Does ATK cap out?

[neighbortaru]

this is very good info here. I'd second the request for a sticky.

[Aeynrieus]

I've been using this formula extensively in the past couple of days, working out, mathematically, just how Monks compare to other job classes. One of the things I've come across is that your statement:

There are two caps on attack.

1.) Attack=999

2.) A more practical cap of (Attack/Defense)*Unknown+random(+/- .5) = 2

You'll never under any circumstances reach condition 2 against an I.T. leveling mob.

... is very inconsistent with the formula that you've presented:

Rate = Centered Damage + Random(+/-.5)

Centered Damage = [(Your attack / Enemy's Defense) * unknown]

With unknown = [(Your attack / Enemy's Defense) * (Your attack / Enemy's Defense)]

or simply:

Centered Damage = (Your attack / Enemy's Defense)^3

According to my calculations, your Attack would only need to be 26% higher than the enemy's defense to cap your Centered Damage. And, in the converse, it'd only need to be 26% lower than the mob's defense in order for you to hit the floor. That's easily attainable against an IT mob. Since the cap on Centered Damage is 2.0, you can see that your Attack would cap at only 126% of the Mob's defense.

Here's an example of what I'm talking about. For the purposes of this equation, let's assume that a level 40 Elvaan Dark Knight/Warrior loaded with good gear (I'll provide a complete list of what I used, if requested) is fighting against a level 50 Galkan Paladin/Warrior (used a Galka PLD because it's pretty much impossible to find out the actual specs of mobs). The Galka Paladin has 63 VIT and, is assumed to have 230 Defense. The Elvaan Dark Knight is eating Meat Mithkabobs on top of his gear. His weapon is a Mythril Scythe +1.

His base stats, plus boosts from equipment/food:

STR 62
DEX 43
VIT 43
AGI 35
INT 32
MND 27
CHR 29
ATT 231
DLY 513
DMG 65
ACC +15

Base DMG:
(STR - MobVIT) + (24 + [3 * {WeaponDMG / 9}])

Base DMG = 44

Centered DMG:
(ATT / MobDEF)^3

Centered DMG = 1.01

Damage Caused:
(Base DMG * Centered DMG) * Random(assume 0)

Damage Caused = 44

*****

Now, let's see what happens when he turns on Berserk (+25% Attack):

Base DMG:
(STR - MobVIT) + (24 + [3 * {WeaponDMG / 9}])

Base DMG = 44

Centered DMG:
(ATT / MobDEF)^3

Centered DMG = 1.96

Damage Caused:
Base DMG * Rate(Centered DMG * Random(assume 0)

Damage Caused = 86

*****

There are two problems here. First, you can see how easy it is to cap your Attack. Second, Berserk does not double your damage. Now, of course we're working with a control here, and the random component isn't thrown in, but I think it's pretty apparent that the way that the way the Damage curve as it is affected by Attack is not right.

Either the formula is broken, or I've misunderstood it. However, I went over your post five times, and the website three times to verify that all my steps were correct. My best estimation is that the algorithm for "unknown" is way off. This entire formula is entirely too linear. I'd expect some kind of logarithmic function as the comparison of your Attack versus the Mob's Defense.

Please feel free to correct me if I've made some kind of mistake along the way, and I'll eat every last word...

[Aneyeisuponyou]

It's a tough read I'd like it stickied so I check on it now and again, that's alot of stuff to remember. ;p

Edit: So I'm lvl 67 ( h2h 236, using spartan +20D )

So my Max dmg is

3+ ( 236/10)+20= 46

24+3*(46/9)=39

So if my str is at least 39 I'm good to go?

[Mithrael]

1st, as it's stated, the "unknown" portion of the equation is unknown. The suggested unknown does show about the right shape of the curve, but it may not happen as quickly as (cap = mob's defense *1.26). You said the graph of (Attack/Defense)^3 is too linear, that it should be more like a logarithm. I disagree. Obviously, it's not linear at all; It's exponential. But what is your logic behind it being logarithmic anyway? If it were a logarithm, it would be true that attack has diminishing returns, but it would also be true that defense has increasing returns. Ask any high level (read: above lv.60) PLD, they'll tell you that adding defense after a while does almost nothing (unless you're fighting HNM with really jacked up attack). If it were logarithmin, to cap attack you're attack would have to be so far beyond the defense of the mob, you'd never achieve realistic damage figures. This isn't really likely either, since it's not all that hard to cap your damage on stuff that's only 10-15 levels below you.

Anyway, what is known is that your rate goes from 0-2. This is easy to see, just go kill too weak stuff with low defense, you won't always cap your dmg due to the random component, but you'll come close enough often enough to see that the equation works.

It's also obvious from any tests that you do on your own, that attack/defense follows a curve upward. To test this, remove all of your attack+stuff, don't eat, don't use berserk. You're a WAR, so you can click on defender (though be careful here because you may start running into the issue where your rate=0 but you're still doing damage due to your STR exceeding the Mob's VIT). Add additional points of attack slowly over the course of several hours watch your real damage per swing (not per second) very closely (you'll need a parser or an fast mathematical mind to keep track of it and average it out), as you increase your attack, your damage goes up greater in proportion to each additional point you add. Of course to pull this test off, you'll have to wait until you're in at least your 50s to have a long enough TNL that you already have your weapon capped and you won't accidentally level in the middle of your tests. If you find that attack +75 doesn't have an exponential relationship to attack+5, attack+10, post your results and we can try to reverse engineer a better version of 'unknown.' I'd be amazed if you saw that the relationship was logarithmic or linear, but as i have not had the opportunity to perform a real lab test like this, I won't rule it out.

Slightly off topic:
The other thing to consider is that you're fighting mobs 10 levels or more above your head in leveling situations. It's possible as a lv.56 Elvaan DRK/WAR to still con Robber Crabs as high defense while eating mithkabobs and using berserk. Further, you should know quite well by now that an "Even Match" mob has quite higher stats than you do. Of course, none of this answers the suspect nature of 'unknown.' But that's why it's marked unknown to begin with...


As far as the equation goes, you don't seem to have it quite right.

You write:

Damage Caused:
Base DMG * Rate(Centered DMG * Random(assume 0)

This is incorrect. It should read
Base DMG + {Weapon DMG * [Center + Random(assume 0)]}


Also, Base DMG is not:

(STR - MobVIT) + (24 + [3 * {WeaponDMG / 9}])

but instead:

STR-VIT ***up to but not exceeding*** (24 + [3 * {WeaponDMG / 9}])

The two are not added together, it's one or the other. you can never exceed (24 + [3 * {WeaponDMG / 9}]) as your base damage, regardless of how much STR you add. Base damage in both cases is -1, since the galka PLD's VIT is higher than the DRK's STR. Remember, Base DMG can be negative.

Your example should read...

Example 1:
Base DMG = STR-VIT = -1
Center = (231/230)^3 = 1.01
DMG from part 2 = 65*1.01 = 65.65
Actual Damage inflicted = -1 + 65.65 = 64

Example 2:
Base DMG = STR-VIT = -1
Center = [(231*1.25)/230]^3 = 1.98
DMG from part 2 = 65*1.98 = 128.7
Actual Damage inflicted = -1 + 128.7 = 127

(Note: I have a very hard time believing a galka PLD only has 230 defense at level 50. I could have easily had ~230 defense minus the two defense boosts PLD get over /WAR, a high defense shield, and the contribution to defense from a Galka PLD/WAR's superior VIT... You also don't have the PLD eating food, which is not a realistic comparison. Even then, I doubt the PLD's VIT is only 63, that's pretty low. However, I left your example alone just to show what the results should have been had the formulas been correct. Change the PLD's defense to 300 and vit to 85 and you'll see more likely (if not completely accurate) results to what you'd see on a leveling mob.)

-----

Aneyeisuponyou,

your max damage would be:

Weapon DMG = 46

Maximum Base DMG = 39

Weapon DMG*2 = 92

Total maximum damage = 46+92 = 138

If your STR is at least 39 above the Mob's VIT, you're good to go. (I say STR+39 on top of your normal stats as a generalization, may need more or less depending on the type of mob). I assume that's what you meant since your STR is probably somewhere in the 60s.

[Aeynrieus]

Originally posted by Mithrael
This is incorrect. It should read
Base DMG + {Weapon DMG * [Center + Random(assume 0)]}

Example 1:
Base DMG = STR-VIT = -1
Center = (231/230)^3 = 1.01
DMG from part 2 = 65*1.01 = 65.65
Actual Damage inflicted = -1 + 65.65 = 64

Example 2:
Base DMG = STR-VIT = -1
Center = [(231*1.25)/230]^3 = 1.98
DMG from part 2 = 65*1.98 = 128.7
Actual Damage inflicted = -1 + 128.7 = 127

Ahhh, now this makes far more sense. Thanks for the clarification.

Originally posted by Mithrael

(Note: I have a very hard time believing a galka PLD only has 230 defense at level 50. I could have easily had ~230 defense minus the two defense boosts PLD get over /WAR, a high defense shield, and the contribution to defense from a Galka PLD/WAR's superior VIT... You also don't have the PLD eating food, which is not a realistic comparison. Even then, I doubt the PLD's VIT is only 63, that's pretty low. However, I left your example alone just to show what the results should have been had the formulas been correct. Change the PLD's defense to 300 and vit to 85 and you'll see more likely (if not completely accurate) results to what you'd see on a leveling mob.)

Quite true. They would have much more than 230 Defense. The reason I used that number is because I already had that example built in another scenario where I was comparing DRK to MNK. It was a time-saver is all. I can tell you that a level 50 Galka PLD/WAR will have 63 VIT, though. At least, according to this:

http://ccwf.cc.utexas.edu/~nhart/Status_calculator.htm

This whole algorithm seems much more viable now that I know the Weapon Damage is added to the Base Damage in the final calculation. The only contention that I still retain is that the algorithm presented for the attack to defense comparison is flawed. I could certainly envision situations where it would probably be fairly close, but not for most.

At the end of the formula, the author presents these three solutions for determining your Centered Damage by examining the enemy:

monster has higher DEF -> your center is 0.0-0.5
monster has normaly DEF(no text) -> your center is 0.5-1.5
monster has lower DEF -> your center is 1.5-2.0

In this particular instance, if the mob checks Low Defense to you, and you active Berserk, you will be capping your attack according to the algorithm: (Your Attack / Mob's Defense)^3. For a mob to check Low Defense to you, your Attack would only have to exceed its Defense by 15%. If we add another 11%, then we've reached the 2.0 cap imposed on Centered Damage. You can see from these figures, why I am skeptical of the validity of the unknown presented by the author.

I think it would be great if we were able to work out a viable alternative unknown that would fit into this function. You were quite astute in pointing out that a logarimthic comparison of attack to defense could potentially result is some astronomical figures if one value were to exceed the other by a wide margin. There would have to be a cap imposed at some point, though where might be difficult to determine. It would take me some time to do some testing to try and create an algorithm for this problem. I just wish we had access to just one monster's actual statistics. We could almost immediately solve this if we had that information...

LIGHT BULB

I just remembered that we do have the ability to figure this out!! We now have PvP!! It's just a simple matter of two combatants entering Ballista with the intention of whacking away at each other and taking the data for this forumla. I think the best choices for this would be a level 50 PLD/WAR and a similiar level DRK/WAR. Have them alternate turning on different abilities which raise or lower defense or attack. Have them switch between ranges of equipment, including +VIT equipment, etc. I think would could nail down this down pretty well with the data that would be obtainable in this fashion. Now, it's just a matter of finding the right people to do it.

[Macht]

You know Mithrael just by goofing around I making my own "Equation" I ended up in similar results you posted in the first page which I saw after I did it. I ended going threw a Sigma function which it's end results came up with.

(m-p) sigma (n=(STR-VIT)) Æ’(n) = (k+1)(1/2(-p+(m-p)^2+m)-1)

I didn't do any in-game testing just goofed around writting the theory on paper and such. k being the Random factor m being attack and p being defense. m and p had more to put in but unfortunatly that seemed like going way overboard on just a playing around theory. :sweat:

Anyway point is if you even do a half*** job on the equation you come close to the same thing without even needing actual data. The actual data should make it easy for someone to create the formula you'd have more then half the variables then, the only matter of finding at that point is just the monsters stats which you can do in a pvp match now.

EDIT:

About the overboard bit, actually the real reason is I got lazy and used my calc to give me the results but it couldn't handle the number of variables to create the formula correctly so I plugged in "dummy" data for the basic variables like STR and VIT the -1 at the end of the equation is the result of that which seems to act as a transpose for the function.

[Aneyeisuponyou]

Ahhhh

I know Kung Fu

[Taoist]

Hmm...
"Show me."
<cue Matrix fight theme>

[Mithrael]

You know Mithrael just by goofing around I making my own "Equation" I ended up in similar results you posted in the first page

How were the results similar?

How were they dissimilar?

How did you come up with similar results without any in-game testing?

How can your results be similar without even taking into account equipped weapon, when equipped weapon obviously profoundly affects damage?

Does this answer any posters' questions in some way I'm not seeing?

Anyway point is if you even do a half*** job on the equation you come close to the same thing without even needing actual data.

Huh?

I don't see how you've made this point at all. What are you actually trying to say, and how does it relate to the potential and practical cap on attack?

[Gman]

Ok, stupid question time. I didn't read the initial web page that was posted where all these formulas came from, since I'm at work and it blocks them out, but please...

Where do the variables "24" and "3" come from? Why is the Formula 24 +3 [weapon power/9]??? Does it have something to do with the STR vs VIT comparison or is it something totally different?

Here's a mind bender for you Mithrael, I'm a RDM and I want to find out which of my swords will do more damage (mainly for farming, so factor in "too weak" mobs rather than IT.)

Fencing Degen vs Bastokan Sword:

Degen: DMG 22
Bastokan: DMG 20 +4 att

Using your formulas, and my STR of 32+5 (37 total), which sword is better? I would work it out myself, but I have no clue what any of that meant

I'm guessing my max attack will be better with the Degen, but my chance of attacking for higher damage consistently will be better with Bastokan Sword? Is that even close?

Thx, and good read btw.

[Mithrael]

The 24, 3, and 9 are just static numbers that are a part of the equation. That part of the equation is just a representation of the maximum amount of strength you can have and still be a benefit.

There is no way to really exactly determine which sword is better, since you don't know the mob's real defense. Delay will play a factor, your delay/60 is the actual amount of seconds your sword has for a delay. Then DMG/(delay/60) will give you a quick and dirty (and also not a real-world) number for damage per second. Attack+4 against a too weak mob probably isn't going to affect much since you should be close to rolling a 2 anyway. My guess (without looking it up since i'm feeling lazy...) is that the degen will have a better damage per second number, so it's probably better off for farming stuff that's 15 or more levels under your level. As things approach your level, the sword with the +4 may boost your damage.

For example, my Barbarian's Scythe has attack+25, but I don't really see it doing significantly more damage than my Darksteel Claymore while farming since I'm already at cap without the +25. I just use whichever I'm more in the mood to see on a given day, though the Greatsword weapons skills are generally more useful and reliable...

[bleegh]

so a quick question clarifying the first part of the equation.

When deciding between a weapon with +9 dmg, and a weapon with +8 dmg and +1str, it would probably be best going with the weapon with +8 dmg and +1 str, because in an exp situation, the +1str will bring you closer to your max base dmg, correct?

Also

As a mnk you're range of power is much smaller than a 2-hander. From attack/defense = 0 all the way up to your max of attack/defense = 2, your range about doubles. So I'd say pump up STR until your STR+ is equal to your base damage cap (the first part of the equation).

Problem is, we don't know what the monster's VIT is, so it's hard to tell how much str we need until we hit our base damage cap. In a sense, it's nice to know the information, but all we can really get from it is a general idea of which stat to boost at a given time. As far as concrete numbers, we're still in the dark, given that monster stats are unknown. But given the fact that base dmg can go negative, while rate cannot, I get the impression that STR becomes more important than Attack on exp monsters - especially so for weaker monks, such as humes, mithra, or taru.

[Mithrael]

When deciding between a weapon with +9 dmg, and a weapon with +8 dmg and +1str, it would probably be best going with the weapon with +8 dmg and +1 str, because in an exp situation, the +1str will bring you closer to your max base dmg, correct?

Just as a clarification, weapon DMG is not a part of base damage.

As a MNK, I'd say STR+1 is probably more useful than DMG:+1. I may or may not change that opinion for DRK, I'm not sure what I'd choose if two weapons had identical delay, one was DMG:76 STR+1 and the other was DMG:77. I'd have to try them out for a while and see which was better, it's a tough call...

But given the fact that base dmg can go negative, while rate cannot, I get the impression that STR becomes more important than Attack on exp monsters

I'd say this is a pretty good conclusion. However, if you have Kote and the HQ version of whichever Bonecrafted earrings for your level (attack+ evasion-), adding additional attack may bring your level up a bit more than STR. But it's really hard to say in the abstract. It'd change quite a bit if you were fighting crawlers vs. crab/beetle, orc/quadav vs. yagudo, etc. Fortunately, you tend to fight the same thing for several levels in a row, so it's not too hard to gear your equip for 2-3 levels at a time.

[Arkadya]

I'm a little confused so let me use an example to see if I have this right. For example, HtH skill of 160 using a 15 damage weapon.

3 + floor(160 / 10) + 15 = 34

Which means I have a max damage of:

24 + 3 x floor(34 / 9) = 31?

"A MNK's goal is good sustained damage, so STR is very important, but you cannot neglect Attack. Once you reach the point your STR+ is equal to the maximum base damage (from part 1), adding anymore STR is pointless. This is why the Kote are godly."

When you say this do you mean the modifier to your strength score through equipment, mithkabobs etc. Or do you mean your totalled strangth score? If it's the latter, wouldn't that render str gear pretty much useless? If it's the former, isn't that a lofty goal to achieve? Or is my math just completely messed up (odds of this.. extremely high!) Or am I using the wrong formula to determine how much strength I need?

Edit: I also havn't looked at my str score since like, level 30, maybe it isn't that hard to get.
Edit 2: Str score of 63 at level 54, it IS that hard to get, at least for a mithra

[bleegh]

he mentioned that it was your str - monster's vit.

if that's the case, there will never be a set number of str until we know exact stats on monsters.

also, even with the added str from food and whatnot, there's no guarantee you'll hit that cap.