Nailing down Haste~

[SamanosukeShiva]

Was the haste from Hasso is 50/512 regardless if it's main job SAM or SJ SAM?

[aurik]

main/sub has no effect on hasso haste. And it's now being said to be 51

[draxyl]

Haven't had a chance to do more testing. But it does appear that collective haste% above 6% don't conform to the 5n pattern. GAX is prob 46/512, and I tested incorrectly... I'll check that again for sure. BB appears to be 61/512. 10% from Hasso appears to be 51/512 not 50/512 (Which makes the error of margin in the test almost nothing .) When I can verify some of the stuff, I'll edit my post.

Oh hay white plains! I'm from Orange County lol.

Washingtonville, OC ... represent.

[Kirschy]

I attempted some Hecatomb to see how it works. They act as negative haste values. As for fast cast, I'm only a DRK. RDM is my lowest job, so I've never had experience with "Fast Cast" outside of Homam Pants and Loq Earring. Homam Pants let me have a recast of 21 seconds on Ni with a ton of Haste stuffs, so I felt it was evidence enough that "Fast Cast" went above the 50% spell recast cap from eq/spells. I did eventualli plan on testing /w a 75 RDM. Af2 Body + AF1 Hat + Loq Earring should be capped Fast Cast and would give some insight. (Maybe there is a difference between "Fast Cast" traits and "Fast Cast" eq.)

And the Hasso tests were done on DRK/SAM.

[VZX]

Bump because there's something bothering me

Despite the validity of the exact haste value, Is there any absolute lowest delay?
How do you verify a weapon delay goes down to 30,40, or 50 delay? If one weapon delay goes down to 40 delay (by theory of 93.75% haste cap), how do you know it's not 30 delay? or how do you know it's not capped at 50 delay?

Sorry if the answer is implicitly implied on previous posts.

[Spekkio]

first of all, it's not that hard to tell how low the delay is. kirschy would simply count the number of frames between swings. with a known FPS, frames/FPS=seconds to swing. in addition, if we're reaching the speed at which we're losing data to the time between each frame, we could simply count how long it took to swing 10 times, add up all the frames between, then divide by 10 to get the number of frames for a single swing. once we have experimental evidence, we take the experimental values and compare them to our theoretical values from our haste model. if they match, our model is correct (technically only for this set of values, but by using a variety of numbers we can demonstrate a trend.) if they don't match, we go back to our theory and revise. it's just like 10th grade physics, doing all of those experiments with the carts to prove our models. does this answer your question?

[VZX]

first of all, it's not that hard to tell how low the delay is. kirschy would simply count the number of frames between swings. with a known FPS, frames/FPS=seconds to swing. in addition, if we're reaching the speed at which we're losing data to the time between each frame, we could simply count how long it took to swing 10 times, add up all the frames between, then divide by 10 to get the number of frames for a single swing. once we have experimental evidence, we take the experimental values and compare them to our theoretical values from our haste model. if they match, our model is correct (technically only for this set of values, but by using a variety of numbers we can demonstrate a trend.) if they don't match, we go back to our theory and revise. it's just like 10th grade physics, doing all of those experiments with the carts to prove our models. does this answer your question?

Verification method is not what I'm interested in. The answer, however, is.

Probably I didn't word it correctly:
"Did anyone do it?"

[Kirschy]

After a certain haste/delay the animation overlaps and it's very hard to get a clear reading. A lot of the delay/fraps tests I did were using the Relic Scythe (999 Delay). While I can get a very decent reading, I can't get very precise data while using all haste at once. (ie; Haste + Eq Haste + SV Marches + Hasso + DB + Catastrophe)

I hold no answer or proof that that there isn't some absolute delay minimum or actual haste cap. What I did test is how Marches, Spell Haste, Equipment Haste affect spell recast. I examined how they affect delay counting frames, and it appears to be similar. I tested Hasso and and desperate blows, alone and together, and their effects on top on large haste quantities.

Knowing the general idea and basics lets us build on it and theorize. Also when I did these tests my Desperate Blows merits were at level II (I opted for 3 Diabolic Eye, 1 Dark seal, 2 Desp Blows.) Since then I've changed to Desp Blows Lv 3, so I could re-run a test with relic scythe and get an approximate haste/delay value. Maybe it might shed some light into what you are looking for.

[Suiram]

After a certain haste/delay the animation overlaps and it's very hard to get a clear reading. A lot of the delay/fraps tests I did were using the Relic Scythe (999 Delay). While I can get a very decent reading, I can't get very precise data while using all haste at once. (ie; Haste + Eq Haste + SV Marches + Hasso + DB + Catastrophe)

Wouldn't it be possible to just do multiple attack rounds until you have two where you can get a clear idea of when it took place, and then divide the total number of frames by the number of intervening rounds? Or alternatively, would it be possible to eat a paralyze potion (which appears to be at or near 100% proc) and use occurrences of "Kirschy is paralyzed." in the chat log rather than animation frames?

[Kirschy]

The problem is that SV is 2hr, I can only fit in 2 sessions of Desp Blows. (With BRD JA + LR Recast Merits). I'll grab a BRD and see if I can get a good uncompressed fraps video to upload. As for Paralyze potion... I have no idea how that works... I never trusted the log that much, but it's worth a try. (In a similar sense can I turn around and get the "Mob is too far away." message and record the frequency of those?)

[Tythera]

I don't think those messages depend on delay as far as I know so that's not a good method of testing it. The paralyzed log might work, since the log inaccuracies might not matter over a long interval like 30s.

[Lordwafik]

The problem is that SV is 2hr, I can only fit in 2 sessions of Desp Blows. (With BRD JA + LR Recast Merits). I'll grab a BRD and see if I can get a good uncompressed fraps video to upload. As for Paralyze potion... I have no idea how that works... I never trusted the log that much, but it's worth a try. (In a similar sense can I turn around and get the "Mob is too far away." message and record the frequency of those?)

Para potions are 100%, every single action is paralyzed, you can't even paralyna yourself (I accidently used one while skilling alchemy, 10 mins later after repeat trying to paralyna myself, a taru paralynad me to save me ).

[Kirschy]

I think the last time I used a Paralyze potion was my "bright" idea to use a stack of those and a Flagellent's Rope Does the Paralyze message pop in relation to attack speed? There's no none up on Fenrir and my alchemy is only 60 so I can't make

[VZX]

Can a parser that reads register be a more reliable source than what you have on the screen?

[Heie]

"The mob is too far away" won't work for sure. I have been lvling rng and it just pops up like every 5 seconds even if I am in the middle of a range attack and my daggers delay for next swing doesn't go down at all while I am range attacking.

I am not 100% sure but I think para would work. You only get the message on your screen when it stops a action. Since people confirmed that it is 100% proc should go off everytime you try and swing. Since you are drk and aren't using dual weild you should have np with that.

GL with this and thanks for the info.

[Zarath]

Can a parser that reads register be a more reliable source than what you have on the screen?

If there isn't a way, timestamp works too for this purpose.

[Spekkio]

i'm not sure if timestamp would give precise millisecond resolution that kirschy needs to figure his(her) numbers, though i'd thought of that. you know how stun can show up as having worn in the log before it hits.

[Lyramion]

So after thinking a bit....


A DRK with Apocalypse could use....

WS Haste 10% (told counts towards GEAR cap)
Turban
Gloom Breast
Homan gand Legs Feet
Blitzring

for cap Gear Haste

Hastebelt (10%)
2x March
Haste

to cap out magic haste WITHOUT BRD 2hr

Hasso
3xDB


Should give you max Haste without having to use a BRD 2hr. Of course if you use that with Souleater and not DRK 2hr you kill yourself .... wouldnt you?

[Cream Soda]

Interesting read

[Seditedi]

So after thinking a bit....


A DRK with Apocalypse could use....

WS Haste 10% (told counts towards GEAR cap)
Turban
Gloom Breast
Homan gand Legs Feet
Blitzring

for cap Gear Haste

Hastebelt (10%)
2x March
Haste

to cap out magic haste WITHOUT BRD 2hr

Hasso
3xDB


Should give you max Haste without having to use a BRD 2hr. Of course if you use that with Souleater and not DRK 2hr you kill yourself .... wouldnt you?

Can't kill yourself with Souleater. The damage is calculated from % of your HP.