Dual wield logic

[DeathsSilkyMist]

Originally Posted by Snaggles [You must be logged in to view images. Log in or Register.]

A monk doesn’t have the luxury of picking two different emote styles of the same delay (ie slash and crush).

Monks do have a few options for same delay weapons with different emotes if you really wanted to. The trick is to unequip Epic Fist. A level 60 Iksar Monk has a 27 Delay Fist without Epic. So you can do bare fist + https://wiki.project1999.com/Club for example. But as I showed in my previous post, you don't need two weapons of the same delay to figure out dual wield chance. You just need the dual wield formula and the double attack formula.

[Defo]

Even at max skill (252) on a monk, dual wield only fires roughly 65-70% of the time, according to years of my parses.

[DeathsSilkyMist]

Originally Posted by Defo [You must be logged in to view images. Log in or Register.]

Even at max skill (252) on a monk, dual wield only fires roughly 65-70% of the time, according to years of my parses.

Generally speaking if you want to do a proper parse, it should be in a controlled environment. From experience I've learned that results can be skewed if you are parsing while playing normally.

As a simple example, you will probably miss a few swings on a fleeing mob. If you add that data to the parses you are reviewing, you have a dataset with less swings than there should be compared to the time of the fight. This would give you a lower dual wield percentage potentially.

Obviously you can't really have a controlled environment on parsing raid mobs, but testing something like dual wield percentage can be done on any mob.

[bcbrown]

Originally Posted by Goregasmic [You must be logged in to view images. Log in or Register.]

Fair points. I mentioned <28 because that's the scenario I ran the numbers on back then, with no dmg bonus you're basically dealing with 2 offhands. I've always had trouble inserting the DW variable into DPS calculations because that page has no confirmed numbers. IIRC I've seen parses more around 65% but I'm really not sure. DSM probably has it in his calculator?

Also, to add another layer of complexity, it seems(?) double attack doesn't fire from offhand until 150 or so (lvl30) so it is probably an incentive to use the better ratio in mainhand until significant damage bonus kicks in to get a bit more bang for your bucks. Low DW chance and even lower DA on OH is why people will say to not bother with 1handers when 2h damage caps are lifted but that is somewhat beyond the point.

Yep, as you said, when in doubt math it out. It may change as you level and the damage bonus becomes more important and DW skill increases.

The DW wiki includes a formula as well as a table for lvl 50/60, and I've found it to be accurate. The last time I parsed DW success rate was at level 51, where I found it to be 65%, matching that table. The example I gave was for monks with a higher skill cap; they should be at ~75% at lvl50 but I don't have a monk to test that.

One method for checking DW success rates was already posted, using fight duration divided by adjusted delay and then scaling by DW and DA skill checks to compute an expected number of swings and comparing to the actually recorded number of swings. That works well, but I prefer a different methodology that with a few assumptions allows you to directly measure the DA and DW success rates. Both, of course, require different weapon types in mainhand and offhand so you can tell which is which.

The central assumption is that both weapons have an adjusted delay significantly longer than one second. In that case, if you have two swings in the same second you can assume that's a double attack. So first you collect all swings or attempted swings, then count how many timestamps have a single swing and how many have two. The ratio is the double attack success rate. Then, you add the counts to find the total number of successful dual wield skill checks. Dividing the fight length by the adjusted delay gives you the expected number of swings if dual wield is always successful, and dividing the actual number of successful DW skill checks by that theoretical maximum gives you the dual wield success rate.

Here's an example from when I was level 51, Infestation offhand, 22% worn haste. There were 405 timestamps with a single piercing swing and 457 with two piercing swings, giving a 53% DA success rate. The fight duration was 1944 seconds, the adjusted delay was 14.75, and so with a 100% DW success rate you'd expect 1317 rounds of attacks. Compared to the sum of 405 and 457, that's 862/1317, or an observed 65.4% dual wield success rate.

This approach breaks down when the offhand is fast enough to potentially swing twice in the same second, so for example I can't use it with my shiny new Barbed Scale Whip.

I hadn't known that DA doesn't fire offhand till lvl30, that's interesting. My baby ranger is 30 already so I guess testing that will have to wait for the next ranger. Another interesting point of trivia for dual wield is that you can skill up while wielding a shield. Also, Jimjam once mentioned that weapon procs can fire even when you fail a dual wield skill check, which is part of why I think it can be a good idea to put a weapon with a good proc but a poor ratio in the offhand.

[Goregasmic]

Originally Posted by bcbrown [You must be logged in to view images. Log in or Register.]

Yep, as you said, when in doubt math it out. It may change as you level and the damage bonus becomes more important and DW skill increases.

The DW wiki includes a formula as well as a table for lvl 50/60, and I've found it to be accurate. The last time I parsed DW success rate was at level 51, where I found it to be 65%, matching that table. The example I gave was for monks with a higher skill cap; they should be at ~75% at lvl50 but I don't have a monk to test that.

One method for checking DW success rates was already posted, using fight duration divided by adjusted delay and then scaling by DW and DA skill checks to compute an expected number of swings and comparing to the actually recorded number of swings. That works well, but I prefer a different methodology that with a few assumptions allows you to directly measure the DA and DW success rates. Both, of course, require different weapon types in mainhand and offhand so you can tell which is which.

The central assumption is that both weapons have an adjusted delay significantly longer than one second. In that case, if you have two swings in the same second you can assume that's a double attack. So first you collect all swings or attempted swings, then count how many timestamps have a single swing and how many have two. The ratio is the double attack success rate. Then, you add the counts to find the total number of successful dual wield skill checks. Dividing the fight length by the adjusted delay gives you the expected number of swings if dual wield is always successful, and dividing the actual number of successful DW skill checks by that theoretical maximum gives you the dual wield success rate.

Here's an example from when I was level 51, Infestation offhand, 22% worn haste. There were 405 timestamps with a single piercing swing and 457 with two piercing swings, giving a 53% DA success rate. The fight duration was 1944 seconds, the adjusted delay was 14.75, and so with a 100% DW success rate you'd expect 1317 rounds of attacks. Compared to the sum of 405 and 457, that's 862/1317, or an observed 65.4% dual wield success rate.

This approach breaks down when the offhand is fast enough to potentially swing twice in the same second, so for example I can't use it with my shiny new Barbed Scale Whip.

I hadn't known that DA doesn't fire offhand till lvl30, that's interesting. My baby ranger is 30 already so I guess testing that will have to wait for the next ranger. Another interesting point of trivia for dual wield is that you can skill up while wielding a shield. Also, Jimjam once mentioned that weapon procs can fire even when you fail a dual wield skill check, which is part of why I think it can be a good idea to put a weapon with a good proc but a poor ratio in the offhand.

Yeah you can absolutely calculate it, just wasn't all that sure it was worth the time before mid 50s and I'm still using 2h almost exclusively.

For DA not triggering on offhand until 150, I wanted to test that on my 16 rogue that just got DA. I read it a couple times but I can't find anywhere that confirms it. Lately I was leveling 1hs and a had a croaking dirk in OH and I can say OH doesn't fire all that much so maybe it can DA but since DW hardly ever triggers and DA on main hand either so maybe it is possible but a statistically insignificant chance.

Didn't know you can skill DW with a shield, that's... counter intuitive. I know you can trigger a proc on a miss but never noticed you could proc on a non swing. I guess it is a byproduct of normalized proc rates, your odds are probably based on swing speed and it pops off regardless of outcome. Much simpler to land on the target value from a programming perspective that way too when you think about it.

For good OH procs on weapons with bad ratios, on one hand you'd have to find a proc worthy of gimping your damage for it but on the other hand for a class like rogue, apparently like 80% of your damage comes from MH (backstabs) so swapping for something like a locustlure OH might be worth the occasional proc if you're losing literally like, 3% of your total damage output vs a 10/18 weap in OH.

[DeathsSilkyMist]

Originally Posted by Goregasmic [You must be logged in to view images. Log in or Register.]

For DA not triggering on offhand until 150, I wanted to test that on my 16 rogue that just got DA. I read it a couple times but I can't find anywhere that confirms it. Lately I was leveling 1hs and a had a croaking dirk in OH and I can say OH doesn't fire all that much so maybe it can DA but since DW hardly ever triggers and DA on main hand either so maybe it is possible but a statistically insignificant chance.

Here is a video of my level 25 Warrior fighting corudoth for 35 minutes.

This video is a bit unusual, as I broke my normal rule of being consistent with buffs. I did 3 casts of Cog Boost (40% Haste), as I was mostly checking max hit.

There were 987 main hand swing attempts and 336 offhand swing attempts over 2139 seconds. My haste would be 62% for 1530 seconds (Cog Boost was 8 minutes 30 seconds duration at my level), and 22% for the rest. The average haste assuming a 50/50 split is 62 + 22 = 84 / 2 = 42% haste. Since the haste wasn't a 50/50 split, 1530 / 2139 = 0.715 * 62 = 44% Haste on average.

1. 336 Offhand Crush Attempts [30 Delay / 1.44 Haste = One attack Every 2.08 Seconds]
2. 987 Primary Slash Attempts [40 Delay / 1.44 Haste = One attack Every 2.77 Seconds]
3. Fight was 2139 seconds long

2139 Seconds / 2.08 Seconds = 1028 Offhand Crush Attempts * ((130 Dual Wield Skill Level + 25 Player Level) / 475) = ~335 Predicted Secondary Crush Attempts on average.

2139 Seconds / 2.77 Seconds = 772 Primary Slash Attempts * (1 + ((130 Double Attack Skill + 25 Player Level) / 500)) = ~1011 Predicted Primary Slash Attempts on average.

Both of the predicted numbers above track with my log results. Also looking at my logs, I didn't see any crush attempts that occured within the same second.

So at least at level 25, it doesn't look like double attack triggers on offhand. Also it looks like 475 is the correct "MaxSkill" value for warriors, as the wiki suggests.

[Goregasmic]

Originally Posted by DeathsSilkyMist [You must be logged in to view images. Log in or Register.]

Here is a video of my level 25 Warrior fighting corudoth for 35 minutes.

This video is a bit unusual, as I broke my normal rule of being consistent with buffs. I did 3 casts of Cog Boost (40% Haste), as I was mostly checking max hit.

There were 987 main hand swing attempts and 336 offhand swing attempts over 2139 seconds. My haste would be 62% for 1530 seconds (Cog Boost was 8 minutes 30 seconds duration at my level), and 22% for the rest. The average haste assuming a 50/50 split is 62 + 22 = 84 / 2 = 42% haste. Since the haste wasn't a 50/50 split, 1530 / 2139 = 0.715 * 62 = 44% Haste on average.

1. 336 Offhand Crush Attempts [30 Delay / 1.44 Haste = One attack Every 2.08 Seconds]
2. 987 Primary Slash Attempts [40 Delay / 1.44 Haste = One attack Every 2.77 Seconds]
3. Fight was 2139 seconds long

2139 Seconds / 2.08 Seconds = 1028 Offhand Crush Attempts * ((130 Dual Wield Skill Level + 25 Player Level) / 475) = ~335 Predicted Secondary Crush Attempts on average.

2139 Seconds / 2.77 Seconds = 772 Primary Slash Attempts * (1 + ((130 Double Attack Skill + 25 Player Level) / 500)) = ~1011 Predicted Primary Slash Attempts on average.

Both of the predicted numbers above track with my log results. Also looking at my logs, I didn't see any crush attempts that occured within the same second.

So at least at level 25, it doesn't look like double attack triggers on offhand. Also it looks like 475 is the correct "MaxSkill" value for warriors, as the wiki suggests.

What's the DA skill max for a 25 warrior? 130? I guess this pretty much confirms what I read. We'd need to find the exact point where it starts triggering but it does confirm that DW is pretty shit if you can 2h mostly uncapped.

Do you "lose" skill points if you delevel? Unless it is already documented somewhere I have a monk that is pretty much exactly 30. I could delevel him, see if it triggers at 28 and then check again at 29 then get a whole bunch of rezzes. Or I could just get my rogue/war to 28 and see but that isn't a priority right now.

[DeathsSilkyMist]

Originally Posted by Goregasmic [You must be logged in to view images. Log in or Register.]

What's the DA skill max for a 25 warrior? 130? I guess this pretty much confirms what I read. We'd need to find the exact point where it starts triggering but it does confirm that DW is pretty shit if you can 2h mostly uncapped.

Do you "lose" skill points if you delevel? Unless it is already documented somewhere I have a monk that is pretty much exactly 30. I could delevel him, see if it triggers at 28 and then check again at 29 then get a whole bunch of rezzes. Or I could just get my rogue/war to 28 and see but that isn't a priority right now.

Yeah double attack max skill is 130 for warriors at 25.

As far as I know, when you delevel, your skills become capped at the max for that level. So if a level 26 Warrior had 135 Double Attack skill, deleving to 25 would mean his Double Attack becomes capped at 130 until he leveled again.

[Snaggles]

When you delevel does your displayed skill via your paper doll skill menu decrease? I’ve never checked it.

The only time I legitimately deleveled (outside making EE’s at 60) was to make a Nagafen draft with my 54 wiz and then went back to click the 10 or so bodies. It still wasn’t worth it in the end.

IMHO, much of the DW vs 2h debates come down to a couple weapon sets which you can easily test under various circumstances. The toughest and most contested one is how they parse against raid targets due to short and inconsistent fights. I guess you could string together enough Vindi’s to make sense of it or just go with “best single x vs best single y”

[Goregasmic]

Originally Posted by Snaggles [You must be logged in to view images. Log in or Register.]

When you delevel does your displayed skill via your paper doll skill menu decrease? I’ve never checked it.

The only time I legitimately deleveled (outside making EE’s at 60) was to make a Nagafen draft with my 54 wiz and then went back to click the 10 or so bodies. It still wasn’t worth it in the end.

IMHO, much of the DW vs 2h debates come down to a couple weapon sets which you can easily test under various circumstances. The toughest and most contested one is how they parse against raid targets due to short and inconsistent fights. I guess you could string together enough Vindi’s to make sense of it or just go with “best single x vs best single y”

On high AC targets a lot of your hits will become minimum hits so I guess it is somewhat down to which setup can optimize minimum hit damage?

Kinda like peacebringer vs IFS. IFS has a slightly better ratio and 4 extra damage bonus but peacebringer is 33% faster. In general peacebringer is probably the better weapon until tstaff.