Evasion Skill vs Evasion

[Kirschy]

Probably one of the longest standing questions in FFXI!

Decided to test it... What I found is that Abyssea is so fucking overcrowded that holding a group of mobs for 30 minutes is virtually impossible. Someone will always steal your mobs.

Anyways, I had at least one decent data set that I wanted to share, and I'm hoping the ease of the test will encourage others to perform similar testing. (Gathering a sample size of 10k took 15minutes.)

Testing was done by using +26 Evasion Skill. Grabbing a group of mobs, and letting them pound on me ~10,000 times and using a parser to determine evasion rate. I then swapped +26 Evasion Skill for +26 Evasion and reparsed using the same group of mobs.

Data:
+26 Evasion Skill: 10,296 - 6235 hit/4061 miss - 39.44% Evasion Rate
+26 Evasion: 4,730 hits - 2782 hit/1948 miss - 41.18% Evasion Rate

I don't know if the sample size is large enough, but this does seem to indicate that Evasion Skill = .9 Evasion. (26 Evasion Skill *.9 = 23 Evasion. 3 Evasion = 1.5% Evasion Rate.)

Obvious things to test:
-An Evasion Rate above 50% before using the Evasion Skill.
-Comparing a large amount of Evasion Skill+ vs Evasion by using a job with high skill such as NIN vs job like WAR using a lot of evasion+.
-AGI's affect on Evasion
-Mambos

Additional info about my testing:
-Used white mandy in Misar.
-Used MNK/WAR with 4 Evasion Skill merits.
-Used 3x Regen Atma.

[Sekundes]

I really have to question anyone stealing mobs from Kirschy...

A couple of things I thought of that might make a difference. Certain abyssea zones have stat boosts that change with how well the zone is doing right? Do any of those affect evasion? Also how about mob level? Considering how hard it is to find uncamped mobs that are good for this test, some might be leveled up partly?

As someone who's leveled nin recently but still has very little knowledge of it this kinda thing interests me a bit.

[Atoreis]

It looks like my test from a year or more ago that was showing same thing as Kirshy's is valid after all, and I got so much hate response lol...
I hope few more ppl will try it to confirm and try to check "Obvious things to test" that Kirshy wrote down.
It was a long time ago but I am sure that my evasion was above 50% when I tested it.
What mobs would you suggest to check on 2 different jobs to check "Comparing a large amount of Evasion Skill+ vs Evasion by using a job with high skill such as NIN vs job like WAR using a lot of evasion+." and be sure that they will stay at the same lvl when you go switch the job? I cant think of anything beside using aggressive mobs and aggro on NIN then use other char on BLM and sleep them then go switch job to PLD ( would be best in my example ) and AoE them for hate. Best would be chigoes but I know it's almost impossible to hold a group :/

[Xera]

I did a similar test a while back. Had around a 10k sample size per evasion type at +15 each.

Results were:
Evasion Skill: 5,292 @ 52.78%
Vanilla Evasion: 5,399 @ 52.90%
Control(5000~ sample): 2,343 @ 45.47%


I'll try to get a 10k sample of 372 Skill and 310 Skill + 62 Vanilla tomorrow. Tried to do it today, but all I ended up doing was kiting some '08 MNK for 10min. <_<

Spoiler: show

372 Skill:
Sample: 12,449
Evades: 8,004
64.294%~

http://dl.dropbox.com/u/18839017/ffx...0_02.00.28.png

[Karbuncle]

I'd help you but no parser i download works.

ever.

I've tried it all

[Byrthnoth]

These results are at the bare edge of potential significance, but I can definitely commiserate with the "people always come and try to take your mobs" because that's what has messed up every test like this that I've run too. Nothing is more frustrating than getting a nice 10k sample control and having some tard Fell Cleave your crowd because "he was almost out of time." Freaking A man, it isn't like you need to build light in Heroes zones. Just use stones! But anyway.

I will try to finish my version of the test tonight. I kept my gear constant on RDM/SCH (>20%) while testing some DA/TA Atmas against EM black mandies in Attohwa. The idea was to test Evasion atma next.

Also, as far as I've seen, you get variable Cruor buffs in Heroes/Scars zones, but they're constant once you put them on. So if you get +40 when you enter the zone, it's +40 until you leave.

KParser: http://code.google.com/p/kparser/

[Sylvrdragon]

Black mandies in Altep are popular AOE targets for BLU and Aeolian Edge. If you pull them, you're gonna have to drag them to some out-of-the-way spot that nobody is ever going to look.

I think Kirchy's white mandies in Misa are the best option if it must be Mandies.

[Xera]

Puks@Aby-La Theine [Easy Prey]


372 Skill (NIN/WAR 354 Base + 8 Merits + 10 gear), AGI: 86+40(124):
Sample: 12,449 Hit/Miss: 4,445/8,004
64.294%~

310 Skill (WAR/NIN Base 310) + 62 Vanilla Evasion, AGI: 76+50(124):
Sample: 15,364 Hit/Miss: 4,803/10,561
68.738%~

[Atoreis]

Nice job Xera.

[Byrthnoth]

How did you hold the monsters while you changed jobs?

[Obsidian]

How did you hold the monsters while you changed jobs?

He was on NIN, a friend was on WAR at the same time. They each had their own group of mobs.

[Byrthnoth]

Eh, whatever. Close enough for me. xD

It would have been nicer if it had been the same monsters. As it is, there could have been a difference in the average level between the two groups. There aren't enough puks to assume that it's evenly distributed really, even if you pulled all of them and each held 50%.

[LinktheDeme]

I was always under the impression that evasion skill did not really help for high level monsters and evasion + did. Maybe it was just placebo but on sky God's back when the cap was 75 evasion skill didn't do nearly as much as evasion + did. My logic was that the level difference was so vast that the floor for evasion skill was so low that improvements were negligible. However evasion + always seemed to give some sort of improvement (I was also under the impression that evasion+'s capped off at some point.).

I'm really bad at explaining things but I guess i'm asking if it is possible to test that if the monsters accuracy/level is so high that it floors your evasion below 5% (But it will be 5% since that is the lowest it can go iirc) Will evasion skill increase your evasion if it is still below 5% or continue until 5%. In comparison to evasion + where if your evasion is below 5% and that it would be 5% still with evasion skill + will evasion + increase it?

I apologize for shitty wording, I'm pretty terrible at flowing thoughts together with the ganja <_<

[Spira]

Really interesting stuff. It's a debate that I've always slights answers for because I quite enjoy Eva solo tanking stuff..
If these preliminary tests end up holding water for most situations, then the way people have been gearing eva has been all wrong (and the prices of gear will change accordingly)
It'd also mean that naked, PUP is the second most evasive job in the game due to it having higher Eva bonus traits than all jobs with higher skill, bar THF

[Karbuncle]

I've been gearing Straight Evasion from the beginning outside of my Boxer's mantle , I don't think my set will change much. but i can see some items (like the new EVA capes) going up in price because of this.

then again we may just be over-estimating how many people will care and/or educate themselves on these findings. need to spread the word!

[Karbuncle]

I've been gearing Straight Evasion from the beginning outside of my Boxer's mantle , I don't think my set will change much. but i can see some items (like the new EVA capes) going up in price because of this.

then again we may just be over-estimating how many people will care and/or educate themselves on these findings. need to spread the word!

[Atoreis]

Really interesting stuff. It's a debate that I've always slights answers for because I quite enjoy Eva solo tanking stuff..
If these preliminary tests end up holding water for most situations, then the way people have been gearing eva has been all wrong (and the prices of gear will change accordingly)
It'd also mean that naked, PUP is the second most evasive job in the game due to it having higher Eva bonus traits than all jobs with higher skill, bar THF

Except maybe You didn't know this but that's pretty much obvious...
Tell me examples of items that ppl were wrong gearing with, please?
It is known from years that evasion and evasion skill were close. Half ppl were saying (me included) that eva skill is 0.9eva and other were saying its 1:1.
PUP was known to be most evasive job after THF since forever +13 evasion was always better then 6 eva skill. That would be true even in 2006 when some ppl tho 1eva skill = 2 evasion.

[Weeks]

Someone needs to collect data on higher level mobs to be sure that evasion skill does not factor into the level correction of the hit rate formula. That's the only believable scenario where the simple skill = 0.9 evasion comparison might be altered.

[Nightfyre]

Really interesting stuff. It's a debate that I've always slights answers for because I quite enjoy Eva solo tanking stuff..
If these preliminary tests end up holding water for most situations, then the way people have been gearing eva has been all wrong (and the prices of gear will change accordingly)
It'd also mean that naked, PUP is the second most evasive job in the game due to it having higher Eva bonus traits than all jobs with higher skill, bar THF

None of my gear changes as a result of this, and I don't think it ever would have. The only people this would change anything for are the people who made wild, baseless claims about skill being overwhelmingly superior to straight evasion+. The only slot I can think of that it has any impact on now is Boxer's vs Fugacity/Blur, and that was generally decided more by whether you wanted additional parry skill or maximum eva (Fugacity)/AGI for TP feed anyway.

[Kirschy]

Sounds like .9:1 ratio is holding same weight. Is interesting stuffs. I plan to get some get extra testing done this weekend to fill in some more holes.

Just to chime in to what a few people have already said, none of this data is game changing. Most of the evasion skill vs evasion options were few and fairly easy to pick. But that doesn't change anything... it's nice to see to see some data to back some long standing theories. Evasion has always been one of those mysteries with more myths and bad experiments than good data.