[Stormblood] [Dragoon Rotation] COME ON AND SLAM

[Frowny]

Yea, seems min-maxing right now gains very little. Using a new calculator that was released on reddit (holy numbers!), shifting 250 DH and 250 DET to crit is only a .22% increase. Still messing around a bit but the 4S chestpiece, if you're shooting for BiS, seems like the good choice.

[aleph_null]

Slightly off-topic, Momo recently posted a dragoon guide which covers a lot of optimization tips, including fight specific uptime optimizations, which I found pretty helpful: https://youtu.be/E-tWGBZZ-5E. I hope Sunny doesn't mind me posting this here, since this thread is the most active 4.x dragoon discussion thread.

[SunnyHirose]

I think it is on topic, thanks for sharing. It is pretty spot-on from a "this is what you do if you are Robocop" perspective. Which is the complete opposite angle from what I'm trying to take, of course; the further your and your raid group's performance is from Momo's (it is a very skilled melee player and a very skilled team), the less I feel that should be the concern. So surely I've made myself a laughingstock in some Discord chat or two, but when it comes to very different recommendations, it's really not about whether the legit speedrunner or some sweaty goonlord on the OF with too many spreadsheets open (and isn't raiding this season) is more credible. It's different goals.

For instance, my sample opener is not there to take you to the top of FFLogs or prevent any awkwardness; it's there for learning the job. So you get to the unavoidable and common BfB -> FT/WC check -> Geirskogul -> Mirage check -> Nastrond sequence almost ASAP, instead of hesitating about whether such-and-such really the best place to BfB (but note the guide never says that, it is rather my hope that one tries it out and starts internalizing that sequence without my input). And where Jump gets awkward, it's at that time you're staring the actual cost in the face, rather than eating it in the opener. Hiding it can be optimal in many situations, in fact, but from the emphasis I'm trying to place, I file that under "clever gambit", not "should do".

In economic terms, the "cost" of something is what you could have done with the same resource (time, for most DPS activity). But also in economic terms, your options to exercise are limited, and further limited by your awareness of those options. So I'm trying to look at it as a problem of calling the shots as they come, and that must give you different conclusions from the problem of fitting the most actions into a time window, aligning your buffs for one last hurrah as a fight closes, etc.

If I did have a criticism: Some of the stuff in the first section doesn't sound factually correct, e.g. you can just hold that Geirskogul for Dragon Sight, no biggie, but maybe there are different opener considerations that are being taken as a given there.

If I had praise: Momo has masterful uptime, observe it closely. I also seem to recall a reddit comment that was good enough to be a guide on its own, but I can't find it now and I may just be misattributing it in my head.

[aleph_null]

Thanks for the insights. Your guide does a really good job at helping people learn the job (e.g. general rotation, openers, cd/ogcd management), Momo's guide gave some good examples at how he optimizes his play further by taking into account the encounter designs. While his play probably requires masterful play from the rest of the group as well, I believe that by seeing what he does we can learn a bit about the thought process behind it, as in why he does this and not that. I agree that this kind of optimization is sorta robotic, and probably requires you to map every single gcd in each phase. It's definitely not for everyone, and it probably only matters when you're squeezing that last % of performance out of your job. Then again, personally what I consider as the best thing about these videos:

If I had praise: Momo has masterful uptime, observe it closely.

So much this. I'm impressed at how good he is at maintaining uptime. Personally I'm not that good, so if I try doing the same thing I'd most likely end up wiping the group or burdening the healers more often than not, but it's interesting to see how some of the really good players do things. Those "wow so you can do that?" moments are probably the main reason why I enjoy watching speedkill runs from Momo's group, Angered, Rushers, etc.

[Orphan_Benn]

Spent a couple of hours using the stat calculator that was put up on reddit , and the assumption that once we hit 3500 combined DT and Crit that Crit takes over as main secondary stat trying to perfect a BiS for us using Ariyala. This was my result - http://ffxiv.ariyala.com/1271B Hopefully this will save others time if they're in need of a BiS.

Things to note on it, looking at our current speeds we don't want to drop below 700 SS or else our rotation gets a bit messy. Persimmon Leaf Sushi will end up being our BiS food due to it's high crit value but currently Pork Kakuni is what most raiders will want to use for the extra Vit, that is until BiS gear starts to make up the Vit. Finally the body is the Genji due to the Crit limit being broken and thus the extra crit makes up for more damage than what the Lost Allagan gives.

Stat weights are close with the difference between builds being around 0.15% difference.

[knotwrite]

I know official stat weights aren't really a thing yet, but does anyone have any thoughts on pentamelded crafted chest vs. the 320 genta piece? That +244 skill speed on the crafted bit makes me uneasy (especially coupled with my +268 SS from my creation lance), but then again we're talking a free 64 direct hit from the extra melds...

I'm sure it's splitting hairs, here, but I want to be certain before I blow through a bunch of materia.

[SunnyHirose]

Tough call, knowing that stat weights are kinda bogus and even if they were valid it'd have to be relative to your existing gear. But basically the entire i320 BiS is pentamelded gear even if only by the tiniest bit. And the i330 chest, if melded, should beat the G6-pentamelded ornate chest by a razor's edge, if that's the next question.

Heck, the SS weight estimate I have, as low as it is, is probably higher than reality, since I averaged auto-attack's damage from arbitrarily-selected logs and then applied an estimated SS buff to it, which I also believe is higher than reality. I can't really prove it either way without the relevant formulas, though.

It should also be noted that Nemekh's sheet downgraded the estimated DH formula coefficient very slightly last week. I can't comment on how accurate it is or was, since I haven't had a look-see at the data and I don't know what sort of means was used to come by that estimate. While it's sufficient to work with, there are two grains of salt I offer:

• If linear regression is the primary tool used, the slope might be biased slightly low. This has been a problem in FFXIV formula estimation before, it's not exactly easy to escape.
• Also, to empirically measure a proc rate with a high level of confidence takes umpteen zillion trials. For example, if you take 20000 swings and get 1000 direct hits, you have not exactly shown that it is a 5.0% rate; you have shown that there is a 95% chance the rate is between 4.695% and 5.305%. That's kind of subtle if you're really hoping to show it's tiered at the tenths place: for the same level of confidence it would take ~184k swings to say "okay there's a 5% chance it's not 5.0%", and ~735k swings plus a similar amount at an adjacent tier to really go "yup, we've got two digits in this humpy-bumpy" (4.9501% to 5.0499%). That's assuming your observed rate is absurdly close to the real rate. Also, it takes more(!) swings the closer the theoretical rate gets to 50%. Naturally, this is tempered considerably by doing testing on more levels of the stat--the slope is a better thing to go by in many respects--but be aware the scope of the problem is a bit annoying like that. I have an aversion to hidden proc rates ._.

[knotwrite]

Thanks for the response!

I gotta say, I was pretty worried about jumping up to 1443 skill speed from around 930, but my numbers are not too shabby (just parsed 90+ on OS1 and OS2). Currently on OS3 it just feels like my cooldowns are aligning a little more cleanly than before. Purely anecdotal, I'm sure, but it seems I was all worked up over nothing.

[LiquidSwordz]

I thought for sure the lost allagan pants would be BiS thought.

[SunnyHirose]

Heavy Thrust Potency increased from 100 to 140.
Chaos Thrust Combo potency increased from 220 to 230.
Potency when executed from a target's rear reduced from 150 to 140.
Fang and Claw Potency increased from 200 to 250.
Wheeling Thrust Potency increased from 200 to 250.

Missed positionals are less harsh on us now, correct positional on uncombo'd CT does 10 less (wat).