Hagun and Onimaru

[ringthree]

An easy calculations leads to this: with haste, 1 march, hasso , dusk gloves, swift belt, byakkos, usu feet (=50% haste), Walhara turban beats usukane head as long as your accuracy 79+%, and becomes all the better that your haste is higher. With 2 marches, its better as long as your acc is 63+% ( and it's even worse with askar: 68% and 54%...). Usukane/askar heads don't come close to ace's helm or turban, since you need to be very close to capped acc for turban to beat ace's (~93% to be exact).

Hasso stacks separately from gear haste I believe.

[WraithLakshmi]

In your situation of 70% haste, walahra turban's 1% haste would be a difference of 3.3% over Askar zucchetto.

.31/.3 = 1. 03333

The str on Askar would increase your fSTR by 1 point, which for my melee setup is an increase of 1.2%.

(75+10)/(75+9) = 1.012

The addition to crit chance from 4 dex is hard to nail down, but for me it would bring my dex from 96 to 100, which based on some parses in a previous post, would be approximately a .75% increase to crit chance.

This means the accuracy from the dex needs to make up 1.35%. This will occur at all values for melee accuracy.

(x+.015)/x = 1.0135
x = 1.111

So, only if you had over 111% melee accuracy would walahra turban beat Askar Zucchetto. We all know this is impossible, meaning Askar Zucchetto always wins over turban. Lower haste situations will just further the gap.

In Ace's helm vs. Askar Zucchetto, you basically come down to the crit chance from 4 dex vs. 4 accuracy. The 2% hitrate is always going to beat the .75% crit chance, so Ace's helm will always be better than Askar Zucchetto.

[Hakamaru]

I dont call a parse something with just one hundred weapon skills as sample...

You know... there are 4 links in that post, all full of parses. Many with much larger sample sizes. If I and countless others are wrong, it should be easy enough to post a counter-example.

[VZX]

In your situation of 70% haste, walahra turban's 1% haste would be a difference of 3.3% over Askar zucchetto.

.31/.3 = 1. 03333

The str on Askar would increase your fSTR by 1 point, which for my melee setup is an increase of 1.2%.

(75+10)/(75+9) = 1.012

The addition to crit chance from 4 dex is hard to nail down, but for me it would bring my dex from 96 to 100, which based on some parses in a previous post, would be approximately a .75% increase to crit chance.

This means the accuracy from the dex needs to make up 1.35%. This will occur at all values for melee accuracy.

(x+.015)/x = 1.0135
x = 1.111

So, only if you had over 111% melee accuracy would walahra turban beat Askar Zucchetto. We all know this is impossible, meaning Askar Zucchetto always wins over turban. Lower haste situations will just further the gap.

Where is the 1.35% coming from?

[Hakamaru]

This means the accuracy from the dex needs to make up 1.35%. This will occur at all values for melee accuracy.

(x+.015)/x = 1.0135
x = 1.111

VZX is right... this part is a little wonky. If you want to compare to haste, then you're interested in how many extra swings you get relative to the swings you are already getting. So... if your ACC is low, you're going to get proportionally more gain from the 1.5% increase in hit-rate (0.75 * 4DEX = 3 ACC -> +1.5% hit-rate btw).

If you want to get an idea of how things work out from just a swings landed perspective, I have a handy spread-sheet for that. Just type into the red boxes (please don't modify any other fields). For Askar, you are interested in 1% haste vs 3 ACC.

AccVsHaste - Online Spreadsheets - EditGrid

[trencher]

You know... there are 4 links in that post, all full of parses. Many with much larger sample sizes. If I and countless others are wrong, it should be easy enough to post a counter-example.

That would be work, it's easier for people to just ignore the math and say "omg i need acc! i once missed a ws!!!"

Parsed Dynamis-Xarc last night.

Melee acc was 71% (blind half the time from demon spam...)
WS acc was 95.7% (WS gear was all +str - Osode, smil+1, Alkys, shura head/legs, rutters, etc)

TADA!

but wait i must be wrong....

[ringthree]

You know... there are 4 links in that post, all full of parses. Many with much larger sample sizes. If I and countless others are wrong, it should be easy enough to post a counter-example.

I think he was getting to the point that parses are not perfect representations of anything. The are a tool just like calculations and hypothesis for determining an unknown.

[ringthree]

That would be work, it's easier for people to just ignore the math and say "omg i need acc! i once missed a ws!!!"

Parsed Dynamis-Xarc last night.

Melee acc was 71% (blind half the time from demon spam...)
WS acc was 95.7% (WS gear was all +str - Osode, smil+1, Alkys, shura head/legs, rutters, etc)

TADA!

but wait i must be wrong....

Actually, you are missing the point. I don't think anyone is questioning the accuracy on Dynamis or Merit mobs, because yes you will be capped on WS's against them. A better test would be against other mobs.

I never really understood why people didn't parse more interesting fights like HNM, Einherjar, Salvage, etc.

[Hakamaru]

I think he was getting to the point that parses are not perfect representations of anything. The are a tool just like calculations and hypothesis for determining an unknown.

I disagree. He looked at the pretty picture in my post and apparently that's all he saw. If you click the links, the first one is even a BG thread in this very section started by Ruke. His parse alone has 300 WSs and over 1400 melee hits. Shinta's parse in the Alla thread (the one I referenced in my first post here) has 200 WSs and demonstrates an enormous differential in melee vs. WS ACC. And of course there are a ton more.

I'm sorry you and mdk have never heard of this WS ACC thing, but it's really OFN at this point. If any of you are going to claim 1st WS hit isn't auto-capped to 95%, let's see some parses.

BTW... Hasso does stack directly with gear haste & magic haste. Why do you think Rune Chopper zergs were so great?

[Zoolander]

Actually, you are missing the point. I don't think anyone is questioning the accuracy on Dynamis or Merit mobs, because yes you will be capped on WS's against them. A better test would be against other mobs.

I never really understood why people didn't parse more interesting fights like HNM, Einherjar, Salvage, etc.

Sure as hell does sound like others are questioning it. You go from 71% acc with ACC gear on during TP and then take off that acc and WS and the acc goes to 95%, aka capped.

How is Einherjar and Salvage more interesting? Those are weak as hell...

[Hakamaru]

I never really understood why people didn't parse more interesting fights like HNM, Einherjar, Salvage, etc.

People do... Kirin would be a pretty awesome thing to gather data on. But anyone worth their salt plants a Feint on him so their mkris/kclub DRKs aren't wiffing up a storm. This will skew the data horribly.

I've parsed Byakkos with my low-ACC WS gear. I come out with capped ACC, but normally a sample size under 30... so why bother posting about it? You really need to look at the evidence from all the places we can easily gather data, then sort of convince yourself with your own parses in regards to higher level HNM type things.

[trencher]

I never really understood why people didn't parse more interesting fights like HNM, Einherjar, Salvage, etc.

Actually I parse a lot of fights. You want to see my ele resist rate on Tiamat?
(It's getting better in case you were worried)

But seriously, what HNM's are you really letting a ton of melee on?
Cerb? Stoneskin...
Khim? runing in to SA gekko and praying he doesnt eat your tp.
Tiamat? the rawr full morigans tarus force her up and down so fast nobody bothers coming on sam.
Ixion? yeah i dont really parse Ix cause i foolishly mostly go Rng and miss 300TP Sides and i dont want any more proof that RNG is broken. But I digress...

We parse a lot of Einherjar, but people come mostly RNG or WAR so not really much help there in terms of SAM or even 1 hit WS acc.



I'm sorry but Dynamis is the perfect place to parse.
we had between 6-8 Sams last night. (go go overheating laptops, so some d/c)
we were killing mobs for 2 hours or so, then did gun for fragment.

I had the other alliance dmg filtered, 3-4 sam in mine were all hovering between 70-80% melee acc and you guessed it 95% WS acc.

1 hit WS acc is pretty much a done deal.

[Bakthi]

I parsed Gurfurlur last night. Oh, laugh it up. But he has some evasion and serious DEF too, meaning that most DDs will have to balance attack and accuracy on him. We wanted a change of pace, and a few people in our LS might actually get a mythic one day (our leader is pretty set on the WHM one).

There were 3 SAMs in my party, all /THF to plant more hate on tanks. I was using Hunter's Roll and Chaos Roll with AF2 hat. None of them should have been using Rana, because we told them not to as you don't want to SC on him, so these should all be the 1-hit WSs.

SAM 1, full Usu: 89.47% melee acc, 75.00% WS acc. Hit 30 WS, missed 10. Now, that's probably a bit of a fluke, but I don't think you want to assert that if he managed to do 160 more WS, he was going to nail every single one of them.

SAM 2, mixed gear: 95.83% melee acc, 81.48% WS acc. Hit 22 WS, missed 5.

SAM 3, mixed gear: 87.95% melee acc, 95.83% WS acc. Hit 23 WS, missed 1.

SAM 4 was...not /THF (don't ask) and not in my party. 91.70% melee acc, 94.87% WS acc. Hit 37 WS, missed 2.

I am not claiming that this proves or disproves anything. SAMs 1 and 2 don't seem to support the idea that 1-hit WS acc is capped. SAMs 3 and 4 seem more in line with it. It's not a ton of data, but well...there is only one Gurfurlur, and it sucks enough to get to. If it helps or is interesting, cool. It's what I have.

I don't really know the TP and WS gear, or food, of all these SAMs, but I can find out if you think it will help. Though if it's supposed to be true that 1-hit WS acc is capped so far above melee acc, I'm not sure that would be relevant. We don't camp any kings, but if you want parses from just about anything else, I can keep that in mind. I try to parse everything, as I think it helps to see how people are doing. Hopefully each person will use the data to help improve themselves. If not...maybe the LS will use the data to improve itself.

Edit: In case it comes up, no, we did not have Feint for Gurfurlur. We needed everyone who managed to show up on a more useful job.

[WraithLakshmi]

VZX is right... this part is a little wonky. If you want to compare to haste, then you're interested in how many extra swings you get relative to the swings you are already getting. So... if your ACC is low, you're going to get proportionally more gain from the 1.5% increase in hit-rate (0.75 * 4DEX = 3 ACC -> +1.5% hit-rate btw).

If you want to get an idea of how things work out from just a swings landed perspective, I have a handy spread-sheet for that. Just type into the red boxes (please don't modify any other fields). For Askar, you are interested in 1% haste vs 3 ACC.

AccVsHaste - Online Spreadsheets - EditGrid

Yeah, I know.

I got 1.35% from:

3.3 - 1.2 - .75 = 1.35%

This is how much walahra turban is ahead before you factor in the accuracy from askar zucchetto. I realize the value for crit hits is a little shakey since research on that subject isn't complete.

I also am aware that the 1.5% hitrate will translate to a different increase in melee dps depending on what your current hitrate is. That's why the next calculation was there:

(x+.015)/x = 1.0135
x = 1.111

This shows at what current hitrate the 1.5% increase from askar zucchetto will translate to exactly 1.35% melee dps. Of course, 111% accuracy is impossible, and near cap some of the accuracy will not be used, so in actuality if you have 93.5% hitrate or lower, askar zucchetto will win. If you have 94% - 95% some of the accuracy will be cut off, making Walahra turban better. Against Greater colibri, this means if your accuracy is 405 or lower, use askar zucchetto.

[Hakamaru]

I parsed Gurfurlur last night. Oh, laugh it up. But he has some evasion and serious DEF too, meaning that most DDs will have to balance attack and accuracy on him. We wanted a change of pace, and a few people in our LS might actually get a mythic one day (our leader is pretty set on the WHM one).

There were 3 SAMs in my party, all /THF to plant more hate on tanks. I was using Hunter's Roll and Chaos Roll with AF2 hat. None of them should have been using Rana, because we told them not to as you don't want to SC on him, so these should all be the 1-hit WSs.

SAM 1, full Usu: 89.47% melee acc, 75.00% WS acc. Hit 30 WS, missed 10. Now, that's probably a bit of a fluke, but I don't think you want to assert that if he managed to do 160 more WS, he was going to nail every single one of them.

SAM 2, mixed gear: 95.83% melee acc, 81.48% WS acc. Hit 22 WS, missed 5.

SAM 3, mixed gear: 87.95% melee acc, 95.83% WS acc. Hit 23 WS, missed 1.

SAM 4 was...not /THF (don't ask) and not in my party. 91.70% melee acc, 94.87% WS acc. Hit 37 WS, missed 2.

I am not claiming that this proves or disproves anything. SAMs 1 and 2 don't seem to support the idea that 1-hit WS acc is capped. SAMs 3 and 4 seem more in line with it. It's not a ton of data, but well...there is only one Gurfurlur, and it sucks enough to get to. If it helps or is interesting, cool. It's what I have.

I don't really know the TP and WS gear, or food, of all these SAMs, but I can find out if you think it will help. Though if it's supposed to be true that 1-hit WS acc is capped so far above melee acc, I'm not sure that would be relevant. We don't camp any kings, but if you want parses from just about anything else, I can keep that in mind. I try to parse everything, as I think it helps to see how people are doing. Hopefully each person will use the data to help improve themselves. If not...maybe the LS will use the data to improve itself.

Edit: In case it comes up, no, we did not have Feint for Gurfurlur. We needed everyone who managed to show up on a more useful job.

This is a great start... and I think this sort of thing deserves a new thread dedicated to 1-hit WS ACC on HNMs. You've got 2 SAMs who are following the expected auto-capped behavior and 2 that don't appear to be.

The thing that really startles me about those number is that the 2 SAMs who appear to not have capped WS ACC, actually are showing WS ACCs less than their melee ACCs. There are enough parses with enormous collective data sets at this point that we can say with certainty that this should not be. Did you ever bust on Hunter's roll? I'm unfamiliar with the Gurfurlur fight / area. Is there any way for PCs to get a negative ACC modifier during the course of the fight? Did any of them fight any add that may have cast utsusemi or used perfect dodge?

From the perspective of 1-hit WSs should auto-cap to 95% chance to hit... it really shouldn't matter what your SAM friends were eating or wearing.

[WraithLakshmi]

Yeah, there must have been some kind of buff/debuff causing them to miss that much. I WS with almost no accuracy on fafnir, and the only times I miss are right after he uses hurricane wing and blinds me. I've always noticed similar things with other mobs, blind can really eat your 1 hit ws's accuracy.

[Weeks]

Years of abusing Besieged on lowbie jobs have made it pretty clear to me that 1-hit WS accuracy definitely isn't always capped. It's just a big ACC bonus, so I guess the next step is to start collecting data and eyeballing whether it's the same across all 1-hit WSes and thus the data can be pooled.

[Hakamaru]

Years of abusing Besieged on lowbie jobs have made it pretty clear to me that 1-hit WS accuracy definitely isn't always capped. It's just a big ACC bonus, so I guess the next step is to start collecting data and eyeballing whether it's the same across all 1-hit WSes and thus the data can be pooled.

It's almost certainly the same for the first hit of any WS (that's not ACC modded with TP... hi2u slugwinder). The fast blade test that showed gorgets were 10 ACC for all hits showed the first hit capped at 95% hit rate. I've done single-wield tests with wasp sting on NIN that show capped WS ACC with ~68% melee ACC and no gear swaps. I've seen spirit taker tests showing the same thing. Whatever the bonus is, it seems pretty universal... although, more significant to 1-hit WSs.

Lowbie job in besieged would be an interesting counter-example to the auto-capped assertion. Although, it still may be auto-capped for situations anyone cares about. It's possible that there's some level difference component that eats away at the 1-hit WS ACC bonus... but I never had problems maintaining capped WS ACC on kirin zergs or byakko.

Parses are key here. Before I posted parses over on Alla ~1 year ago, everyone seemed to know for a fact that WS ACC followed melee ACC precisely.

[Weeks]

My hunch is that ranged WS don't get an accuracy bonus period (And Slug/Side gets the TP-dependent penalty), but this could just be selective memory and will need to be parsed.

I'd say start collecting WS data for 1-hit weapon skills, eyeball whether the data can be pooled, and once the ACC bonus is nailed down, test to see whether it applies to the first hit of multihit WS too.

[Bakthi]

The thing that really startles me about those number is that the 2 SAMs who appear to not have capped WS ACC, actually are showing WS ACCs less than their melee ACCs. There are enough parses with enormous collective data sets at this point that we can say with certainty that this should not be.

Yeah, I think it's pretty weird too.

Did you ever bust on Hunter's roll?

I think I did once, but that only lowers my accuracy, not anyone else's; everyone else just doesn't have a roll. I used Fold on it immediately anyway.

I'm unfamiliar with the Gurfurlur fight / area. Is there any way for PCs to get a negative ACC modifier during the course of the fight?

Pleiades Ray does inflict Blind (along with almost every other status effect), but he only gains access to that during the last 20%, and only used that move once around 10%. It is possible that one or more of them didn't get Blindna.

Did any of them fight any add that may have cast utsusemi or used perfect dodge?

A few troll adds were fought, but trolls don't have THFs or NINs. It's possible they WS'd into Diamondhide, but that shouldn't show up as a miss. Um...it's possible one of them WS'd into a troll RDM's Blink, but I think I would have heard bitching about it! I can ask though. We only had 3 adds to fight ever, which happened due to people getting tossed off the cliff. In my wildest statistical fantasies, I can imagine that 1 WS each from 2 SAMs was somehow eaten by Blink, max. Maybe.