Yey! Another Haste Topic!

[VZX]

The chart kinda off me thinks (or it's just me who doesn't use image editing program to precisely track the y and x value)
Anyway
The last 2 dots on graph shouldn't exist no? Since we practically can't hit 95% haste mark. (and hence, the maximum Y value is 15.0588...)

[Kirschy]

I was an idoit and bought Excel 2007 a while back, and had trouble making the graph. The last point it displays is 95% - Attacking 20x. (And yea it's theory passed 93%ish) The chart is mostly for illustrative purposes. If Person A and person B use the same weapon, and Person B has 75% in Attack Delay Haste, Person B will attack 4x times for every 1 attack from Person A. But that disregards lots of other factors. The graph was just meant to show that Haste is not very impressive or beneficial at lower haste values. In situations where you may not be recieving Haste or March x2, options like Hasso, Desp Blows, or Catastrophe will not be as beneficial as they are when stacked in large values. In contrast at higher values of haste, it's very important. A mage who decides to cut out the 2nd march for another song, or neglects to haste when haste values are high can be detrimental.

[Valyana]

What I was trying to say is that Fast Cast recast reduction is applied either before or after haste recast reduction, you can't just add it into the Haste number. I haven't proven that myself, but that's what I hear from Red Mages, and it made sense with my Rostrum Pumps testing (below). It also works with your numbers:

If eq cap plus 118/1024 March plus Homam Pants just knock Dread Spikes recast down to 111 sec, applying fast cast separately we get

180*(1024-256-118)/1024*(1024-20)/1024 = 112.03
180*(1024-256-118)/1024*(1024-21)/1024 = 111.91

So Homam Pants are 21/1024 Fast Cast (2%). This is consistent with your other Homam Pants numbers, assuming the 3% Haste = 30/1024:

180*(1024-30)/1024*(1024-21)/1024 = 171.31, 9 sec off
180*(1024-30-50)/1024*(1024-21)/1024 = 162.70, 18 sec off
180*(1024-30-50-113)/1024*(1024-21)/1024 = 143.22, 37 sec off

I think this seems more likely than 3% Haste being different on Homam Hands vs. Pants.

My Rostrum Pumps testing:

Mazurka recast (base 60)
March(485, 110/1024)+Pumps: 53
March(486, 111/1024)+Pumps: 52

60*(1024-110)/1024*(1024-11)/1024 = 52.98 -> not 11/1024
60*(1024-111)/1024*(1024-9)/1024 = 53.02 -> not 9/1024

60*(1024-110)/1024*(1024-10)/1024 = 53.03
60*(1024-111)/1024*(1024-10)/1024 = 52.97

so Rostrum Pumps (and, I assume, Loq Earring) are 10/1024 (1%) Fast Cast.

[Gergall]

Great topic!

I bet that delay reductions (such as from that 2h grip and from the Haste Samba) are also applied either before or after haste. Is this already known?

[Valyana]

I bet that delay reductions (such as from that 2h grip and from the Haste Samba) are also applied either before or after haste. Is this already known?

Dual Wield and Sword Strap delay reduction, yes. I haven't been keeping up with DNC research, so I don't know how Haste Samba works. Does it reduce TP/hit like DW and Sword Strap do?

[Wizerd]

I bet that delay reductions (such as from that 2h grip and from the Haste Samba) are also applied either before or after haste. Is this already known?

Dual Wield and Sword Strap delay reduction, yes. I haven't been keeping up with DNC research, so I don't know how Haste Samba works. Does it reduce TP/hit like DW and Sword Strap do?

No.

[Valyana]

BTW, my testing with a +0 March horn doesn't match the chart in the OP; I got 79/1024 at 491 skill and 80/1024 at 492, with Faerie Piccolo giving an additional 32/1024. That's consistent with your +1 horn numbers always being 16/1024 lower than your +2 numbers.

Raw numbers (Victory March, Wind instrument, +0 March, Hasso Dread Spikes recast):

504 skill - 4:08
492 skill - 4:08
491 skill - 4:09
466 skill - 4:09

270*(1024-79)/1024 = 249.17
270*(1024-80)/1024 = 248.91

[Kirschy]

Your fast cast idea seems to be on target. I'll go back and mess around with the fast gear available to me, and modify my original post. You're also right about the march+0 tests. I made a mistake and assumed the wrong boundry.


req flute, +0 march, +41/1024(homam hands+blitz ring)
482 skill = 21 seconds off = 84/1024
485 skill = 21 seconds off = 84/1024
486 skill = 22 seconds off = 85/1024
500 skill = 22 seconds off = 85/1024

That made me believe march had been capped, but it was an error =/. I will fix up the chart and retest a few things to be certain. Thanks for catching me on that~

[Xrave]

Although probably not the case (and doesn't really matter), I would be interested to know if Speed Belt and Sonic Belt 6%'s are the same or different.

[RKenshin]

Very nice job on the tests Kirschy!

I can finally toss my Relic Scythe I've been carrying around for doing my own tests too. D: I've had one the last 2-3 months waiting for some free time, but you've just about covered everything I wanted to play around with. Only thing left is figuring out the DNC haste.

[Kirschy]

If I ever find a dnc /w Haste Samba, I won't hesitate to run another delay test. Would be nice to know how the ability works, if it stacks, and if there is a cap on Attack Delay.

Looking back on Valyana's calculations, depending on how/where you truncate/floor, the fast cast from homam could be 20/1024. You seem to be right on target in regards to it not stacking /w Haste, and being applied seperately. I think next I have the free time I'll play with fast cast from RDM and /RDM, including AF1 Head/AF2 Body/Loq Earring to see how it all stacks whatnot.

[Nameless]

Go go kirschy, shouldn't be that hard to find a dancer with it, so many are lvl 75 now a day. I have partied with a lvl 75 dancer, and did tried to do 100 chop with it at the end of our merit party, but I think the bird died too fast and the dancer didn't get the original idea of why I wanted to taste out haste samba, so I believe it didn't kick in til the bird is almost dead.

Then I went on to kill another bird, and at this time my last resort has already wore, so you see I am 15% haste short, but interestingly I was hitting at about once every sec, or slightly slower than that (from eyeballing). So I am thinking it might stack with the JA hastes, which makes sense cuz if it is indeed 5%, then I would be at 83.3% haste and on a 500 delay weapon that's 83.5 delay so about 1 hit per 1.4 sec, but I can't really tell if it stacked on top of 93.3% cuz it didn't kick in til LR wore.

[Skyylya]

I have nothing to add to this thread.

I just want to express how glad I am that people like you exist in this world that can, and do this kinda stuff.

Numbers make my head hurt

[Xanthu]

Figured this would be the place to bring it up. Just to check, at base singing/instrument skill/no merits of 450 Victory March would be 75/91/107? And does Advancing March do the same kind of scaling?

[willriker]

does this information still hold to be correct after this update, or has the decrease in cap haste truncated the graph listed by the OP? If so, what is the new haste cap?

[Tajin]

The haste cap is 80% now IIRC, and wow >.> its in the first page of adv. too ._.

[mallister]

hard cap of 80%, regardless of sources, seems to be the new ceiling.

and yes, there has been much wailing, gnashing of teeth, and flaming going on about the whole thing. depending on where you sit on the issue, it's a much needed adjustment, a pointless nerf, or a well deserved shot in the head to something that shouldn't have existed.

[Kirschy]

I don't have an unmeritted BRD to test base 75 skill /w. Advancing March caps at 96/1024 /w +2, which appears to be at base 75 skill. Victory March appears to have no cap, and adding more skill increases the amount of haste.

The new haste cap only applies to total attack speed gained from haste. For all intents and purposes you could just draw a line at the 80% mark. In relation to the chart, at capped haste, you'll attack 5 times as fast as if you had 0% haste.

Just to reiterate, the cap(80%) only applies to DRK /w Desperate Blows merits, and a MNK/SAM /w Hasso during SP. If you don't fall into that catagory you are completely unaffected.

[Quizno]

Actually, depending on your point of view, it could be all three.

[kenshyn]

Still no real total testing with Haste Samba? Really like to know where it fits in especially can it increase the cap or not.

some say 5% haste or 5% increase in attack speed of the weapon.