Accuracy

[Khamsin]

So, I was in a lowbie party earlier with p.charm, venerer/jaeger rings, sole sushi, a capped A+ skill weapon, and the mobs checked as low evasion... and I still missed over half my attacks.

Is there a particular reason behind this? If I miss that much with good accuracy gear and sushi, I can't imagine how horrible my hit rate would be without sushi or p.charm or venerer/jaeger rings. I probably would be missing 90% of my attacks in that case.

It just doesn't make sense! Anyone with some knowledge of the mechanics and math behind all this have any insight?

[Thorshammer]

lowbie = level sync? If so... no wonder... this will gimp the fuck out of you, no matter how much acc you have.

[Klain]

If you're over camping, level correction.

[Kongolo]

try to zone or relog, or even restart the whole game.

..but couple the worst luck u got, in the end its all about %.

/sigh

[mdkuser]

low evasion means 80% accuracy minimum I think.

[arshesney]

Depends mainly on what level you were sync'ed to (guess sj, level by the gear you listed), that level range is "borderline" for sushi, where it is not a waste anymore, yet no that big of a boost.
The check message doesn't account for level correction, it's a mere acc/eva comparison. For every level it is something like -4 acc to you (not sure if that's the actual number), and an IT is at least 6 levels higher.

Base hit rate is 75%

Low Eva means that you have at least 10 more acc than mob's eva:
75% + (Acc - Eva) x 0.5%

Then Level correction is factored in:
75% + (Acc - Eva) x 0.5% + (Your level - Mob's level) x 2

If you barely were in the "low eva" zone whould mean a 68% hit rate (if the mob you were fighting was 6 levels higher)

[Zelduh]

From my experience accuracy is retarded at low levels. You could get a 75 BRD to soul voice double madrigal you and a COR to hit XI on Hunter's. You'll still miss a lot.

[Phla]

as stated above.. most likely level correction, sometimes lowbie parties overhunt on mobs... either way low level accuracy is pretty bad regardless, it happens to everyone lol, give it till around 40-45ish, more options open up and your hit rate becomes pretty decent

[The Stig]

Evasion is retarded too at low levels, I remember fighting an EP gob in the dunes with nearly 40 evasion in gear+food, I was level 29:

Emp Pin = 10
Battle Gloves = 3
Kingdom Trousers = 3
2x Dodge Earring = 6
Night Cape = 3
Beetle Harness + 1 = 1
Jack O Lantern = 10

39, and I was still getting hit regularly. This stuck out in my mind because I clearly remember being a bit annoyed about it.

[Tsuko_Asura]

Keep in mind, while lower level parties do tend to over-camp, accuracy really isn't an innate issue during that period. If you're fighting the things that you should be fighting, the level difference should actually be lower than what you grow accustomed to in the later levels, which is why your accuracy is fairly high in places like Qufim.

[Kirschy]

I haven't had a chance to fully prove it yet, but I feel there is a "fDEX" portion of the accuracy function used in determining hit rate. Using /check only shows you the ratio of your accuracy to the enemies evasion. This is similar to /check of defense which is only shows you the ratio of your attack to enemies defense (only half of the equation used in determining damage dealt.)

The following test you can perform to give you an idea:
Use a character at level 15 and fight a VT /w low DMG weapon. Have a BRD + COR in the party and use 2HR Madrigal/Rng Roll, even toss on a sushi. Parse hit rate. It won't be 95%. Add a DRK and have them use Abs-agi, or have another BRD use DEX+ songs and re-parse hit rate. It'll be much higher or capped.

There are other explanations for why this works, which is why I can't say it's definite. It's very hard to test due to nature of accuracy needing large sample sizes for accurate data. It also requires the use of many characters =/. Also, it would have to explain why lower levels have this problem, but higher levels can hit 95% having an fDEX around 0.

[warrior2582]

So would Dex rings actually be better if you are using sushi with a brd and/or cor in lower lvl parties?

[Max™]

*has known this for years, it is why he preaches the glories of Shield Break, because it reduces the ENEMY eva by a flat number before worrying about level correction issues*

[Narol]

But Shield Break means greataxe, means you miss out on the holy shit awesome that is sword and its pretty Red Lotus Blade.

Then again I haven't been to Dunes or even Qufim in years. I have no idea if low-level gimps are the same as always or what. Either way, Shield Break is useful all the way up to Raging Rush, and it makes me sad to see WARs not using it at all.

[VZX]

I haven't had a chance to fully prove it yet, but I feel there is a "fDEX" portion of the accuracy function used in determining hit rate. Using /check only shows you the ratio of your accuracy to the enemies evasion. This is similar to /check of defense which is only shows you the ratio of your attack to enemies defense (only half of the equation used in determining damage dealt.)

The following test you can perform to give you an idea:
Use a character at level 15 and fight a VT /w low DMG weapon. Have a BRD + COR in the party and use 2HR Madrigal/Rng Roll, even toss on a sushi. Parse hit rate. It won't be 95%. Add a DRK and have them use Abs-agi, or have another BRD use DEX+ songs and re-parse hit rate. It'll be much higher or capped.

There are other explanations for why this works, which is why I can't say it's definite. It's very hard to test due to nature of accuracy needing large sample sizes for accurate data. It also requires the use of many characters =/. Also, it would have to explain why lower levels have this problem, but higher levels can hit 95% having an fDEX around 0.

If what you're saying is true, we lol at elvaan dex then.

But Khamsin, can you mention where did you partying and what level were you? I remember Yuhtunga/Yhoator is quite an overkill of level difference.

Generally speaking, you push your level difference further when you were leveling up compared to when you are meriting.

[Suiram]

The thing that I am skeptical about the "fDEX" idea is that DEX and mob AGI are already factored into /check. If adding 3 acc will change /check message, then so will adding 4 or 6 DEX (depending on weapon). Is the idea here that DEX works in two parts, once as a part of accuracy, and again for this separate component? But what inconsistencies would this address that the current model does not (current model being basically what arshesney wrote above)? Maybe some WS issues but I don't think we've ruled out a flat acc bonus have we? I don't know what to make of the thing with the low-level experiment except that possibly the bonuses, which feel like they should be suffient to cap, simply aren't, while the extra dex puts them over the top.

[Kirschy]

I'm using a parallel of the attack function out of simplicity and the fact you can model damage equations this way. (Note I'm purposely simplifying the following equations.)

(dINT + spell DMG) * MAB/MDB Ratio = Magic Damage

(fSTR + wpn DMG) * Attack/Defense Ratio = Attack Damage

It wouldn't seem a far stretch to assume that accuracy works in a similar way.

(fDEX + ???) * Acc/Evasion Ratio = Hit Rate %

While I'm not saying this is definetly how accuracy is calculated, I think it's a more accurate model than just basing it on accuracy vs evasion. While adding 4-6 DEX can change a /check evasion, adding 4-6 STR can change a /check defense. Adding DEX or STR also adds an amount of Accuracy or Attack respectively. The /check only helps in determining one half of the equation in attack, and I'm making the jump to infer it works the same way for accuracy. It explains why lower level jobs are almost incapable of 95% accuracy rate, and an apparant accuracy cap that can only be changed by DEX/AGI changes.

It doesn't however explain how an accuracy % is formed from the equation with boundries of 20%-95%. I also admit I haven't put much time into testing testing and fleshing out the theory~