combat stats, proc rates, and a lot of questions

[bobjonesp99]

trying to better understand the combat statistics in p99. i have a general understanding of weapon damage formulas and proc rates, but there is some things that aren't very clearly defined that i am trying to get a better understanding of. specifically i am talking about a warrior in this example, but presumably the same applies to other classes (with some minor exceptions).

proc rate:

i have read and understand that proc rates are defined on a per minute basis. as i understand, they are normalized to your 'per swing' rate, ignoring things like haste, weapon delay, and double/triple attack.

i have read that proc rate is 0.5 to 2.0 ppm from 85 dex to 255. this would imply 85 to 170 dex equated to 0.5 to 1.5 ppm and 170 to 255 dex equating to 1.5 to 2.0 ppm.

i have read that proc rate is dex/170+0.5 ppm. this would mean 0 dex equate to 0.5 ppm, 85 dex to 1.0 ppm, 170 dex to 1.5 ppm, and 255 dex to 2.0ppm.

can anyone confirm either of the above statements to be accurate with some level of certainty?

i have read 'the secondary proc rate is equal to half the primary proc rate'. is it actually half? or is it based on your dual wield chance? in other words, if you just reached dual wield level, are you actually getting half the proc rate for the secondary? i have also heard some people suggest that their secondary seems to proc MORE than 50% of the primary... which begs the question of is it based on dual wield chance.


dual wield:

i have read the formula for dual wield is [dual wield skill + level] / [max skill]. i have read the max skill could be anywhere from 400 to 600 depending on class and that is is unclear what is used in P99.

assuming max skill 600 means a level 60 warrior with 240 dual wield skill would dual wield 50% of the time.

assuming max skill 400 means a level 60 warrior with 240 dual wield skill would dual wield 75% of the time.

can anyone confirm either of the above statements to be accurate with some level of certainty?

another question is, does the dual wield attack act as a 'proc' on the primary attack? i.e. is it that the dual wield is attacking only when a successful dual wield is rolled on the primary (and then separately adjusting the number of attacks to account for the damage/delay of the secondary)? or does the dual wield act as a delay modifier (i.e. like haste in reverse) reducing the secondary delay by dividing by the dual wield modifier?


double attack / triple attack:

parsing this is all complicated due to double/triple attack which adds another couple variables.

i have read the formula for double attack is [double attack skill] / [max skill x 1.05]. i have read the max skill could be anywhere from 400 to 600 depending on class and that is is unclear what is used in P99.

assuming its 600 means a level 60 warrior with 245 double attack skill would double attack 38.888% of the time.

assuming its 400 means a level 60 warrior with 245 double attack skill would double attack 58.333% of the time.

although not a skill that you can train, most warriors confidently state that triple attack is a real thing. i have read the formula for triple attack is [double attack skill x 0.5] / [max skill x 1.05]. this implies half the rate for triple attack as double attack.

assuming its 600 means a level 60 warrior with 245 double attack skill would triple attack 19.444% of the time.

assuming its 400 means a level 60 warrior with 245 double attack skill would double attack 29.167% of the time.

one thing that is unclear to me, is if triple attacks only apply to primary weapon attacks, or also apply to secondary weapon attacks. from my limited parsing, i did not see any triple attacks on the secondary, but i am still unclear if that is accurate or a sample bias. i am also not entirely convinced triple attacks are an actual thing for dual wield weapons and may just look that way due to hasted delays < 10 giving multiple primary rounds (with successful double attacks) in a single second. most warriors seem convinced it is a thing, so i have included it for completeness and discussion.


some parse data:

i parsed a level 60 warrior with max skills, 58 percent haste, and was equipping a primary infestation (9/18) + secondary frostbringer (12/22).

across the 83 second fight i saw 1085 primary attempts compared to 585 secondary attempts; keep in mind total attempts includes dual wield chance, double attack chance, and triple attack chance.

based on those weapons delays and haste i am calculating an estimated ~54.5 primary swings per minute and ~42.9 secondary swings per minute; keep in mind this does not account for dual wield chance, double attack chance, and triple attack chance.

from the actual attempts per minute for primary and secondary, i can back calculate the attempts per swing.

what i saw was that attempts per swing for the primary was ~1.54 and for the secondary was approximately ~1.13; keep in mind this is independent of haste and the delay of the weapons... it is calculated as dual wield chance (100% for primary) + dual wield chance x double attack chance + dual wield chance x triple attack chance.

this results in a non-unique formula since we don't know exactly what the dual wield chance, double attack chance, and triple attack chance are due to the unknown 'max skill' used in those calculations. we can assume values and try and come up with estimates. i was able to get a very close approximation to the damage per second, and number of attempts for primary and secondary by assuming max skill = 400 for dual wield and double attack, and assuming that triple attack is not real (only observed to be true due to hasted weapon delays < 10). this implies a 58% double attack chance, 75% dual wield chance. that results in primary attempts per swing of ~1.58 and secondary attempts per swing of 1.19. have a 75% dual wield chance would also give some weight to the prior statement that some warriors suggest proc rate for secondary weapons is >50% of the primary weapon.

all of that said, i am sure someone else has gone down this rabbit hole and can provide some additional insight to many of the questions above.


white damage hate:

one last question regarding white damage hate.

i have read the formula for white damage hate is [weapon damage + bonus weapon damage] / [weapon delay]. i understand this is independent of player attack rating, target ac, randomness in actual weapon damage, and any damage mitigation that gets applied. i also understand this is impacted by haste. my question is, do double/triple attacks get accounted for as well? or is it just the initial 'swing'? as you can see it all ties into the above calculation.

i look forward to the potential discussions!

[Allishia]

Everything is random /nod...

Proc 5x in a row one minute, then sometimes go an entire mob w/o a single proc.

[DeathsSilkyMist]

I am not 100% sure on this (if someone could confirm that would be great), but for a time there was also a proc adjustment on certain weapons to prevent them from being too overpowered. This means they would actually proc less than average. I think this was still not common, my brain says Earthshaker has/had a proc restriction. So if those are still in place that would be another consideration to look at, but in general it should be the same. I doubt Frostbringer or Infestation has that.

I would just keep doing parses since you have the capability to do so. For proc research I would just focus on counting the procs specifically, instead of trying to count the white text, since it does have a lot of extra swings due to double attack, tripple attack, etc. As long as you have two different procs in your hands, you can easily sum up an average amount of procs per minute for both hands with enough data.

[Snaggles]

I considered parsing the same delay 1hb and 1hs (like splintered club and rusty short sword) with my ranger for this same curiosity. Offhand swings in general are quite a bit less unless you have a slow AF mainhand. It's part of the reason with the upgrade to 2h on blue (scaling dmg bonus via delay) you better have amazing 1h's to consider them for dps. At least comparing apples to apples.

Proc rate is PPM as you mentioned. MH will proc twice as much on average but as Allishia said random is as random does. No weapons proc more than others. Slow weapons will have a higher proc per swing average though (earthshaker etc) but if you time it per minute they wont proc more than a quick weapon. On average.

White aggro as I understand is possible damage done which includes your damage bonus. Basically dps minus the actual dps. The prob is most the best dps 1h's have low aggro procs or none at all so it's something to consider. The Feverblade is one of the few that strikes a good balance.

Edit: All weapon proc rates are normalized on blue. Most aggro capped (relatively) for spells you cannot cast. Some are flagged non-pet weapons like the Earthshaker and weighted axe.

[SlankyLanky]

Everything is random /nod...

Proc 5x in a row one minute, then sometimes go an entire mob w/o a single proc.

i think this is important to hear for people wanting to roll a new war or people who are getting to the level where its actually important for the tank to be the one whos getting hit. dex makes a big difference, but so does rng when it comes to getting a proc on a mob.

[Allishia]

i think this is important to hear for people wanting to roll a new war or people who are getting to the level where its actually important for the tank to be the one whos getting hit. dex makes a big difference, but so does rng when it comes to getting a proc on a mob.

https://wiki.project1999.com/Forlorn_Totem_of_Rolfron_Zek

Get those, can usually land 2 clicks before vindi gets in camp. Most tov mobs can land 1-2 before dragon turns on me from the bumpers...cheap recharge too.

I'm max 255 dex and procs are still random. Clickys are your best bet lol. When mob turns to eat a monk that doesn't know how to fd, you can pop another click :p

[jdmchris]

You’re trying way too hard bro

[Videri]

i have read that proc rate is 0.5 to 2.0 ppm from 85 dex to 255. this would imply 85 to 170 dex equated to 0.5 to 1.5 ppm and 170 to 255 dex equating to 1.5 to 2.0 ppm.

No. It's 0.5 to 2.0 ppm from ZERO dex to 255 dex.

i have read that proc rate is dex/170+0.5 ppm. this would mean 0 dex equate to 0.5 ppm, 85 dex to 1.0 ppm, 170 dex to 1.5 ppm, and 255 dex to 2.0ppm.

Yes. You got it right in this part.

[Videri]

i have read 'the secondary proc rate is equal to half the primary proc rate'. is it actually half? or is it based on your dual wield chance? in other words, if you just reached dual wield level, are you actually getting half the proc rate for the secondary? i have also heard some people suggest that their secondary seems to proc MORE than 50% of the primary... which begs the question of is it based on dual wield chance.

Don't mix up dual wield chance with it. It's like so:

Let's say you're wielding a 10/30 weapon in each hand. That means you swing every 3 seconds. That means you swing how many times per minute? 20.

Let's say you have 85 dex, so 1 proc per minute mainhand. Divide 1 proc per minute by 20 swings per minute and you get 0.05 procs per swing; in other words, 5% chance to proc per swing.

Each time you swing your mainhand weapon, you have 5% chance to proc.
Each time you swing your offhand weapon, you have a 2.5% chance to proc.

So...don't mix up dual wield chance with this. It's just a percentage that is applied every time you swing. And remember, you might still proc three times in a row!! It's just probability!


HOWEVER, if you get slowed, let's say by 50%, now you're gonna swing half as much...but you'll still have the same procs-per-minute...so now you'll proc twice as much per swing. 10% chance to proc on each mainhand swing, 5% chance to proc on each offhand swing. See what I mean?

[bobjonesp99]

Don't mix up dual wield chance with it. It's like so:

Let's say you're wielding a 10/30 weapon in each hand. That means you swing every 3 seconds. That means you swing how many times per minute? 20.

Let's say you have 85 dex, so 1 proc per minute mainhand. Divide 1 proc per minute by 20 swings per minute and you get 0.05 procs per swing; in other words, 5% chance to proc per swing.

Each time you swing your mainhand weapon, you have 5% chance to proc.
Each time you swing your offhand weapon, you have a 2.5% chance to proc.

So...don't mix up dual wield chance with this. It's just a percentage that is applied every time you swing. And remember, you might still proc three times in a row!! It's just probability!


HOWEVER, if you get slowed, let's say by 50%, now you're gonna swing half as much...but you'll still have the same procs-per-minute...so now you'll proc twice as much per swing. 10% chance to proc on each mainhand swing, 5% chance to proc on each offhand swing. See what I mean?

i understand the distinction between procs per minute and procs per swing and how the game normalizes delay such that your proc per minute is constant (either haste or slow).

what i am saying is that, your secondary weapon swings less frequently than your mainhand and as a result the procs per minute of the secondary weapon is less than that of your mainhand. log data for hours of fighting green cons on a 60 warrior with two different procing weapons shows the frequency of the secondary weapons proc was ~54% of the primary weapons proc. based on the formula for dual wield chance on a 60 warrior, it calculates as ~50% chance.

[Videri]

i understand the distinction between procs per minute and procs per swing and how the game normalizes delay such that your proc per minute is constant (either haste or slow).

what i am saying is that, your secondary weapon swings less frequently than your mainhand and as a result the procs per minute of the secondary weapon is less than that of your mainhand. log data for hours of fighting green cons on a 60 warrior with two different procing weapons shows the frequency of the secondary weapons proc was ~54% of the primary weapons proc. based on the formula for dual wield chance on a 60 warrior, it calculates as ~50% chance.

And now we are getting to the point where I do not have answers. I think it's close enough.

[Videri]

Wait a second

i parsed a level 60 warrior with max skills, 58 percent haste, and was equipping a primary infestation (9/18) + secondary frostbringer (12/22).

across the 83 second fight i saw 1085 primary attempts compared to 585 secondary attempts; keep in mind total attempts includes dual wield chance, double attack chance, and triple attack chance.


83 seconds? That's waaaaaaay too small a sample size to be conclusive.

If you're going to try to get this far into the formulae etc, you're going to need far more than 83 seconds of data. Perhaps one of the resident statisticians can estimate an amount of parsing time necessary to get actually useful results, but I'd bet you need to parse for hours.

Secondly, all this stuff is governed by formulae hidden in the server and client. That doesn't mean we're going to get the same results every time. Even if you run the same experiment over and over, you'll get different numbers every time.

[bobjonesp99]

Wait a second



83 seconds? That's waaaaaaay too small a sample size to be conclusive.

If you're going to try to get this far into the formulae etc, you're going to need far more than 83 seconds of data. Perhaps one of the resident statisticians can estimate an amount of parsing time necessary to get actually useful results, but I'd bet you need to parse for hours.

Secondly, all this stuff is governed by formulae hidden in the server and client. That doesn't mean we're going to get the same results every time. Even if you run the same experiment over and over, you'll get different numbers every time.

this should have read "across the 83 second fight i saw 116 primary attempts compared to 67 secondary attempts" (0.58). in this example is was referring to a single fight. the example fight was consistent with the much larger data that i quoted as 1085 primary attempts and 585 secondary attempts (0.54).

[Videri]

Even 1000 swings isn't enough of a sample size. :(

There's plenty of room for variation that could explain the difference between 0.50 and 0.54.

[Snaggles]

It's not accurate enough for a thesis paper but close enough for a general perspective that the offhand does about half the work, maybe even less since no damage bonus. Maybe without triple attack and on a longer parse the MH it's a bit closer to that 60/70% range.

As for the rest (procs linked to swing), it's a matter of if you are mainly curious about the science or hoping it will bring a functional difference to how someone would gear out a warrior. I don't think anything will change. The best dps and proc will always go in the MH and the 2nd best in the offhand at least if proc aggro is capped. The only time I saw people deliberating over this was with the WESS and Infestation; whether to lose a poison counter worth of aggro for less white damage aggro.

[ScottBerta]

So being slowed doesn’t actually increase chance of proc but it doesn’t lesson the chance as well? I was under the impression that slowing yourself before you use OT hammer would get it to proc faster.. which made little sense to me.

[kaev]

Proc chance is time-based, roughly 1/minute mainhand with 100dex. So when you are slowed the chance to proc per swing becomes higher to keep the chance per minute the same. So if you're a melee with high 2hb skill trying to hammer off some wimpy cloth caster take off your haste item and if poss get a slow so you don't beat the poor clothie all to hell waiting on your proc.

[mattydef]

I’ve been playing this game for a long time and one thing I can say with certainty is that the Paladin / SK epic weapons proc far more than 2 ppm.

[jolanar]

So being slowed doesn’t actually increase chance of proc but it doesn’t lesson the chance as well? I was under the impression that slowing yourself before you use OT hammer would get it to proc faster.. which made little sense to me.

Slowing to make OT hammer proc is pretty pointless. It's still 1 ppm. You swing less but proc more. You might get more procs on the first swing, but over time the proc rate will still even out.

Slowing to increase proc rate of weapons when you can only get a few swings off between casts (basically only applies to shamans) does make some sense in specific situations.

[booleansphinx]

dual wield:
i have read the formula for dual wield is [dual wield skill + level] / [max skill]. i have read the max skill could be anywhere from 400 to 600 depending on class and that is is unclear what is used in P99.


EQEMU is likely not 100% true to P99s formulas, but I'd bet it's pretty close.
In the EQEMU,
Double attack is (Skill + Level) / 500
Dual Wield is (Skill + Level) / 375
Parry and Dodge is (Skill + 100) / 4500
Riposte is (Skill + 100) / 5000

I never looked into procing, it gets pretty hairy in the code.
https://github.com/EQEmu/Server/blob/master/zone/attack.cpp

The wiki has a lot of mention about the denominator ("max skill") changing based on class, but I think that's unlikely. The different skill caps given to classes only really make sense if the "max skill" is the same.

another question is, does the dual wield attack act as a 'proc' on the primary attack? i.e. is it that the dual wield is attacking only when a successful dual wield is rolled on the primary (and then separately adjusting the number of attacks to account for the damage/delay of the secondary)? or does the dual wield act as a delay modifier (i.e. like haste in reverse) reducing the secondary delay by dividing by the dual wield modifier?

Your offhand attacks the same was as your main hand, except before the attack goes through it makes a skill check. If it fails, the offhand goes on cooldown based on weapon delay and no attack occurs. Similar to what happens if you are out of range with your auto-attack on.

[Jimjam]

Your offhand attacks the same was as your main hand, except before the attack goes through it makes a skill check. If it fails, the offhand goes on cooldown based on weapon delay and no attack occurs. Similar to what happens if you are out of range with your auto-attack on.

Are you talking about p1999 here? If so upon what are you basing this statement?

Ever played a high level ranger with Wyrmslayer and Lupine Dagger?

[booleansphinx]

Are you talking about p1999 here? If so upon what are you basing this statement?

Ever played a high level ranger with Wyrmslayer and Lupine Dagger?

I can't figure out how to edit my post, but I can see the wording is confusing.

Your offhand attacks the same way as your main hand, on an independent timer. Except, before the offhand attack goes through, it makes a skill check. If it fails, the offhand goes on cooldown based on its weapon delay (as if it had swung), but the offhand attack does not occur. This is similar to what happens to your main hand if you are out of range with your auto-attack on.

[Snaggles]

The last Scout roll I did for my 60 ranger I used a BFG (20/50) and a Swiftwind (13/21). This was the amount of hits landed (not sure on misses):

67 seconds
BFG main hand = 36 hits (73.2% dps)
Swiftwind offhand = 51 hits (26.8% dps)

I disc'ed so the skill cap for archery should have been the same as 1h. In short, the offhand swings independently of the main.

[Snaggles]

I’m not refuting what Boole said, it actually makes sense. Just added some extra data for the MH/OH swing dynamics and overall dps difference.

[bobjonesp99]

EQEMU is likely not 100% true to P99s formulas, but I'd bet it's pretty close.
In the EQEMU,
Double attack is (Skill + Level) / 500
Dual Wield is (Skill + Level) / 375
Parry and Dodge is (Skill + 100) / 4500
Riposte is (Skill + 100) / 5000

I never looked into procing, it gets pretty hairy in the code.
https://github.com/EQEmu/Server/blob/master/zone/attack.cpp

The wiki has a lot of mention about the denominator ("max skill") changing based on class, but I think that's unlikely. The different skill caps given to classes only really make sense if the "max skill" is the same.



Your offhand attacks the same was as your main hand, except before the attack goes through it makes a skill check. If it fails, the offhand goes on cooldown based on weapon delay and no attack occurs. Similar to what happens if you are out of range with your auto-attack on.

appreciate comments like this; i just plugged these updated forumlae into my calculator and its coming very close to the attempt per primary and attempt per secondary for my parsed data sets.

still unclear on if triple attacks are a thing for 1h, dual wield and/or 2h, but could probably be tested with 60 war and high delay weapons without haste. i think one of the current issues is inputs in logs are per second and if a hasted delay faster than 10 you can have multiple inputs on the same second so it looks like >2 attempts for primary for example.

[Grumph]

I just did a quick little parse with a 25 delay main hand wepon at 100% haste. (so should be 1.25 secs per round...)

out of 533 main hand rounds...

41% were single attack
54% were double attack
5% appear to have been triple attack.


I also saw 64% double attacks with off hand.

Hilariously: (54+5) + 64 ... averages out to 61% chance to double attack. which falls in the line with the formula posted above.

Additionally. Said formulas predict an 80% chance to swing off hand. And I saw a 77% swing rate. Seems reasonably within confidence interval based on sample size.