Nocturnus helm Proc Rate

[Kirschy]

Ran a quick test, here's what I found: (Note: Parser does not count Noct Absorbs, so I needed to go through my log file and count them manualli~)

DRK/WAR /w 41% PDT + Shadow Mantle + Pescatora + Defender.

1441 hits against me
72 Absorbs.
1513 Total hits + absorbs.

72/1513 = 4.76%

Of the 1441 hits, 81 proc'd 0 from mantle.

I never absorbed for 0, so it appeared as though mantle + helm could not proc at the same time~

I did not absorb any tp moves, but there were only 6 Big Scissors on me, so thats not conclusive of anything~

My avg damage taken was 16.02 damage. My average absorb was 32.75 damage. I feel when you absorb double the damage dealt. (This would be similar to Shadow Lord's Soma/Magic Walls which absorb 2x the damage you'd deal.)

[Nightfyre]

DRGs everywhere came buckets.

If you assume Shadow Mantle takes priority over Nocturnus Helm, 5% proc rate on the Mantle puts the absorb rate right at 5%. Probably within confidence interval anyway but meh. Thanks for testing!

[Khajit]

How does the proc rate on mantle interfere with the helm? I know you said it cant absorb for 0 but i mean which has priority etc?
Eg. 10% proc on each and wearing both means A procs 9% instead of 10% because proc B landed first.

[Kirschy]

I am not sure how they interact. My mantle tests showed it was ~ 1/16(6.25%). Mantle proc was lower during this test which makes me think there's some kind of interaction going on, *shrug. I didn't bother testing them separate, because I'll be using them both together majority of the time~

[miokomioko]

Chigoes!

[Foldypaws]

My avg damage taken was 16.02 damage. My average absorb was 32.75 damage. I feel when you absorb double the damage dealt. (This would be similar to Shadow Lord's Soma/Magic Walls which absorb 2x the damage you'd deal.)

Alternately, when you absorb damage, -pdt doesn't apply (as you were wearing 41% and have a low number of total absorbs, it should be within the confidence interval) And I think that matches up better with the magic absorb piece, which IIRC didn't seem to double heal.

[Spekkio]

Alternately, when you absorb damage, -pdt doesn't apply (as you were wearing 41% and have a low number of total absorbs, it should be within the confidence interval) And I think that matches up better with the magic absorb piece, which IIRC didn't seem to double heal.

i'd been asking in the other thread regarding if mdt/resist are rolled before or after absorb effect on mail. if you roll for absorb before damage mods, a resist tanked mob could almost be tanked afk.

[Kirschy]

Alternately, when you absorb damage, -pdt doesn't apply (as you were wearing 41% and have a low number of total absorbs, it should be within the confidence interval) And I think that matches up better with the magic absorb piece, which IIRC didn't seem to double heal.

That was another thought, yea~ I forgot to mention that I chose the 2x absorb theory because the highest absorb was 95, and the highest damage taken was 46 damage from a crit. It seemed way more likely that the Double Damage thing was occurring than ignoring pdt or defense. But yea, I don't know for sure~

Which reminds me to add. Everytime you absorb HP it look likes this:
[16:17:38]The Robber Crab hits Kirschy.
[16:17:38]Kirschy recovers 39 hit points!

It never shows a critical occuring, but there are "critical-like values" that deviate from the average.

[Onorgul]

You could get an odd number from doubling the damage so long as it was x.51 damage or above. Flooring the final result is generally the last step.

[Foldypaws]

You could get an odd number from doubling the damage so long as it was x.51 damage or above. Flooring the final result is generally the last step.

SE actually generally likes to floor it's values several times. I know that when calculating WS damage, the WS mods is applied to stats, then floored, then the level correction, then floored again, and then the pdif. (fTP may also be floored as well, I forget.)

That said, I don't think there's a floor involved at the "apply elemental staves" step, so if this x2 is similar to that x1.15, you could get odd numbers, yeah

[Nightfyre]

100% sure fTP doesn't floor. If it did, the first hit of Guillotine would do 0 damage every time and YGK would be much weaker, among other examples.

It's a good point though, SE likes to floor early and often.

[Foldypaws]

100% sure fTP doesn't floor. If it did, the first hit of Guillotine would do 0 damage every time and YGK would be much weaker, among other examples.

Er sorry, no I didn't mean the fTP value itself, I meant the damage value you get when you multiply D + WSC by fTP. (WSC itself is, and (D+WSC)*fTP*pDIF is, but I forget if the intermediate step is.)

[Nightfyre]

Ah, nevermind then. I'm not sure if it floors or not but I'd guess it does. That would be really annoying to check given the two randomizations in the damage equation.

[Masamune]

Honestly, i don't buy the "SE likes to floor multiple times".
With the years, you guys litterally confounded the lil geniuses that put in place the whole dmg and stats system, with real SE devs.
Those guys were probably like us, and decided need effectively multiple flooring at some points only in their calculations for them to match as close as possible observed dmg values.
Another scenario could be perfectly something like one single (and hopefully) simple direct equation giving exact dmg values (and not approximated) ?
Just posting for the 2 posters above me.

Kudos on your tests and well-formed conclusions, Kirschy ^^

[Rena]

Magic damage is floored at every single step (and this is very easy to prove given the lack of randomness in magic damage, resists and day/weather aside, which are very predictable), except when combining WSC before level correction for magic WS/BLU nukes (and possibly MAB/MDB (though it is floored after this is calculated again)), which is exactly how WSC on physical WS supposedly is calculated. It's more reasonable to assume the same goes for physical damage than not, unless you actually find some evidence to the contrary.

[Norellicus]

Honestly, i don't buy the "SE likes to floor multiple times".
With the years, you guys litterally confounded the lil geniuses that put in place the whole dmg and stats system, with real SE devs.
Those guys were probably like us, and decided need effectively multiple flooring at some points only in their calculations for them to match as close as possible observed dmg values.
Another scenario could be perfectly something like one single (and hopefully) simple direct equation giving exact dmg values (and not approximated) ?
Just posting for the 2 posters above me.

Kudos on your tests and well-formed conclusions, Kirschy ^^

Or, mathematics could be doing exactly what it's supposed to and producing observed results based on a number of variables, and over time we've refined those variables to be more precise.

Flooring is a perfectly normal method that, given the seemed simplicity of SE's coding, makes it easy to pass values from one function to another without having to store and recall a bunch of floating point integers.

[Kirschy]

Or, mathematics could be doing exactly what it's supposed to and producing observed results based on a number of variables, and over time we've refined those variables to be more precise.

Flooring is a perfectly normal method that, given the seemed simplicity of SE's coding, makes it easy to pass values from one function to another without having to store and recall a bunch of floating point integers.

Gonna have to agree here. As someone who has taken her share of computer architecture/programming courses, it's easy to accept that values are truncated when multiple operations are being done on them. My experience messing with damage calculations (back when pdif capped), and haste calculations has shown that SE doesn't like floating point operations. (Which would make sense, since tons of those operations are done per second every time actions are taken.)

I've also come to accept that sometimes we can't get exact functions, and just need an good model /w a high prediction rate. 4.76% for proc rate on helm isn't exactly a great number. It could be 5% /w margin of error, it could be some obscure x/1024 number, or even interacting /w shadow mantle in some way. Regardless, it's sufficient enough to me to say "When I'm about to get take some serious physical damage, I feel using Noct Helm will give me the best overall benefit."

[Masamune]

Gonna have to agree here. As someone who has taken her share of computer architecture/programming courses, it's easy to accept that values are truncated when multiple operations are being done on them. My experience messing with damage calculations (back when pdif capped), and haste calculations has shown that SE doesn't like floating point operations. (Which would make sense, since tons of those operations are done per second every time actions are taken.)

Ok, thank you for explanation. But then another question comes to my mind :
If we assume SE used floorings (or truncations), would it make sense to see a few only variables actually not floored/truncated ? (asking that while thinking about fSTR or Ratio&pDIF)
I didnot test with ALL operations truncated to 1 decimal (like TP function) or entirely floored at each step (like Rena example with Magic calculations). Thank you guys for the idea, i hope it work :s.

[Kirschy]

Ok, thank you for explanation. But then another question comes to my mind :
If we assume SE used floorings (or truncations), would it make sense to see a few only variables actually not floored/truncated ? (asking that while thinking about fSTR or Ratio&pDIF)
I didnot test with ALL operations truncated to 1 decimal (like TP function) or entirely floored at each step (like Rena example with Magic calculations). Thank you guys for the idea, i hope it work :s.

I think tp is stored as a non-decimal number since floating point operations use a ton of cycles to perform basic math. For example 150.3 tp is stored as 1503, but displayed as 150.

I know fSTR can only be a whole number, as it was easy to see when normal hits capped to 3.0 pdif.

I have no clue on pdif though

[Baelorn]

Flooring is a perfectly normal method that, given the seemed simplicity of SE's coding, makes it easy to pass values from one function to another without having to store and recall a bunch of floating point integers.

You lost me here.