As much or as little as you like. Black Mages always get the full benefit of Spellspeed, allowing it to be a matter of taste.

If you wanna go fast, go fast. If you wanna go slow, go slow.

---

Just know that Crit takes a long time to ramp up to equality against SpS (without any DHit, it takes almost 4k stat for Crit to nominally outperform SpS, assuming you can make optimal use of the latter). (Heck, it takes almost 3.5k just to outperform Determination.) If you want to use Crit at all, it's best to ramp it and DHit as high as possible, which may not leave much room for SpS.

Given that I love speed, I'd probably just go Speed>DHit>Det>Crit until later tiers just allow for too much Crit not to swap entirely over to it.

Quote:

---

Originally Posted by Shurrikhan

Just know that Crit takes a long time to ramp up to equality against SpS (without any DHit, it takes almost 4k stat for Crit to nominally outperform SpS, assuming you can make optimal use of the latter). (Heck, it takes almost 3.5k just to outperform Determination.) If you want to use Crit at all, it's best to ramp it and DHit as high as possible, which may not leave much room for SpS.

Given that I love speed, I'd probably just go Speed>DHit>Det>Crit until later tiers just allow for too much Crit not to swap entirely over to it.

---

The "crit is worse than other stats until X tier" is based on bad information from Stormblood that got parroted at ShB's release. It only takes 17 points of crit to increase crit rate/damage by 0.1% (expected DPS value is a bit more difficult to calculate because of crit's exponential nature) compared to 6 points of Dhit to increase Dhit rate by 0.1% (24 points to reach an expected 0.1% DPS increase). Additionally, Determination takes 26 points to increase damage by 0.1%, and SpS/SkS take 26 points to reduce the GCD by 0.01 seconds (0.4% DPS if all damage is from GCDs, relative value goes down as more damage is from oGCDs or from sources unaffected by your class's chosen speed stat i.e. RDM's Enchanted Weaponskills or NIN's Ninjutsu).

So for BLM the expected DPS value of SpS far outweighs any other stat, but for classes like RDM, NIN, or MCH, crit quickly overtakes any other option thanks to its exponential nature.

All the spell speed all the time.

In addition to improving base dps, the lower your gcd is the less it feels like casting underwater and the easier it is to sneak in extra spells while doing mechs.

See Nemekh's post on the next page instead. I had apparently been, again, listening to the rigorously confirmed data.

Quote:

---

Originally Posted by Leidiriv

The "crit is worse than other stats until X tier" is based on bad information from Stormblood that got parroted at ShB's release. It only takes 17 points of crit to increase crit rate/damage by 0.1% (expected DPS value is a bit more difficult to calculate because of crit's exponential nature) compared to 6 points of Dhit to increase Dhit rate by 0.1% (24 points to reach an expected 0.1% DPS increase). Additionally, Determination takes 26 points to increase damage by 0.1%, and SpS/SkS take 26 points to reduce the GCD by 0.01 seconds (0.4% DPS if all damage is from GCDs, relative value goes down as more damage is from oGCDs or from sources unaffected by your class's chosen speed stat i.e. RDM's Enchanted Weaponskills or NIN's Ninjutsu).

So for BLM the expected DPS value of SpS far outweighs any other stat, but for classes like RDM, NIN, or MCH, crit quickly overtakes any other option thanks to its exponential nature.

---

Calculate it at 17 per .1% chance/increase then. Using that amount, the expected damage increase across 1000, 2000, 3000, and 4000 additional stat points (in isolation; no DHit helping Crit and vice versa) for each stat would be...

Critical Hit: 3.06, 6.91, 11.4, 16.8
Direct Hit: 4.15, 8.33, 12.5, 16.7
Determination: 4, 8, 12, 16

That said, Direct Hit and Crit Hit synergize, DHit helping Crit more than Crit helps DHit -- and more importantly buffs during burst which increase your Direct Hit or Critical Hit chances drastically increase the efficacy of Crit Hit's multiplier bonus -- so the actual values much more quickly outperform Det.

The nearest intervals to those points for SpS are 940/1041, 1955/2057, 2970/3072, and 3986/4088. We'll average from each end. Thundercloud will faintly reduce SpS's actual efficacy, but only by a trifle. We'll mark each with "~" to note this.

Spell Speed (BLM): ~4.38, ~8.93, ~13.9, ~19.3

We agree that SpS is going to be the most efficient stat for at least at couple tiers before Crit overtakes it, but Crit starting off weak is not just a result of misinformed parroting. In itself, is weak until one has sufficient DHit+Crit. Of course, one needs to account for how much Crit/DHit getting from raid buffs, too, and that's where it gets complicated.

In SB, it took, what... almost 4k total between the two for Crit to finally eclipse DHit as the most efficient stat?

Spell speed is def. king.
More spells in less time, easier to edge out mechanics preventing cast cancelling, and instant cast spells have less recovery time.

But it sure feels good seeing those direct crit foul, despair's and xenoglossy's.

I really, really wish there was a viable way to have different sets for casters at endgame. As a RDM main, I definitely do not want spell speed which forces me to use crit/dhit gear on my BLM when I play it. Not that it's bad like that, but I would definitely prefer a full SpS build for him without having to sacrifice ilevels.

Quote:

---

Originally Posted by Shurrikhan

<snip>

---

I really don't want to dig into this right now having not slept with the patch night in mind. But this is misinformation with regards to crit vs, especially with regards to the farce that was Stormblood era misinformation thanks to the likes of Omono and that one reddit thread that was a thorn in my side, that it took 14! months for the community as a whole to recover from that DH is king and crit is bad until x nonsense. It really has not helped that certain notable YouTubers decided to revive the statements in their 5.x era videos. It undoes so much of our hard work each time.

You need to look at absolute damage resultants for max damage potential. Do I really have to make a big disclaimer in bold on my website about it D: there's a reason why so many BiSes prioritise the stats in the manner they do, and early enough in the expansion instead of requiring gear from the end raid tiers. It was the same in Deltascape having Crit biased BiSes that I made for RDM and SMN back then which I still have listed on akhmorning.com in the historical gear sets page.

This expansion I am getting so tired of people stating the likes of "Crit's worthless until you get 3200" over and over. Enough with history repeating needlessly.

Anyway, with regards to BLM their high crit set is very competitive, but the max speed set enables some edge case rotation considerations that slower sets cannot, but those are concerns for people speedrunning more than the typical player. Go with what's comfy enough for you OP, however, if you want dedicated sets that have already been built to aim for here are what we list at The Balance:

"Fast/Min Crit BiS" (Mathematical Best)

"High Crit BiS" (Comparable w/ Crit Comp)

"SMN/RDM-Friendly BiS" (can use the SMN/RDM BiS w/o swapping melds)

Relative Gear Comparison sheet

Quote:

---

Originally Posted by Nemekh

snip

---

Correct me if I'm wrong, but I believe the relative damage spread sheet for gear does not account for rotational variances possible due to sps.

It seems like dh does have more weight than crit to a certain value, but it doesn't take until 3k+ to see crit overtaking dh. More like 1.6k ish, and depending on party composition, which is easily obtainable.

Though for blm of all classes, the difference in substats performance is relatively negligible, and you can go for whatever sps you find comfy and likely not see a massive change in outcome comparing against other max ilvl sets. +/- like... 100 dps.

Personally I enjoy higher sps for removing more RNG from your clears and making it much more reliable of an outcome. With crit, it either feels amazing and all of that 30% crit was on xeno, despair and f4, or terrible with every crit being on your f1s, b3s, f3s, and b4s. And your output can swing anywhere between those 2 scenarios. And my luck is kinda horrible.

The sheet's not a sim, but it tries to give a fair model to how stats impact dps given the pps model of blm. Furst would be better to go over the nuances behind it as it's not one I've worked on directly beyond all my early sub stat work for the game in general. Something like Yunalesca's BLM AI which came up with the megumin rotation is more along the lines of something that can better account for rotational variances. However, for the intents and purposes of seeing what gear options are stronger, especially without changing the GCD value, the sheet more than serves its purpose as a gear comparison tool.

Yes, DH can be better and more impactful earlier on, although being a linear function it does suffer from relative diminishing returns just as Det does. This is DH vs Crit that will always be correct, the full range and when it's within more realistic bounds. Translate those to points with a 3300 level mod and Lv. 80 values. Can solve with the surface diagram or just solve by plugging the numbers in and working it out that way.

The problem is when people use the data and calcs incorrectly to come up with those conclusions. The goal isn't about getting the best mileage out of your points, yes that matters, but what ultimately matters is what combination is achieving the greatest total damage possible. Speed muddies the water, but if all that's being looked at is Crit, DH and Det for DPS jobs? That's really easy to resolve.

It is easy to resolve ignoring speed. However blm being the only job (except maybe sam?) that can use speed well, it kind of forces it to be a factor. The gear this tier made it almost too easy to stack sps up and it scales so well, but most of that gear seemed to favor pairing it with dh than crit.

The sheet doesn't account for rotational gimmicks at certain speed values, and the min crit set still outperforms the crit set on average. That being said, while having a slightly lower average, people still like the crit set because the variance is higher. So if you go for a lot of runs in a fight you can fish for "the god run" more easily with the crit set.

The rotational gimmicks you can pull off with high enough sps are usually not that easy to do in actual fights without putting yourself at risk to drop eno.

In any case "crit outscales dh at X" is just an inherently flawed thought. You don't look to outperform a single stat, you try to maximize your total expected damage. That means you try to find the maximum of a 4d manifold which is bound by the total amount of substats. The problem is that gear choices dictate a major part of the substats and as such we aren't just looking at a circle around 0 with radius Σ(substats). Finding this maximum is non-trivial and we bruteforce a solution for gearsets by going through a huge part of the about 20^9*10^3 permutations in gear and melds. You could also just settle for an analytical solution since the problem is just an optmization problem (https://en.wikipedia.org/wiki/Lagrange_multiplier) but once again gear based substats and tiering makes this non trivial.

In any case, instead of blindly following magic numbers you are way better off just using the tools the community already created.