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