cratio for two handers

**Cadsuane** · 2009-03-18 01:15

BG wiki and ffxiclopedia both list the level corrected att/def ratio for melee as being: cRatio = Ratio - 0.050 x level difference.

The 2h update in August 2007 said this:

Two-handed weapons will now receive larger benefits from a character’s attributes, and will also inflict more damage when facing higher-level opponents or enemies with high defense.

and the subsequent nerf in November said this:

* The following changes have been made to two-handed weapons:

○Bonuses from attributes have been reduced.
○The method for calculating damage has been readjusted.

So after the first update the level correction function was presumably changed, was it changed back to be the same as 1h, or is the wiki information bogus and the cratio penalty for 2hers is now somewhere between 0.050 and 0.025? Or does no one actually know?

Also were pdif distributions changed for 2hers and subsequently changed back or are they still different and the research has been swept under a rug somewhere?

**shaddix** · 2009-03-18 01:21

I dunno, lol. I think the chance of having accurate information was destroyed when the banhammer commeth.

**Araelus** · 2009-03-18 02:00

This is relevant to my interests, especially regarding http://www.bluegartr.com/forum/68018...testing-4.html. Right now we are still treating 2hand level difference penalty as dLV*0.05, even though there were some odd numbers in the testing, since it seems to work out properly. But someone who knows more about what Wraith meant by

the multipliers for pDIF(max) aren't in the same ranges. However, the multipliers for pDIF(min) for levels 73, 74, and 75 are all around .75. In fact, the pDIF(max) values all appear to be equal to pDIF(min) + .6. So it would appear that pDIF(min) is calculated first, and pDIF(max) is based on that result, backwards from how it used to be done.

would raise an eyebrow at this.

**Cadsuane** · 2009-03-18 16:46

I'm unfamiliar with these formulas so feel free to point out any shoddy math.

Let's say I'm a 75 melee with 455 attack, 84 total STR and a 94 DMG weapon, meleeing Kirin, a level 92 mob with about 450 defense. We'll say he's been angon'd and dia III'd.

Finding Kirin's reduced defense:

initialdef * (1-(diaeffect + angoneffect)) = def
450 * (1-(53/512 + 51/256)) =~ 314

Then the level corrected ratio:

ratio - (0.050 * leveldif) = cratio
455/314 - (0.050 * 17) =~ 0.59

According to this graph: (Image

DIF.gif - FFXIclopedia - a Wikia Gaming wiki)

The upper bound for pdif distributions at cratio 0.59 is plateaued at exactly 1, which makes finding the max easy.
Now we mutiply this pdif value with the D value for this melee and can predict approximately what the max damage on a non-crit melee swing should be.

base damage = Weapon + fstr

This is actually where I don't know how to continue because I don't know what to guess for kirin's VIT attribute, though I can guess it's pretty high. But thankfully, I have a real parse that tells me the max this particular melee hits kirin for in the course of 45 minutes or so under these circumstances is for 110 damage. If we solve for fstr then:

damage = pdif * base damage
110 = 1 * (94 + fstr)
110 - 94 = fstr
16 = fstr

And that high a value for fstr on kirin is absurd for a melee with 83 STR.

So is the reported def value for kirin lower than 450?
Does Kirin have about 30 total VIT?
Did I miscalculate the def down? (are stacking defense downs multiplicative or additive?)
Are the reported pdif distributions wrong?
Or is the 2h cratio formula wrong?

**Araelus** · 2009-03-19 18:00

Reported def value for Kirin is closer to 600.

**Cadsuane** · 2009-03-19 18:30

That makes even less sense. Let's be generous and assume the melee was getting two minuets instead of the one I assumed, for 503 attack.

reduced defense: 600 * (1-(51/256 + 39/256)) =~ 389

cratio: 503/389 - (0.050 * 17) =~ 0.44

maxpdif: 0.4 + (1.2)(0.44) = 0.93

damage (94 + fstr) * 0.93 = 110

and fstr = 25.2

EDIT: corrected the defense down calculation. Getting a bit closer to sanity.

The 110 is a non-crit, of course, so I'm not making that mistake. And there's no further means of reducing defense in the game, and I'm certain the melee in question had one BRD and was using sushi, so there's no means again to improve att further. The max crit that parsed was for 206. Something's seriously fucked.

**Araelus** · 2009-03-19 22:23

Ok... I'm reading your math and looking at the charts and I can't figure out where it all goes wrong. I'm going to tidy up the variable names when I quote, but otherwise try to follow your math from start to finish. One conclusion I could make is that your parse data is wrong because the values for max normal hit and max critical hit produce different fSTR values.

Spoiler: show

TL;DR Spoiler maths is HUGE.

This would indicate the existing cRatio → pDIF formulae are incorrect.

Something is going wrong here, and I tend towards it being an input rather than the formulae. But at this point there could be something wrong with the math itself.

There is one solution, if all the inputs you provided are correct, and that is a lower than expected defense value for Kirin. This would mean breaking the pDIF(max) = 1 plateau, and since he is hitting a higher pDIF than you expected, his fSTR is lower than you calculated.

Another solution is that 2h have different tiers for cRatio (perhaps 4 instead of 3) or a built-in multiplier to cRatio that is then subject to a second level correction.

**Cadsuane** · 2009-03-20 19:32

Normal monsters of level ~82 have 3.99 DEF/mLV, and HNMs of both level 87 and 95 have ~6.32 DEF/mLV. This would suggest Kirin's defense is ~580. But let's suppose the 450 defense value just for the sake of your argument.

If Kirin has higher defense the cratio would be lower, which makes my inputs entirely too generous. I made the scenario more plausible in the second example, with 600 defense and two minuets instead of one.

The entire 45 minute parse with max melee non-crit hit of 110 damage was with both of these effects in play the whole time? Ok, then, Angon + Dia III is 90/256 defense down. 450*(166/256) = 291.796875

If the parser reads out the highest recorded hit made by that melee, it stands to reason that it would've happened when dia III and angon had both been stuck. There's no guarantee that the max normal or max crit damage represent maxpdif, but the non-crit in particular is probably pretty close. Besides, the problem is that higher numbers than you'd expect are coming out, and if even higher ones are possible it's only further from the prediction drawn by accepted damage formulae.
I'm a bit confused that you decided to run over the numbers I didn't correct.

pDIF = Damage Dealt / Base Damage
1 = 110 / Base
Base = 110
94 + 16 = 110
fSTR = pSTR - mVIT
This implies that Kirin's VIT is 67. That would be the same as a lv82 Greater Colibri.

I'm pretty sure this is not how you calculate fstr.

Let's use the 206 critical for a cross-check of the math.

Critical Hits provide pDIF +1, so if pDIF(max) was 1, CRITpDIF(max) is 2.

206 = 2 * (94 + fSTR)
103 - 94 = fSTR
fSTR = 9

Still confused why you decided to use the input that underestimates kirin's defense, though it would only end up being a loss of 0.07 cratio if you had used my second set of inputs.

So from this we realize that it's possible the pDIF value for his maximum normal hit was actually higher than 1. This would make the actual fSTR lower than 16. Then again, for pDIF to output a value above 1 on a normal melee hit vs. Kirin, Ratio must be above 1.7 (based on required cRatio and Kirin's level correction). This requires Kirin's defense to be below 267!

As you've pointed it out, it's unlikely that pdif max was higher than 1, particularly since this projection already assumes kirin has been affected with the largest possible amount of defense down available to us in the game.

I've got an excell sheet to do this math now and for the minimal data I have, yeah, the fstr values don't match up. Two possibilities I submit is that the pdif distribution is actually not just a vertical translation of +1 for critical hits, or that the 206 crit is actually not the highest amount of damage you can hit for. But to get the same fstr on a crit it has to be for exactly 228 damage, and it might be the case that these higher numbers never procced when we're only talking about a sample of 16 crits, many of which would not have happened during the angon effect.

The thing about the fstr value in this case is that in a TP set with less than 100 STR, it's supposed to be impossible to cap it on anything less than a negative level con for any rank of weapon, and on kirin I'd expect fstr to be negative if anything. A value that's not only positive, but close to cap or past it, is definitely a good suggestion that something is seriously wrong.

**ringthree** · 2009-03-20 19:35

Originally Posted by shaddix

I dunno, lol. I think the chance of having accurate information was destroyed when the banhammer commeth.

Because it was being investigated by all those super awesome people that were banned?

Two hand cRatio and pDif have been wrong on the wiki ever since the two handed update, I proved that on my blog, but no one has tried to find out the new calculations.

**Cadsuane** · 2009-03-20 19:42

Can you show us your proof?

**Araelus** · 2009-03-20 20:03

My spoilered maths prove RingThree is correct. Using the data you gave me and the existing formulae produce self-referentially contradictory results.

**ringthree** · 2009-03-20 20:18

Originally Posted by Cadsuane

Can you show us your proof?

Yes, I tested it over a year ago so it will take a minute to shuffle through blog posts.

Here it is: http://ringthree.blogspot.com/2008/0...mind_7172.html

The three or four posts before that was a discussion of the mechanics of melee damage, I was going to use my testing to explain it better but it just proved that the equations were wrong.

**Cadsuane** · 2009-03-20 20:48

I was pretty tired of having obsolete damage math thrown in my face anyway.

There's a strong example for HNMs and for merit mobs, then. Someone who isn't me should edit the wikis or maybe try to find a new formula. Maybe move this to advanced so it can fall off the first page quicker?

**Araelus** · 2009-03-20 20:59

Thanks for the maths consideration, RingThree. I'll talk to VZX and see what we can do about redetermining cRatio tiers. I'll have to search for a level 75 mob where we can set Ratio (and thus cRatio) to 1, fSTR to 0, etc. And then get 10,000 + results for a distribution curve.

**shaddix** · 2009-03-20 22:07

Originally Posted by Araelus

Thanks for the maths consideration, RingThree. I'll talk to VZX and see what we can do about redetermining cRatio tiers. I'll have to search for a level 75 mob where we can set Ratio (and thus cRatio) to 1, fSTR to 0, etc. And then get 10,000 + results for a distribution curve.

I'll paypal you 10 bucks for solid info on 2hander damage formulas!

**Araelus** · 2009-03-20 23:24

Right now the only obstacle from me starting tests is getting a parser that will output my damage per hit data in a usable form.

Step one will be to find a level 75 mob with a known VIT and DEF. This will remove level correction and fSTR from the picture and let me be assured of Ratio.

I could work using Greater Colibri, and I will if I have a parser, but it will be a little shakey, using Excel to fit more than a few variables.

I'm sure someone out there is more prepared than I am, but I'm motivated to sitting down and getting thousands of attack rounds of data so long as I don't have to write the damage by hand.

**shaddix** · 2009-03-21 06:16

You don't use kparser?

**Araelus** · 2009-03-21 10:59

If kparser does what I mentioned with individual attack round data output, then I will use it.

This thread? http://www.bluegartr.com/forum/50852...parser-19.html

Reading everything I can on VZX's website. Still looking for stat information on common mobs.

**Thunder** · 2009-03-21 11:42

Kparser does show the damage of each individual attack you've performed during the duration of the parse. It's under Offense Detail, I think.

Although.. usable form? It just lists every single digit of damage you've performed and then next to each digit a number telling you how many times you've hit for that high. Not sure if that's what you mean by usable or if you meant something like a parser that already puts it into a spreadsheet for you.. or something.

**Araelus** · 2009-03-21 11:53

Having it output the damage value and a frequency (what you said) is fine. Pasting it into Excel shouldn't take too~ much time... I hope.

Based on RingThree's tests, pDIF does not conform to a normal distribution, but I still want to have all the data points between pDIF(min) and pDIF(max) for the sake of statistical rigor.

I still lack exact defense/VIT values for, say Steelshells (level 73 - 76 which I'll have to account for in the data collection after each kill based on xp) or more preferably for some of the tests, a lower level mob such as Colibri (Thickets, level 71-73) since those should not have any level correction effect to account for.

To start, dealing with a lower level monster is great if it caps out fSTR (remember I am trying to eliminate as many variables as possible since we have no idea where the Ratio tiers are located and Excel's solver functions can only do so much) in which case I do not need to know VIT, just an actual defense value.

Thread: cratio for two handers

Thread Tools

cratio for two handers

Similar Threads

Two hander Pdif question

Latest Threads

Up & Coming Threads

Hottest Threads