Relics Xtimes damage proc rate testings

[Masamune]

After a few researches on forums, and 2 PMs (@ Kirschy and Shamaya), it appears clearly that 0 test has been conducted correctly to hear something better than :

"Eyeballed ~X%"

Also, nothing tell us clearly if all relic weapons have same Xtimes dmg procrate.

So i'm proposing to you relic owners to better know your relic stat making it dealing wtf numbers at random times... by just opening your parses (if you kept them) and 2 formulas.


Methodology:

- How to discriminate between normal hits and relic enhanced hits: use critical data recorded in KParser "Offense details" tab, under criticals section. Indeed since normal crits caps @ pDIF 3.15 for 1H and 2H weapon (under capped conditions), a crited relic proc(even @ lowest pDIF) is above highest normal crit. (Credits to Kirschy for the method)

- formulas:

a. you need to know your base damage:
Base DMG = WeaponDMG + fSTR = WeaponDMG + (yourSTR - monsterVIT +4)/4
(WeaponDMG is the base dmg displayed on your relic description in game)
I can help in determining monsterVIT if needed.

b. Highest normal crit damage :
HighNormalCrit = BaseDMG * HighCappedCritPDIF = BaseDMG * 3.15

c. Lowest relic crit damage:
LowRelicCrit = floor(floor(BaseDMG*X) * 2.655), for 1H relics
LowRelicCrit = floor(floor(BaseDMG*Y) * 2.891), for 2H relics
(this is mainly to make sure there is indeed no overlapping between normal crits and relic proc'd ones)
Check wiki to get your relic corresponding X or Y damage multiplier.

- Using parses :
In "Offense Details" tab, scroll down to Criticals section, then search for HighNormalCrit calculated earlier (and additionally verify if LowRelicCrit > HighNormalCrit).
Once found, start counting from immediate damage value just below HighNormalCrit in the list, all until the highest Relic proc'd damage.
Then in "Offense" tab, under "Melee damage" section, checks the total number of crits displayed.

- Posting results here :
Please post in the following format

Relic:[Name of your Relic]
Monster fought: [Name of mob parsed]
Your Attack: [post the attack you had during the parse, this is to make sure you were indeed under capped conditions versus mob fought]
Data: [#RelicCrits/#TotalCrits]

- Important Notes:
pDIFs written above are applicable only under capped conditions(capped cRatio & fSTR), since they have been determined on lvl0 mobs.
Those pDIFs have been also tested only for 1H and 2H weapons, i'm sorry i don't know for Ranged weapons... so useless to post Yoichinoyumi/Annihilator data
If you notice an overlapping between HighNormalCrit > LowRelicCrit, just count from damage above HighNormalCrit (have to ignore values between), and say it in your post.
You can use several parses, but post their results separately.
THFs and DRKs, take care with SA/TA/Souleater'd damages skewing your parse.
If your parse contains more than one mob type/level(for example @ MMJSP camp), make sure to post results for each mob separately (different VITs).



Current results listing:

Amanomurakumo
Corwen: 10/147 = 6.8% +- 4%


Apocalypse
Kirschy: 83/990 = 8.38% +- 1.726%
Corwen: 8/118 = 6.78% +- 4.54%

Bravura


Excalibur


Gungnir


Guttler


Kikoku


Mandau


Mjollnir


Ragnarok


Spharaï
Vyvian: 34/1000=3.4% +-1.12%

Yoichinoyumi
Hanayaka: 15/495 = 3.03% +/- 1.51%




TOTAL
150/2750 = 5.45% +- 0.85% => Relic proc rate = [4.6%;6.3%]

[Hanayaka]

I didn't use the method you detailed above, but this is an old parse from a Dynamis - Xarcabard right after I got my Yoichinoyumi.

I used Logger to dump all my data into a .txt, then ran my own program to strip out damage totals. I looked at all of them above a certain range, then rescanned the log to validate that they were triple procs. A bit more cumbersome - but hopefully it helps.

Hanayaka's ranged attack hits the Kindred Beastmaster for 446 points of damage.
Hanayaka's ranged attack hits the Kindred Dark Knight for 348 points of damage.
Hanayaka's ranged attack hits the Kindred Warrior for 201 points of damage.
Hanayaka's ranged attack hits the Kindred Ninja for 504 points of damage.
Hanayaka's ranged attack hits the Vanguard Eye for 286 points of damage.
Hanayaka's ranged attack hits the Kindred Dark Knight for 382 points of damage.
Hanayaka's ranged attack hits Marquis Cimeries for 319 points of damage.
Hanayaka's ranged attack hits the Kindred Black Mage for 363 points of damage.
Hanayaka's ranged attack hits the Kindred White Mage for 267 points of damage.
Hanayaka's ranged attack hits the Kindred Bard for 301 points of damage.
Hanayaka's ranged attack hits the Kindred Red Mage for 366 points of damage.
Hanayaka's ranged attack hits the Kindred Paladin for 183 points of damage.
Hanayaka's ranged attack hits the Kindred Ninja for 380 points of damage.
Hanayaka's ranged attack hits the Kindred Summoner for 275 points of damage.
Hanayaka's ranged attack hits the Kindred Ranger for 397 points of damage.
Misses: 54
Hits: 495
Crits: 36
Relic Crits[Shown Above]: 15
Total: 549

Based on these totals, 15/495 or 3% of the time, Yoichinoyumi Triple procs.

I will do some more controlled testing when I get the chance, but perhaps this might help if someone posts a similar number.

[ringthree]

I don't think your testing scheme is a good one, no offense.

You are looking for a variable proc in a variable proc. Although Crits may be the easiest way to separate the relic hits from normal hits (I don't even know if that is true) it is not going tell you an accurate amount of relic procs because your crit chance could change from mob to mob.

Wouldn't it just be easier to go around and pummel level 0 mobs where all the mobs stats are going to be static and you will easily be able to tell the difference between the relic hits and normal hits without needing crits?

[Therin]

Wouldn't it just be easier to go around and pummel level 0 mobs where all the mobs stats are going to be static and you will easily be able to tell the difference between the relic hits and normal hits without needing crits?

Not easier, but way more accurate.

You could probably just use Lesser Colibri, honestly. I doubt there would be any variation in critical hit rate on those and your regular hits would probably be static enough to use.

[Ratatapa]

you forgot claustrum

[Shamaya]

It's pretty easy to tell a triple just based off of normal hits as well. Hmm, idk. I really can't be fucked to mess with this atm, as much as I'd like to >,<. I'd prefer something that did save the log. Then calculate pDif for w/e mobs, or otherwise use another method to tell,and then just find out what my max non-relic hit should be. Then just look through the log and highlight or tally every time I see a triple proc *shrug*

[Murdock]

you forgot claustrum

/clap

[Masamune]

I didn't use the method you detailed above, but this is an old parse from a Dynamis - Xarcabard right after I got my Yoichinoyumi.

I used Logger to dump all my data into a .txt, then ran my own program to strip out damage totals. I looked at all of them above a certain range, then rescanned the log to validate that they were triple procs. A bit more cumbersome - but hopefully it helps.



Based on these totals, 15/495 or 3% of the time, Yoichinoyumi Triple procs.

I will do some more controlled testing when I get the chance, but perhaps this might help if someone posts a similar number.

The biggest problem with this method is you don't know how to accurately discriminate relic procs from normal and especially normal crit hits. at least yu didnot explain with which pDIFs yu used to scan for that "certain range".
On top of that, i doubt you were under capped conditions on DynaX mobs, which i really can't calculate their VIT and defense.
So sorry, unexploitable data.

Originally Posted by ringthree
Wouldn't it just be easier to go around and pummel level 0 mobs where all the mobs stats are going to be static and you will easily be able to tell the difference between the relic hits and normal hits without needing crits?

Not easier, but way more accurate.

You could probably just use Lesser Colibri, honestly. I doubt there would be any variation in critical hit rate on those and your regular hits would probably be static enough to use.

For Therin, yes yu are right: ringthree idea is what i typed 1st, until i recalled most relic owners parses their epeen @ Gcoli camps usually, so i modified my post in a more general way. Also, from my bad experience when asking for help for RR crit rate testing, i learnt poeple almost immediately hit their browser back button as soon as they read something involving testing on lvl0 mobs... or they put lolcomments or trolls(cf claustrum here that starts already...).
But regarding Lesser Colibris, they are semi-good preys: the lvl65 one has higher stats than the lvl63&64 ones, so tester gonna have to be careful when posting his results, ie choose correct mob on top of KParser window.
And finally, im more and more thinking i should revert the post into what i typed originally, like Ringthree suggested: parse under capped conditions only ! (to be sure pDIF used to discriminate normal and relic procs are correct)

Your thoughts ?

EDIT: regarding question why i don't ask for normal relic procs but only the crited ones, is because Kirschy noticed she had overlapping ranges between these 2, but not for crits.

[Hanayaka]

The biggest problem with this method is you don't know how to accurately discriminate relic procs from normal and especially normal crit hits. at least yu didnot explain with which pDIFs yu used to scan for that "certain range".
On top of that, i doubt you were under capped conditions on DynaX mobs, which i really can't calculate their VIT and defense.
So sorry, unexploitable data.

Ranged crit hits do not increase damage nearly as much as melee crit hits. We are talking the difference between a normal ranged attack (100), a ranged critical (125~145) [no where near a melee crit], and a ranged triple proc (290+++).

The way I determined whether things were triple procs was by scanning the log for unusually high damage attacks. After I got a list of them, I reopened the log to see what my damage looked like normally.

Example:
Hanayaka's ranged attack hits the Kindred Bard for 101 points of damage.
Hanayaka's ranged attack hits the Kindred Bard for 98 points of damage.
Hanayaka's ranged attack hits the Kindred Bard for 106 points of damage.
Critical hit! Hanayaka's ranged attack hits the Kindred Bard for 130 points of damage.
Hanayaka's ranged attack hits the Kindred Bard for 301 points of damage.
Hanayaka's ranged attack hits the Kindred Bard for 115 points of damage.

My parse results would have picked up the "Hanayaka's ranged attack hits the Kindred Bard for 301 points of damage." line, and then I scrolled through the log to verify that it was indeed a triple proc. It is very easy to see that that line was indeed a triple proc.

While I was not under capped conditions, I still think it is very easy to determine when my base damage suddenly spiked to three times.

[Kirschy]

The reason to use only your criticals, is to make it so none of us need to go out and mindlessly hit mobs for hours on end. Since most of us like to parse our merit parties (and hopefully save the results!) we can get data just through normal routine. Here's an example of of an exerpt from a coli camp:

  Melee Crits
      250:    1
      256:    1
      263:    1
      265:    1
      267:    1
      285:    1
      293:    4
      298:    1
      300:    2
      301:    4
      302:    3
      304:    2
      307:    1
      308:    2
      309:    1
      310:    2
      314:    1
      316:    2
      317:    1
      318:    2
      319:    1
      320:    3
      321:    1
      322:    1
      323:    1
      324:    1
      325:    2
      326:    2
      327:    2
      328:    2
      331:    1
      332:    2
 ^    335:    1
      336:    4
      337:    1
      338:    1
      339:    3
      340:    1
      341:    2
      342:    1
      344:    1
      348:    1
      349:    3
      350:    1
      351:    1
      352:    3
      353:    1
+     354:    5
      356:    4
      357:    1
      358:    2
      360:    3
      361:    4
      362:    1
      364:    1
      365:    1
      381:    1
      612:    1
      615:    1
      630:    1
      702:    1
      710:    1
      714:    1

There is a large discrepency between the 381 crit and 612 crit. Makes it easy to spot which were relic procs, and which were not. In this particular case there was 6 Relic procs/107 total = ~5.6%. With only 107 hits, there's a high margin of error, but if enough of us collected data on our relic proc rate we would probably get a fairly accurate proc rate fast. We'd also be able to see if the proc rate is the same across all weapons. I support Masa in collecting this data~

Just for reference, here is a small cut from the non-crit portion:

      225:    5
      226:    7
      227:    7
      228:    6
      229:    6
      231:    6
      232:    7
      233:    8
      234:    1
      235:    5
      236:   12
      237:    6
+     238:   15
      239:    7
      240:    2
+     241:   15
      242:    9
      243:    4
      244:    6
      245:    9
      246:    7
      247:   11
      248:    8
      249:    9
      250:    5
      251:    4
      252:   13
      253:    5
      254:    3
      255:    3
      256:    3
      257:    7
      258:    2
      259:    3
      260:    6
      261:    4
      262:    1
      263:    4
      264:    3
      266:    5
      267:    2
      268:    1
      269:    3
      270:    4
      272:    6
      273:    1
      274:    1
      275:    4
      276:    2
      277:    1
      278:    1
      283:    1
      290:    1
      301:    2
      302:    1
      304:    1
      309:    1
      313:    1
      315:    1
      323:    2
      327:    1
      331:    1
      335:    1
      337:    1
      341:    1
      348:    1
      350:    1
      358:    1
      359:    1
      365:    1
      366:    1
      367:    1
      373:    1
      375:    1
      381:    1
      389:    1
      390:    1
      394:    1
      396:    1
      397:    1
      398:    1
      399:    1
      403:    1

Min was 118, max was 543. Buffs don't always stay up, food gets eaten, lower relic procs overlap with higher regular hits. It's rough to deduce the proc rate on normal swings, very easy with crits. While a 10k sample size in controled conditions would be ideal, I can't see anyone doing it. This test seems like the next best thing~

[ringthree]

Not easier, but way more accurate.

You could probably just use Lesser Colibri, honestly. I doubt there would be any variation in critical hit rate on those and your regular hits would probably be static enough to use.

Sorry I used easier to mean an easier way to get accurate results. Though, some people enjoy clobbering bunnies for hours, so who am I to judge.

Even if you used Lesser Colibri you are still trying to get a variable out of a variable. If you are doing it from every hit (crit or otherwise because you would easily be able to tell on level 0 bunnies) then you are going to be exactly accurate to your sample size. If you use crits, even in a situation where you know the % of crits you are going to be massive decreasing the reliability of the tests.

[ringthree]

The reason to use only your criticals, is to make it so none of us need to go out and mindlessly hit mobs for hours on end. Since most of us like to parse our merit parties (and hopefully save the results!) we can get data just through normal routine.

I totally understand why you are doing it the way you are doing it, but the accuracy is going to be so poor there is going to be very little difference between your results and the situation we have now of people just eyeballing the number.

[Masamune]

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?

[Kirschy]

Ranged crit hits do not increase damage nearly as much as melee crit hits. We are talking the difference between a normal ranged attack (100), a ranged critical (125~145) [no where near a melee crit], and a ranged triple proc (290+++).

The way I determined whether things were triple procs was by scanning the log for unusually high damage attacks. After I got a list of them, I reopened the log to see what my damage looked like normally.

Example:
Hanayaka's ranged attack hits the Kindred Bard for 101 points of damage.
Hanayaka's ranged attack hits the Kindred Bard for 98 points of damage.
Hanayaka's ranged attack hits the Kindred Bard for 106 points of damage.
Critical hit! Hanayaka's ranged attack hits the Kindred Bard for 130 points of damage.
Hanayaka's ranged attack hits the Kindred Bard for 301 points of damage.
Hanayaka's ranged attack hits the Kindred Bard for 115 points of damage.

My parse results would have picked up the "Hanayaka's ranged attack hits the Kindred Bard for 301 points of damage." line, and then I scrolled through the log to verify that it was indeed a triple proc. It is very easy to see that that line was indeed a triple proc.

While I was not under capped conditions, I still think it is very easy to determine when my base damage suddenly spiked to three times.

This method works, but it's a lot of work to go through a big parse and check out each mob. With constant def down abilities like dia/angon being using during dynamis, not to mention buffs/mob range changing, and a big diversity in mobs, you need to hand pick the relic procs out. It's a daunting task~

[Masamune]

I totally understand why you are doing it the way you are doing it, but the accuracy is going to be so poor there is going to be very little difference between your results and the situation we have now of people just eyeballing the number.

That's the risk indeed Ringthree.... it's that or 0 contributor ?

[Thorny]

I totally understand why you are doing it the way you are doing it, but the accuracy is going to be so poor there is going to be very little difference between your results and the situation we have now of people just eyeballing the number.

It seems like you don't understand that crits and X procs are independant of each other. As sample size rises, it becomes irrelevant that you're just looking at crits instead of your whole set of attacks. You're trading off slower results in real time for the convenience of being able to do it in a merit party, personally I think it's a very smart way of doing it.

[Shinryuu]

What Thorny said. Merit a lot, parse a lot, look at parses, count numbers, done.

[BerenTebogo]

I think what RT isn't understanding is that for 2h weapons the range is very skewed, and even 1h weapons sometimes (I remember having to pause and figure out if xxx hit was a 3x or not with bluefagdag)

Crits make it easier it's just a bit more tedious I guess? @25% base crit and 8% xhit rate you're looking at 1/50 hits where (for the purposes of mathffff) where you're going to get crits instead of 2/25 for the xhit damage.

What (i believe) mas is suggesting is to test on lower level shit where we're 100% sure of cap on crit to maximize the frequency. To be honest if you don't already have a stockpile of parses it's an annoying task.

To contribute to the thread, rounded off, 7.7% overall 3x proc rate on Colibris (kikoku, no info on brovura yet) (taken from 3 parses, not big sample size) and I have no contributing information for 3x crit frequency.

I just woke up so if I got the purpose of the thread wrong just slap me.

[Thorny]

What (i believe) mas is suggesting is to test on lower level shit where we're 100% sure of cap on crit to maximize the frequency. To be honest if you don't already have a stockpile of parses it's an annoying task.

They're suggesting tests on lower level shit because you can parse normal hits as well as crits and get your data that much faster. The downside is that you get no benefit besides testing. Additionally, compared to colibri, you're getting fewer swings which closes the gap a bit(albeit still some).. you lose out on haste/marches/samba if you're soloing tiny mandragoras and have to run around reengaging far more than if you had a BRD pulling. In practice, parsing crits in a merit party will be pretty close to parsing all hits on lv1 mobs in speed at gathering data, but the merit party option benefits you more and is less boring.

[BerenTebogo]

Ok, that makes sense then. Still out of it ~_~