Spell interruption down cap and Aquaveil Tests

[Argettio]

wait so this is literally saying at 102% spell interrupt down, you won't get interrupted ever?

That's what the testing suggests.

Although it is only really mages that can reach the high a number.

[Lyramion]

Unpractical yet fun NIN math:


NIN/RDM

Nikkariaoe -25%
Willpower Torque -5%
Druid's Rope -10%
Yasha Hakama +1 3%
Yasha Jinpachi +1 6%
Yasha Samue +1 11%
Yasha Sune-Ate +1 6%
Yasha Tekko +1 4%
Magnetic Earring 8%

78% Spell Interuption

Caps with Aqua Veil, no merits needed, if you have merits can NQ all the Yasha stuff


> Invincible for ICHI !

[Starr]

Don't need to sub RDM with SCH doing Aquaveil either

[Spekkio]

this also has implications for tanking jobs in general if a sch is available to accession aquaveil on them. if the target is sufficiently mean (though i'm not sure what would qualify off the top of my head) it may justify the use of a sch's time/strategems to post av on the tanks. with a 5 minute duration, i believe, for aquaveil, keeping it up would not be a major burden and would allow at least semi-practically for a nin (or maybe a pld, i haven't added things up yet) to get close or reach the cap.

[Argettio]

PLD interrupt gear is a bit limited

Without losing TP you have:
Relic legs
Karasutengu Kogake
Magnetic Earring
Silver Obi +1
Knightly Earring
Willpower Torque

= 55%

Losing TP would get you another 10%.

[Siniroth]

Factor that in with basic observation that you should have had over the years playing FFXI to show that you get interrupted less after skilling up, and a reasonable conclusion can be drawn with very little doubt.

That's not proof and i've never experienced that.

I'll agree it's not proof, but wat? You've never experienced a lower interrupt rate on spells cast that you have a higher skill on? I know of the weakness of seeing something if you're looking for it but after so many years of playing I agree with Dawin, it's quite reasonable to conclude that higher skill causes less spell interruption (if you're not wearing a shit-ton of -interrupt gear, I've never really bothered with it so I can't tell you with, but your testing seems to cover that)

[Suiram]

Yinnyth was nice enough to let me kill him /w 1k needles and test Earth Staff + Lv 70 JSE Earring for BLM. (These work in /256ths btw.) Earth Staff: 51/256, JSE Earring: 76/256. So fixing my mistakes and re-doing my numbers:

/256 system:
(Estimating values of course.)

Club: (25%) 64/256
Shield: (10%) 25/256
Torque: (5%) 12/256
Earring: (8%) 20/256
Belt: (6%) 15/256
Feet: (15%) 38/256
Merits: (8%) 20/256

Total: (77%) 194/256(~75.78%)

Aquaveil would need to be at least: 62/256 (~24.21%)

I was looking at this yesterday when pchan first posted the gear he used. I started off assuming exactly the values of 256ths that you wrote, as well as 30/256 for the 12% pieces. The problem I observed is that it doesn't work out for all of the %s, where you'd want to be under 256 for 76 (with aquaveil), and 98~101%.

From this starting point, while we get the 102% gear at 258/256, the 101% gear is exactly 256/256, so there should have been no interrupts in that case either. Similarly, using the starting assumptions for the values of 256, the 76, 77, and 78% gear sets are 192, 194, and 196/256, which means that if Aquaveil is 64/256, they are at 256, 258, and 260, respectively.

So basically, there's a problem at 101%, and at 76% w/ Aquaveil. Now, normally, I would suggest that the issue with this is just that the value for one of the pieces is off by 1/256. But all of these values (aside from 25% which goes evenly) are reached by multiplying 256 by the appropriate % and rounding down already. So you'd think if there was a discrepancy it'd be caused by them rounding up. Like 10% of 256 is 25.6, so some damage taken pieces are 25/256 but occasionally you get 26/256, but never 24/256.

That's not to say it can't happen though; my prime suspect is karasutengu because there's no damage taken piece with 15% to compare to, and it's in both of the sets that cause a problem with the initial conjecture.

I'd like to see the following 76% build:
Club (25%), Club (25%), earring (8%), 10% belt (10%), merits (8%) + Aquaveil (whatever, probably 25%).

If there are no interrupts there, that would suggest that the issue is with the feet, provided that Aquaveil is in fact 25%. If not, we could do similar swapping to try to see where the problem is. If we want to confirm Aquaveil first, it might be useful to redo the 101% and 102% tests exactly, but replacing 1 club with aquaveil and see if there are still interrupts on 101 and none on 102.

Of course, if none of that works out then we could start looking at it maybe being 1024ths or something else altogether. Some day maybe can even figure out how much interrupt each level of Choral Roll gives. Might come in handy some day? lol


Anyway pchan, if you feel like doing more testing this would be great:

1: club + torque + earring + NQ body + 6% belt + karasutengu + Aquaveil (shouldn't interrupt if Aquaveil is 25%)
2: club + earring + NQ body + 10% belt + karasutengu + Aquaveil (should interrupt if Aquaveil is 25%)
3: club + club + earring + 10% belt + Aquaveil (depending on whether or not it interrupts will help in determining issue on karasutengu)

[Nermie]

I was looking at this yesterday when pchan first posted the gear he used. I started off assuming exactly the values of 256ths that you wrote, as well as 30/256 for the 12% pieces. The problem I observed is that it doesn't work out for all of the %s, where you'd want to be under 256 for 76 (with aquaveil), and 98~101%.

From this starting point, while we get the 102% gear at 258/256, the 101% gear is exactly 256/256, so there should have been no interrupts in that case either. Similarly, using the starting assumptions for the values of 256, the 76, 77, and 78% gear sets are 192, 194, and 196/256, which means that if Aquaveil is 64/256, they are at 256, 258, and 260, respectively.

So basically, there's a problem at 101%, and at 76% w/ Aquaveil. Now, normally, I would suggest that the issue with this is just that the value for one of the pieces is off by 1/256. But all of these values (aside from 25% which goes evenly) are reached by multiplying 256 by the appropriate % and rounding down already. So you'd think if there was a discrepancy it'd be caused by them rounding up. Like 10% of 256 is 25.6, so some damage taken pieces are 25/256 but occasionally you get 26/256, but never 24/256

Earlier I suggested a 8-bit system that uses the mantissa for a 128-basis implementation. Such a system would base everything on 128 and not 256. Under this paradigm the 101% Spell Interrupt Down items would result in 99.21875% spell interrupt down while the 102% Spell Interrupt Down would result in 100%.

101% Set
Club: 25% => (0.0100000) => 25% (32/128 )
Club: 25% => (0.0100000) => 25% (32/128 )
Magnetic: 8% => (0.0001010) => 7.8125% (10/128 )
RDM AF1 Body: 10% => (0.0001100) => 9.375% (12/128 )
Belt: 10% => (0.0001100) => 9.375% (12/128 )
Feet: 15% => (0.0010011) => 14.84375% (19/128 )
Merits: 8% => (0.0001010) => 7.8125% (10/128 )
Total: 101% => (0.1111111) => 99.21875% (127/128 )

102% Set
Club: 25% => (0.0100000) => 25% (32/128 )
Club: 25% => (0.0100000) => 25% (32/128 )
Torque: 5% => (0.0000110) => 4.6875% (6/128 )
Magnetic: 8% => (0.0001010) => 7.8125% (10/128 )
RDM AF1 Body: 10% => (0.0001100) => 9.375% (12/128 )
Belt: 6% => (0.0000111) => 5.46875% (7/128 )
Feet: 15% => (0.0010011) => 14.84375% (19/128 )
Merits: 8% => (0.0001010) => 7.8125% (10/128 )
Total: 102% => (1.0000000) => 100% (128/128 )

Yes, I am aware that it is generally believed that other variables are stored on a 1024 or 256 basis. I am (at this point) recommending this system for Aquaveil / Spell Interrupt only.

Yinnyth was nice enough to let me kill him /w 1k needles and test Earth Staff + Lv 70 JSE Earring for BLM. (These work in /256ths btw.) Earth Staff: 51/256, JSE Earring: 76/256. So fixing my mistakes and re-doing my numbers:

/256 system:
(Estimating values of course.)

Club: (25%) 64/256
Shield: (10%) 25/256
Torque: (5%) 12/256
Earring: (8%) 20/256
Belt: (6%) 15/256
Feet: (15%) 38/256
Merits: (8%) 20/256

Total: (77%) 194/256(~75.78%)

Aquaveil would need to be at least: 62/256 (~24.21%)

Using the 128-system I described before this set would be 75% Spell Interrupt down:

<Item>: <Defined %> => <8-bit binary representation> => <Resulting decimal due to truncation>
Club: 25% => (0.0100000) => 25%
Shield: 10% => (0.0001100) => 9.37500%
Torque: 5% => (0.0000110) => 4.68750%
Earring: 8% => (0.0001010) => 7.8125%
Belt: 6% => (0.0000111) => 5.46875%
Feet: 15% => (0.0010011) => 14.84375%
Merits: 8% => (0.0001010) => 7.8125%
Total (sum of 8-bit binary representation): 77% => (0.1100000) => 75%

Aquaveil being 25% makes much more sense now.

[Suiram]

The main issue with the above analysis is that merits are most likely the same value for each upgrade, rather than going by what the % would be and determining it from there. So in 128ths, a 2% spell interruption merit would be either 2/128 or 3/128, and 4/4 would be 8/128 or 12/128, both of which cause issues unless some of the values on the armor pieces are different, as well.

[Nermie]

Yeah I thought about that too but I'm assuming that the merits are added together before creating the binary representation.

2% => 1.56250% (2/128 )
4% => 3.90625% (5/128 )
6% => 5.46875% (7/128 )
8% => 7.81250% (10/128 )

I know it sounds silly but this game was programmed in 2000-2002 for the PS2?

[Suiram]

Yeah I thought about that too but I'm assuming that the merits are added together before creating the binary representation.

2% => 1.56250% (2/128 )
4% => 3.90625% (5/128 )
6% => 5.46875% (7/128 )
8% => 7.81250% (10/128 )

I know it sounds silly but this game was programmed in 2000-2002 for the PS2?

Right, I'm saying it doubt works that way, which then indirectly causes issues for the 128ths idea, at least as it's presented above.

Shell 5 works how I'm suggesting spell interruption likely does. It says in the merit help text (I'm fairly sure, not ingame atm) that it adds 1 per upgrade. And each upgrade gives 2/256, even though 5/5 would be 12/256 instead of 10, if it worked the other way.

My understanding is that Desperate Blows (and snapshot?) works in a similar way, but Kirschy can confirm this later hopefully (hopefulli? )

[Nermie]

I guess this could be tested by using 1 Merit, 2 Merits, 3 Merits, 4 Merits in Spell Interrupt with corresponding gear to reach 102% in each scenario.
(e.g. 3 Merits + RDM AF+1 Body vs. 4 Merits + RDM AF Body)

[Kirschy]

Right, I'm saying it doubt works that way, which then indirectly causes issues for the 128ths idea, at least as it's presented above.

Shell 5 works how I'm suggesting spell interruption likely does. It says in the merit help text (I'm fairly sure, not ingame atm) that it adds 1 per upgrade. And each upgrade gives 2/256, even though 5/5 would be 12/256 instead of 10, if it worked the other way.

My understanding is that Desperate Blows (and snapshot?) works in a similar way, but Kirschy can confirm this later hopefully (hopefulli? )

Testing Desperate Blows to precision is near impossible due to the margin of error in FRAPs testing But yea, I strongly believe the the merits would be the same per upgrade, like shell is.

I'm doubtful of the /128 system too, but it's best we keep an open mind until someone figures out a way for us to test this to some precision.

Onyx Cuisses have Damage Taken+14% -_- Plastron +1 has Damage Taken+15%... Anyone know someone /w one!? (Maybe Mawdac at KI would wanna test it ~.~)

[Nermie]

I understand your hesitation. It's tough to suggest that the system can be unique for each variable. The only reason I came up with 128 as a basis is because i tried all possibilities 2^7 through 2^10 and 2^7 (128 ) described the 102% versus 101% discrepancy best.

[Byrthnoth]

Well guys, look at it another way.


Skill affecting interrupt rate
Though his tests nowhere near had the sample size to prove this, look at the interrupt rate just 1% before he became uninterruptable. Utsusemi's interrupt rate is way over Stoneskin's, and he's on RDM/NIN. Stoneskin has more skill, makes sense, etc.

Another option (again)
At 76 + Aquaveil, he has a 33% interrupt rate on Utsusemi.
at 98%, he has a 50% interrupt rate on Utsusemi.
At 99%, he has an 18% interrupt rate on Utsusemi.
At 100%, he has a 5% interrupt rate on Utsusemi.
At 101%, he has a 36% interrupt rate on Utsusemi.
Severely weird, insufficient sample size, etc. Not really obviously approaching 0 though. It could be that, like pDIF, the spell interrupt rate is fed into a piecewise defined system of equations. Could be that this system of equations isn't defined for negative numbers, thus it returns 0 or an error that the system simplifies to 0.

You could probably make use of this with Spellcast macros as long as you're cool with losing Terra's staff.

[pchan]

don't do any interpretation on the% of interrupt below cap in my tests, there are various stuff to consider :

- sample size does matter a lot, with just 50 spells (=~300 hits) you have +/- 5% error and with 20 spells that's +/-8%
- you can be interrupted multiple times during a spell, but the chatlog will only count one obviously

edit : and the most obvious stuff , assuming you get interrupted on every single spell you cannot find more than 1/6 rate since you will find ~6 hit per spell and 1 interrupt..

[Max™]

Call me crazy cause I don't cast anything but whatever spells Pld (+ /Rdm) and Nin (+ /Nin) use... but on War/Nin I notice that being hit at different times during the cast appears to effect it in a way based on the skill.

On Nin I notice that I can be hit during a broader portion of Ichi.

On War/Nin:
[0=10=20=30=40=50=60=70=80=90=100]
-[Safe]----[Interrupted zone]----[Safe]-

On Nin/War:
[0=10=20=30=40=50=60=70=80=90=100]
----[Safe]---[Interrupted]---[Safe]-----

On Nin/Rdm + Yasha/Nikka/Aquaveil:
[0=10=20=30=40=50=60=70=80=90=100]
------[Safe]----[Bad]-----[Safe]-------

[Chen]

to RyaWHM, i'm 99% sure that if your main job has no skill in the spell

i.e.
nin/rdm casting SS
whm/nin casting utsu, you're going to be at the bottom cap (0 or 5 idk)

which would explain your earlier comment about whm/nin being interupted with aquaviel and low ninjustu skill against mobs at crags (i'm saying skill isn't a large factor, it's more about sub only spells)

apologies if i've missread your comments or this already been answered.

[Spekkio]

to RyaWHM, i'm 99% sure that if your main job has no skill in the spell

i.e.
nin/rdm casting SS
whm/nin casting utsu, you're going to be at the bottom cap (0 or 5 idk)

which would explain your earlier comment about whm/nin being interupted with aquaviel and low ninjustu skill against mobs at crags (i'm saying skill isn't a large factor, it's more about sub only spells)

apologies if i've missread your comments or this already been answered.

doubtful. otherwise rdm/whm would get rocked casting reraise, but i'd bet if you tried that, you'd get roughly as many through as on whm as rdm has native skill, despite not having reraise natively.

[Belgaer]

Homam legs are over 3% haste

I've never seen anything supporting this. o_O


Anyway, how is this proof that skill doesn't affect spell interruption? It doesn't directly add spell interruption rate down, I think, is what you mean. There is a difference. I believe skill influences where your spell interruption rate chance caps on the low end (IE: Like accuracy% can only go as low as 20%, spell interruption rate most likely has a point where it can't get any lower, and skill increases this.).