Relics Xtimes damage proc rate testings

[ringthree]

Going along with the theme of testing methodology, you have to always ask yourself is your data collection actually feasible. That is, can you actually get enough adequate data using this method? Because you are dealing with a relatively low "success" on relic proc, you are going to require a pretty large number of attempts.

You can easily get thousands of swings with parses with little to no effort. You have to spend hours, even days, collecting this on bunnies. One is just not feasible - you will be hard pressed to find people with relics that are willing to do this to that extent. Hitting bunnies does indeed control for more possible confounding variables, but it's not feasible so it's not a good test. Parse data controls for less potential confounding variables, but there is some logic behind not controlling for said variables.

Did you just say that while hitting bunnies is better it's not feasible. You know what isn't feasible, getting two bards and a mage to not have to go to the bathroom for more than 20 minutes at a time.

Killing bunnies is actually a lot more feasible. It doesn't require anyone else, and is equal to about 80% haste when it comes to testing speed (see the previous discussion on this) while eliminating random variables.

I think the word you are looking for is "boring". But then again, a lot of people think meriting is boring too. /shrug

Sometimes you have to sacrifice precision for practicality.

That's an interesting comment coming from someone that tested the enmity of Flee.

[Thorny]

Did you just say that while hitting bunnies is better it's not feasible. You know what isn't feasible, getting two bards and a mage to not have to go to the bathroom for more than 20 minutes at a time.

Hastes last 2 minutes after you put 3 of them up back to back. Marches last 2 minutes if you put them up back to back. You're female or doing it wrong if it takes longer than that to piss.

Killing bunnies is actually a lot more feasible. It doesn't require anyone else, and is equal to about 80% haste when it comes to testing speed (see the previous discussion on this) while eliminating random variables.

Half your time engaged is a very conservative estimate for bunnies, that aside I'll keep in mind that you love wasting time there next time we find something that needs rabbit testing.

I think the word you are looking for is "boring". But then again, a lot of people think meriting is boring too. /shrug

You walk away from merits with something more than the knowledge you've provided data for the community.

[FailureMidgard]

just got an idea: i know an apoc and a mandau users, if i could get meripos with them and parse them that would work ?

I would strongly suggest against using this method for mandau.

1. Solo SA/TA.
2. Offhand crits (could half crit #, but that's meh).

[Kaeko]

Did you just say that while hitting bunnies is better it's not feasible. You know what isn't feasible, getting two bards and a mage to not have to go to the bathroom for more than 20 minutes at a time.

Killing bunnies is actually a lot more feasible. It doesn't require anyone else, and is equal to about 80% haste when it comes to testing speed (see the previous discussion on this) while eliminating random variables.

I think the word you are looking for is "boring". But then again, a lot of people think meriting is boring too. /shrug



That's an interesting comment coming from someone that tested the enmity of Flee.

What is not feasible about this is actually convincing people with these rare items to perform such tests for you. The fact that it is boring prevents you from getting their help to the point of attaining reasonable sizes. This is again, an issue of the collector. Just because you are willing to do it doesn't mean others are. Those that would be willing to go to such great lengths for a single value are few and far between, and finding one with the relic you want to test is much harder.

There is just an element of common sense in this. You rarely get to perform the "perfect" test just due to time and resource limitations. That's why developing an "easy" test such as the one the OP has devised is so important.

Something like this is almost completely different that enmity testing. This requires large sample size - mine did not. Hence, it's perfectly OK to be anal about things since I only have to perform the test 1-2 times. Here, you're looking for large sample sizes, and hence, time and resource limitations play a huge factor. In addition, I didn't need a rare item to perform anything. Every situation is different and you just have to know when something probably isn't going to work out.

And going along with the idea that you have to sacrifice precision sometimes, I have in fact sacrificed a good amount of precision in my testing, namely for VE. There are 2 types of VE tests - one with tremendous precision but takes a long time, and another with maybe +/- 10 VE but can be done in seconds. Guess which I chose to use for maybe 99% of the values we use now? Margins of error are always going to be there - it's just a matter of what is acceptable to you.

EDIT: sorry for derailing.

[RyaWHM]

Ringthree: all I saw was...

me go WHM

me go WHM too

Masamune: great thread idea, I'm really excited to see how this turns out. I only wish I could help test Quick, someone gimme 30 m.silver!

[Izzy]

I had a hard enough time doing Mystic Boon testing with this shit. So many poor Bumblebees died... Rya seems like she has more patience than I do haha.

[ringthree]

Hastes last 2 minutes after you put 3 of them up back to back. Marches last 2 minutes if you put them up back to back. You're female or doing it wrong if it takes longer than that to piss.


Half your time engaged is a very conservative estimate for bunnies, that aside I'll keep in mind that you love wasting time there next time we find something that needs rabbit testing.


You walk away from merits with something more than the knowledge you've provided data for the community.

Lisa: It's just a joke.
Homer: Oh, I get it! I get jokes!

[ringthree]

What is not feasible about this is actually convincing people with these rare items to perform such tests for you.

Por moi! Why I do declare! *blush*

The fact that it is boring prevents you from getting their help to the point of attaining reasonable sizes. This is again, an issue of the collector. Just because you are willing to do it doesn't mean others are. Those that would be willing to go to such great lengths for a single value are few and far between, and finding one with the relic you want to test is much harder.

You give me a relic, I will test it until I am blue in the face.

There is just an element of common sense in this. You rarely get to perform the "perfect" test just due to time and resource limitations. That's why developing an "easy" test such as the one the OP has devised is so important.

Oh that bunny! It's going to be a tricky kill, but I am doing it for the greater good! *wipes brow*

Something like this is almost completely different that enmity testing. This requires large sample size - mine did not. Hence, it's perfectly OK to be anal about things since I only have to perform the test 1-2 times. Here, you're looking for large sample sizes, and hence, time and resource limitations play a huge factor. In addition, I didn't need a rare item to perform anything. Every situation is different and you just have to know when something probably isn't going to work out.

Dude, you tested the enmity of Flee... come on... Just sayin'!

And going along with the idea that you have to sacrifice precision sometimes, I have in fact sacrificed a good amount of precision in my testing, namely for VE. There are 2 types of VE tests - one with tremendous precision but takes a long time, and another with maybe +/- 10 VE but can be done in seconds. Guess which I chose to use for maybe 99% of the values we use now? Margins of error are always going to be there - it's just a matter of what is acceptable to you.

EDIT: sorry for derailing.

Dude... seriously... dude...

Double your derail, double your fun.

[Thorny]

?

[cpu]

I'll handle horn testing.

[Shamaya]

I would strongly suggest against using this method for mandau.

1. Solo SA/TA.
2. Offhand crits (could half crit #, but that's meh).

Agreed. It's very unlikely that I'd go through the effort of doing this /war in a meripo and not use sa/ta (though thinking about it I suppose my lowest possible sa/ta should significantly enough outdo my highest possible standard 3x crit; and messing w/ the calculator, my lowest 3x proc sa/ta should be significantly different from my highest normal sa/ta.. but something feels amiss)

As for me, I'd rather go to robber crabs with an alt char to dispel (or some alternative) than worry about the meripo thing :/. I could at least kick back and listen to some podcasts or music while I do this and not worry about being gimp or messing something up.

I see the sense in the OP idea though for other weapons.

[Masamune]

Agreed FailureMidgard, indeed i didnot thought about offhand skewing #crits...

..and i don't recall any 1H relic owners not going dual wield besides Guttler and Mjollnir...

Then guess all 1H relic owners are better off to follow "conventional" way of "massacring bunnies"...

[Izzy]

..and i don't recall any 1H relic owners not going dual wield besides Guttler and Mjollnir...

wat?

[Hanayaka]

mmm ok Hanayaka, added your data, let's hope you got 0 overlapping with normal crits.

From your example, looks like ranged gets a pDIF range lot narrower than melee's, but thats just eyeballing :s

EDIT: did you get any
"Critical! Hanayaka's ranged attack hits the Kindred [job] for [WTF] points of damage." over the 36 crits i see reported?

No I didn't have any wtfCriticals

[Vodou]

Yoichi normal 482 (16) crit 34 (5)

Anni normal 227 (12) crit 19 (2)

Amano (birds) normal 187 (7) crit 34 (2)

Amano (mamool) normal 352 (23) crit 77 (3)


So I get Yoichi 4.07%, annihilator 5.69% and amano 5.34%.

[Masamune]

erm... why didnot you follow the format suggested to post results Vodou ?

i mean can't exploit those values sorry...


Sidenote: i made some important reflexions on the subject and... i think i opened this testing too early. Without correct pDIFs under non capped conditions, which is the case of the OP, i'm 80% sure relic owners will miscount their relic procs from their parses... i should have think a bit before opening such test...

Also, other reason is i'm not sure that all relic owners knows that #hits displayed by KParser is actually #crits+#normals...

[Shinryuu]

Agreed FailureMidgard, indeed i didnot thought about offhand skewing #crits...

..and i don't recall any 1H relic owners not going dual wield besides Guttler and Mjollnir...

Then guess all 1H relic owners are better off to follow "conventional" way of "massacring bunnies"...

Actually, club and axe have the best offhand weapons available (KC, Temperance). I'd sooner singlewield Mandau (/WAR or whatever) or Excalibur (even /NIN, I guess, DW slows Excal down) than those two.

[Masamune]

np i was mistaken, im mostly worried about skewed results (wrong discrimation between Lowest relic procs and highest normal hits/crits) which will result in lower result than real (ie could get 5% while its actually 10%...)

[Vodou]

erm... why didnot you follow the format suggested to post results Vodou ?

i mean can't exploit those values sorry...


Sidenote: i made some important reflexions on the subject and... i think i opened this testing too early. Without correct pDIFs under non capped conditions, which is the case of the OP, i'm 80% sure relic owners will miscount their relic procs from their parses... i should have think a bit before opening such test...

Also, other reason is i'm not sure that all relic owners knows that #hits displayed by KParser is actually #crits+#normals...

I am dumb I guess. I thought if I saw:

180, 185, 178, 522

that I could interpret that as 4 hits of which 1 was a triple proc.

I guess you should exclude my results.

I could go out and smack Lesser Colibri with my Spharai, but I guess I'll wait until I can figure out how to determine what a triple proc looks like.


Ninja edit: I followed what was posted for Yoichi results, I'm assuming those are no good as well, or I totally do not understand what we are doing.

[Masamune]

... but I guess I'll wait until I can figure out how to determine what a triple proc looks like.

Yea that's a good thing to do, yu guessed the main prob of the testing. you can get a few pieces of answers @ main OP.