Equipment Haste Values

[Kirschy]

Twilight Belt (7%)
Migawari: Ichi = 1:51 (69-76/1024)
Migawari: Ichi +91/1024 (464 +1 Instr) = 1:41 (154-162 or 63-71/1024)
Migawari: Ichi +92/1024 (465 +1 Instr) = 1:40 (163-171 or 71-79/1024) ∴ 71/1024

Iga Hakama +2 (7%)
Migawari: Ichi = 1:51 (69-76/1024)
Migawari: Ichi +91/1024 (464 +1 Instr) = 1:41 (154-162 or 63-71/1024)
Migawari: Ichi +92/1024 (465 +1 Instr) = 1:40 (163-171 or 71-79/1024) ∴ 71/1024

Iga Zukin +2 (6%)
Migawari: Ichi = 1:52 (60-68/1024)
Migawari: Ichi +73/1024 (450 +0 Instr) = 1:44 (129-136 or 56-63/1024)
Migawari: Ichi +76/1024 (471 +0 Instr) = 1:43 (137-145 or 61-69/1024)
Migawari: Ichi +92/1024 (465 +1 Instr) = 1:42 (146-153 or 54-61/1024) ∴ 61/1024

Iga Kyahan +2 (4%)
Migawari: Ichi = 1:55 (35-42/1024)
Migawari: Ichi +79/1024 (491 +0 Instr) = 1:46 (111-119 or 32-40/1024)
Migawari: Ichi +80/1024 (492 +0 Instr) = 1:45 (120-128 or 40-48/1024) ∴ 40/1024

Tiercel Necklace (1%)
Migawari: Ichi = 1:58 9-17/1024)
Migawari: Ichi +92/1024 (465 +1 Instr) = 1:48 (94-102 or 2-10/1024)
Migawari: Ichi +93/1024 (478 +1 Instr) = 1:47 (103-110 or 10-17/1024) ∴ 10/1024

Posting a PM from Mizunos. Most of this is verification of expected values, except for T Necklace being 10/1024. I must have messed up my earlier testing which had it at 11/1024. Good stuffs to know~

[Foldypaws]

So if 6 & 7 % haste items are coming in at 61 & 71 rather than 60, it looks like it picks the largest fraction of 1024 that wouldn't go over the decimal value.

Anyone have solid testing on a 5% haste piece? By that theory it should be 51/1024, as that's less than 5%.

Anyone have data that would disprove this general theory? (i.e. a 6% haste piece that's actually 60, or a 4% that's 41?) Wiki lists W. Turban as 50 and Blitz ring as 11, but It had Tiercel as 11 as well, and that's apparently wrong, so...

(This affects how close my DRG is to the 256/1024 haste cap <_< )

I'll test that gear myself if I need to.

[Kirschy]

So if 6 & 7 % haste items are coming in at 61 & 71 rather than 60, it looks like it picks the largest fraction of 1024 that wouldn't go over the decimal value.

Anyone have solid testing on a 5% haste piece? By that theory it should be 51/1024, as that's less than 5%.

Anyone have data that would disprove this general theory? (i.e. a 6% haste piece that's actually 60, or a 4% that's 41?) Wiki lists W. Turban as 50 and Blitz ring as 11, but It had Tiercel as 11 as well, and that's apparently wrong, so...

(This affects how close my DRG is to the 256/1024 haste cap <_< )

I'll test that gear myself if I need to.

I've tested Turban and Byakko's pants at 50/1024 for sure. Velo Belt has been 61/1024 a while. I tested DRK AF3+2 helm to be 61/1024 also.

My original testing showed T Neck at 11/1024, but it's possible I was using an incorrect march tier. Mizuno's went through doing his own March Chart before running those tests, so I'm inclined to believe I'm the one who made the mistake. Which means there's a chance Blitz ring could be 10/1024 too. I guess I'll see if I can run a test on it real fast, if not maybe someone else can.

EDIT: Also to add, Black Belt is 122/1024.

[Kirschy]

Tested Blitz:

Victory March /w March+4 = 160/1024

If Blitz Ring is 10/1024 then: (1024-170)/1024 * 180 = 150s(2min30s)
If Blitz Ring is 11/1024 then: (1024-170)/1024 * 180 = 149s(2min29s)

Result: 149s (2:29)

This makes me fairly sure my original Blitz Ring tests were correct.

[Foldypaws]

Blaaah. Looks like gear is going to have to be tested individually then, there's no firm conversion. Well, expect DRG +2 hat tests soonish, when I can finagle a bard.

[Foldypaws]

Allright. Gear tested.

DRG Hat +2 = 61/1024
DRG Feet +2 = 51/1024
(Homam hands + feet + Victory +2 = 188, + feet = 238 (46 second RR) or 239 (45 second RR). Got 45 seconds twice. 57 seconds with feet only, confirming 51/1024)

Timarli Dastanas = 40/1024
Ace's Hose = 40/1024

DRG Feet +2 give me hopes that Goading is 51/1024 as well, though I don't have one yet to test.

[Masamune]

There is something i don't understand about those haste tests :

Where is the proof haste% values are actually expressed in /1024 format instead of FLOOR(InitialRecast*Haste%) ?
I just tried with Migawari test above and that method match o.O

[Kirschy]

Goading Belt:

NIN /w Usu Feet + Dusk+1 Hands + Goading Belt-
(1024-20-40-50)/1024*120= 107 sec
(1024-20-40-51)/1024*120= 106 sec

Result: 106 sec (1:46)

So yea, 51/1024 for Goading Belt. That's surprising, good call on the prediction.

(Just to double check I did Turban in place of goading. Result: 1:47, so my original tests seem to hold, Turban is 50/1024, Goading is 51/1024.)

@Masa:

There's no proof that all haste% values are expressed in /1024. Instead it's better to think of the /1024 system as a model to accurately predict recast values. So far the /1024 system has held up perfectly, and there's been no examples to contradict it. But to answer your question more thoroughly, Victory March tiers are the best proof for a /1024 system.

An example of why the printed haste can't be trusted is the one in this post where Turban and Goading belt both show 5% haste in the description, but two different recasts can be achieved when they're interchanged. Using a high recast spell like Dread Spikes makes it easier to see the small differences.

@Yugl, I should probably write something up in the Test Instruments thread, but recording hits via fraps or using time stamps is unfortunately fairly inaccurate. There's an error margin of about +/-1.5% which makes it a great way to confirm basic ideas like Delay Caps, and Delay Addition via JA or WS. But it's not a good tool for measuring precise values.

[Byrthnoth]

How would I go about testing Dancer and Warrior AF3+2's precise Haste values?

I think I have or can get all the relevant Haste options. If you write a comprehensive explanation, mind adding it to the test instrument page?