Damn you, Auto-Attacks!!

Valky had a Theory back in Beta Phase 3. His assumption, if he was correct, was that the only stat that scales differently between Auto-Attacks and Ability Attacks was Determination. Our original formula which was made by EasyModeX, as well as the Japanese Formula, were both constructed with this in mind. But, this rule does not fit my data samples what so ever.

Let’s look at the solver I’m using, and yes EMX, this is your original solver, albeit modified slightly.

Continue reading “Damn you, Auto-Attacks!!”

Advertisements
Damn you, Auto-Attacks!!

Determination Update v2

If you managed to make it through my previous Determination Update post, you would of seen that I posted an updated Ability Damage Formula for the Dragoon.

((WD*0.0402502+0.9209679)*(AP*0.0447422+2.6089895)*(DET*0.0011331+1.8536433))

I wanted to streamline it a tad bit more. I was messing around with a few more data sets, and I realised that that the +X parts of Weapon Damage and Determination were fairly static. They were always accurate to the thousandth decimal place. I also noticed how close these numbers actually were to 1… And I thought to myself… “are they just multipliers and nothing more?”

Oddly, my assumption was accurate.

((WD*0.0437052+1)*(AP*0.0763653+4.4581293)*(DET*0.0006116+1))

Or, a rounded version of:

((WD*0.044+1)*(AP*0.076+4.46)*(DET*0.00061+1))

This formula above streamlines all of the excessive numbers I used and also provides me with a damage variation exactly similar to my previous formula.

I still need to gather low AP/High DET data samples and run it by this formula, but it’s looking like this is the expression I’ll be sticking with.

Now that I’ve established the final Expression will be (WD*A+1)*(STR*B+C)*(DET*D+1), the next thing I’ll be doing is continuing my work on Job Coefficients. As we all know, each job affects damage differently and what I’d like to do next is find out whether jobs affect the value of stats individually, or if it’s just another external coefficient.

Determination Update v2

Determination Updates – Half Way there

It’s been a while since I did my last update. I thought I’d share some of my work thus far. I’ve spent the last couple of weeks gathering additional armour pieces so I can eventually have a massive spread of data for Determination information. I’m also working on obtaining more Bard, Ninja and Warrior FCOB armour pieces, so I can assess Job Coefficients more carefully and whether Jobs add an additional Coefficient to your Main Stat, Weapon Damage, or both (as a multiplier at the end).

Anyway, here’s what I’ve got so far.

Continue reading “Determination Updates – Half Way there”

Determination Updates – Half Way there

Influence of Weapon Damage and Strength on Damage

The first stage reworking the Damage Formula is to go over and validate all previous Theorycrafting work, which, will decide how I approach data collection. One of the most important things is understanding the influence of Strength on the Damage Formula. Why? Well, according to Valks work, we can still continue to deal damage even with 0 Weapon Damage, meaning Strength on it’s own, directly influences our damage. This also means regardless of our stats, we will always have a base value of damage.

If there is no “0” Weapon Damage value in the game, then we do not have a base damage value. If we do not have a base damage value, then Strength is not, or doesn’t hold any independent properties which can alter the damage formula. This is a very important mechanic, as it absolutely influences how the Damage Formula is shaped.

If Strength doesn’t have its own damage value, then the Damage Formula would simply be something like:

  • WD*STR*DTR

If Strength operates how we currently assume it does, then we’ll have a formula which resembles:

  • WD+(STR*WD)*DTR.
  • or, ((WD*STR)+STR)*DTR

Both formulas allow Strength to be influenced by Weapon Damage, but one makes it so Strength has a base value, whereas the former doesn’t.

Continue reading “Influence of Weapon Damage and Strength on Damage”

Influence of Weapon Damage and Strength on Damage

Reconsidering Determination

Before we begin, I’d like to give a special thanks to Arisue Neetsha for making me reconsider how Determination works, Sunny Hirose with supporting evidence, as well as an active poster who goes by the name Kaesebrezen.

Determination is a stat that has largely been misunderstood for the most part. No one actually knows, even to this day, how it works, or how it’s implemented into the Damage Formula.

Originally, Valk who created “Version 1” of our damage formula, used these correction factors for our damage formula:

  • Auto-Attacks: + (DTR-202)*0.11
  • Weaponskills: + (DTR-202)*0.035.

Yes, this was incorrect. The main thing that Valky never managed to understand is how other stats impacted the effect of Determination. Has he used a “+DTR”, he basically stated that Determination was an independent value which was added on top of the remainder of the damage formula… which is incorrect.

Continue reading “Reconsidering Determination”

Reconsidering Determination

Job/Classes & their impact on in-game Damage.

Back during Beta Phase 3, Valky had a Theory. He assumed that every job had a unique damage formula, or a modifier for the damage formula. He noticed that during his testing with the Monk, the Bard, Dragoon, Warrior and Paladin all gave him different damage ranges, even though his formula was 100% accurate for his Monk. This kind of makes sense and I can see why Square Enix might do this. If every job has their own damage formula, or a modifier for the Damage Formula, it makes it extremely easy to balance out Job Damage – Too easy in fact.

I’ve been reading up on the Japanese Theorycrafting recently as well. The Japanese took Valkys work to heart and used it as the basis of all of their Theorycrafting. What they’ve done on the other side of the world actually resembles Valks work quite a lot. Up on reading numerous, badly Google translated blogs and web pages, I noticed something… and it got me thinking.

Was Valky correct all this time?

One thing I’d like to point out now is that our Damage Formula was constructed dismissing this belief. It was made as an “all-purpose” damage formula. It just simplifies everything a lot more.

So, I got kind of bored and I said “Fuck it! I’m going to play with my Training Dummy some more”… And I found something which kind of surprised me.

I parsed both a Dragoon and a Ninja spamming a 150 potency skill, with a set which gave me the exact same same stats:

  • WD: 52
  • STR: 365
  • DET: 224

And to my bemusement… There was a difference. A noticeable difference. The Dragoon’s min/max damage values were 2 more than that of the Ninja. Now, I thought this was Bullshit at first and could of been caused by Rounding Errors as I’m using low-magnitude data to compare. But, if the damage formula was consistent for all Jobs/Classes, then the Rounding Errors would of been exactly the same for both…

So I then repeated the test with a Lancer and a Rogue. Again, using a 150 potency ability and the exact same stats as above, my parses are completely different, The Lancer and Rogue were both below that of Jobs, but the Lancer was still above that of the Rogue.

So, I then recorded data for the Paladin, Warrior, Marauder and Gladiator, then put all of my information into one neat table, as you can see below.

Yeah. There was definitely a difference in damage values. What amused me the most was the Gladiator, which was completely below any of the other Jobs/Classes. One thing I noticed was that jobs/classes might well be “tiered” or “grouped”, which uses the same formula or modifier. You’ll see that a few Jobs/Classes, such as the Lancer and the Warrior had the same min/max values. This could be rounding errors, who knows? But that’s why I’m gathering more data. For now, here’s the “Tiers”:

  • Tier 1) Dragoon
  • Tier 2) Ninja
  • Tier 3) Warrior & Lancer & Marauder
  • Tier 4) Rogue

I’m not going to go too in depth into the mathematics for now. I’ll do that later on. But for now, I’m posting all of my observations and findings so people can see them.

So yes, this does in fact confirm that different Jobs/Classes have different damage formulas, or they impact the damage formula differently. Maybe it’s a modifier, maybe it’s not. We’ll see. But for now, I need more data.

Oh! Also here’s a download link for the spreadsheet which recorded all of my data if you want to take a look.

http://www.filedropper.com/jobdmg-52wd365ap224det

Job/Classes & their impact on in-game Damage.

Why we need a new Damage Formula.

There’s many reasons for this.

The most important reason… Is that it’s just not that accurate. Take this for example:

Hi Sunny!

On the right, we have our parse of a 360 potency skill, Full Thrust, then we have our Minimum + Maximum calculated. Min+Max/2 = Average damage of our parse. Damage in this game has a damage range of +-5%, or *1.05-*0.95 damage ranges. You know your data has reached “accurate” values if you do Max/Min damage and you get a number close to 1.1053.

Our parse has an average of 847.2166667 damage. This does not include Crits, as it would skew our data.

Then at the bottom, we have our expected damage which uses the current Damage Formula… Our Damage Formula gives us a value of 835.8588004 damage. Thats 1.01358826% less than what the actual number should be.

That’s the first issue with the damage formula. It’s inconsistent. Yeah, maybe back during Patch 2.0-2.2, the damage formula was accurate. But that’s mainly because our Damage wasn’t that high. Because the damage wasn’t high, this greatly caused in-game rounding errors. The formula we currently use only matches a certain data set which only fit both Dragoon and Black Mage damage, which used low-magnitude data. The lower the numbers, greater the worse rounding data will be.

The next issue, is that the Damage Formula uses something known as a “Correction Formula”. Instead of identifying a pattern of the linear scaling of Determination from level 1 -> 50 and beyond, our data simply starts from our base of 202. This correction formula was present in both Valkys original formula, EMX’s formula (which we use) as well as the Japanese Damage Formula.

This will cause MANY issues throughout the years as we go through more and more Expansion packs and level increases.

Image