Bar spell tests

[Aleste]

So, I did some testing on bar-ailment spells, and offhand I'm feeling it's pretty inconclusive. Any ideas where would be a good idea to post the results so someone else can look at it and see if there's any correlation?

Edit:// Screw it, it'll go in here.

Following Byrth's post on the official forums regarding testing methods for bar-ailment spells
(http://forum.square-enix.com/ffxi/th...l=1#post278946)

Spoiler: show

"So, I have thought about testing this in the past and come up with a method (though I am too lazy to do it myself) that might help you answer some of the questions you posed.
* Take your character with a low base damage weapon and find a monster that casts Ice Spikes whose paralyze additional effect you are generally susceptible to.
* Hit it with your weapon and count the number of hits it takes to get paralyzed.
* Do this quite a few times. Like, start with ~100.
* Cast Barparalyze and repeat the test (keep your stats the same and note the enhancing magic).
* Calculate your average number of hits to proc Paralyze without and with Barparalyra, and use the Magic Hit Rate equations to figure out how much M Eva the spell gave you.

Problems:
Ideally you would use monsters that are a constant level, which is not easily possible unless they are EM. Barring that, you could compensate by collecting the same number of samples of each type from each enemy. So you do 3 rounds without barparalyze, 3 rounds with, monster is dead and you go to the next monster."

------------------------------------------------

All tests were done on 99WHM/SCH with 78+57 INT and 109+17 MND TP gear. (resist: silence job traits)
Cruor buffs were HP and MP only.
Atma choice was Vicissitude, Mounted Champion and Minikin Monstrocity (Regen+35, Refresh+10, Def+40, VIT+50, INT+50, MDB minor, ENM-, Enm decrease-).
Buffs used: Protect5, Shell5, Reraise, Light Arts, Aff: Solace, and the odd regen4.

Tested in Abys-Grauberg on Sensenmenn's (blm skeles up near azdaja) ice spikes and were waited on until no further depop/repops occured (to help ensure lowest level possible)
Care was taken to ensure that there was no-one around when testing, and when people did turn up, that any further testing that day was halted.
Testing was also done on pretty much any day that isn't iceday, nor lightsday (colossus's torques enhancing magic changes on lightsday)
Results were recorded by physically counting the number of hits required before paralysis occured, and timestamp (or is it bufftimer?) used to monitor the duration of the debuff.
The testing method used was the one that Byrth posted on the official forums.

Control Test
No bar-blizzara or bar-paralyzra was used for this test.

Spoiler: show

I expected this kind of distribution, however I did not expect the variance of duration to vary so much. Even when compared on a single mob, clear splits between expected resisted durations were not observed.

Bar-ailment test
For this test bar-paralyzra was used, whilst bar-blizzara was not.
Bar-paralyzra was cast in 503 enhancing skill gear.

Spoiler: show

Distribution seems approximately the same although a larger sample size would probably help define the curve more. Whilst average duration only decreased by 2.22 seconds.


As it stands I'm feeling pretty inconclusive that bar-paralyzra is doing what it is supposed to do, and I'm not sure it's worthwhile collating a larger sample size. Is the testing method incorrect? Is there something I'm blatantly doing wrong?

Edit2:// Spoilered page-breaker results

[dasva]

Well part of the problem I think is paralyze is kinda odd on duration in the first place. Who knows maybe it actually only lowers chance to have it landed. I'd do a bar spell against something with more defined durations if you want to test if it effects duration.

[Motenten]

While you've got a plot of how many hits it takes to proc, what you actually want is a plot of the duration frequencies, for what you're trying to test. Once you see that distribution, you can see if it seems to move between with/without the bar spell. I noticed quite a few ~20-second durations in the second plot, and almost none in the first. Just because the average barely shifted doesn't mean nothing's happening.

[Ophannus]

I wouldn't do it using Ice Spikes. Ice Spikes may have a solid accuracy that ignores resistances. I can attest to being a 75DRG/37BLM and consistantly and frequently paralyzing VT-IT Mamools/Colibri with Ice Spikes. Hell even at 90/99 with /RDM Ice Spikes kicks in super frequently against things I solo that are IT in/out of abyssea whereas casting Paralyze would outright get resisted 99% of the time. For the same reason that certain merit abilities like Aura Steal ignoring Dispel resistances or Tachi Gekko silencing with 100% accuracy on everything except mobs with 100% silence immunity or Blade/Weapon/Shield Bash stunning 100% of the time except targets immune to stun, I think Ice Spikes bypass resistance and have a fixed % chance to proc paralyze with each time a target activates the spikes and is independent with nothing to do with enhancing skill, probably only enhanced by RDM relic legs or something. Also the potency on Ice Spikes seems static and I've seen paralyze proc ridiculously often on VT-IT things 3-4x in a row up to 9-10x on a mob with /RDM's spikes. Seems as potent, if not more potent than Paralyze 2.

On the other hand, with barsilence on against Worms in Dyna-Tavnazia, fighting for 2 hours, Sound Vaccuum was always evaded(510 enhancing RDM/DNC) when I forgot to recast it, it would land every time. I'm certain you'd need to redo these tests using some other method such as a mob that frequently casts silence.

All status resists and elemental resists rise each level at the rate of a C skill. They used to show in the equip menu but SE changed it so that the values are still there and rise each level but the menu would only show 0 unless you add more, So you can use that as a base for your study and assume at 99 a naked player without buffs and no resistance traits has 373 resistance vs all elements and status ailments.

[Aleste]

@Motenten I graphed it up for you.

Blue bar is the control test, red bar is the bar-ailment test.


Thoughts on graph:
I'd imagine with a larger sample size it would even out those large spikes, but I'm still debating whether or not the testing method is slightly flawed at this point in time before continuing.
I noticed that there is a bit of a shift towards the lower durations with bar-ailment up, although this is balanced out by a larger quantity of larger durations which is causing the average to remain approximately the same.
I figure with capped skill (assuming at this point that it does scale up to 500 skill, as hinted towards by Byrth in the OF thread) that the effect would be more clean cut.

@Ophannus
I had considered that as one of the potential reasons as to why my testing came back inconclusive... although I'm hesitant about doing any tests involving capped skill and silence as /sch gets the resist-silence trait and it's impossible to hit capped skill without /sch on whitemage. I suppose I could abyssea-burn up redmage though.

Another method I had considered was aggroing a low level (ice?) elemental on a low level mule (via magic aggro, so I shouldn't get hate from buffing the mule from my main) on the hopes that due to the lower level, the elemental will have fewer spells to possibly cast (lowest level one should have bliz1, frost, bind and para?). Though I'm still mulling it over.

[dasva]

Could get something that spams sleep or blind too.

[Raelia]

Look through some of the November '09 NMs for the element-specific-casters like Bastet. Find one decently low enough level that it won't nuke you dead and you could probably record the resist rate of everything it casts. I think most won't even come into melee range, but just spam spells at you for hours...