What's Better?

[Nexxus]

i'm interested also about dumping merit switch STR for AGI (as main job nin (95% time) and using Kannagi).
Also, now i got my last AFv1+1 item (plaited...) What the best to get first: AFv1 body (mostly for stun spike) or feets (dusk/dawn flee & AGI+6 for Hi)?

[Foldypaws]

Er, while that's technically true, it's not relevant. You're interpreting it as having some sort of diminishing returns.

Crit damage enhancements are cumulative, +5% is +5% regardless of how much you had already (unless you're hitting the cap).

ARG. NOT THIS AGAIN.

You can't consider gear in a vaccuum. It absolutely *does* matter how much you already had. You have to consider the *marginal* damage increase.

+5% crit damage is only worth +5% damage to crits if you had 0% crit damage before.

If you can add 4% damage to your crits by adding attack, that attack is *better* than 5% damage to crit if you have more than 24% crit damage already, and *worse* if you have less.

[Yugl]

My post *did* include non-body changes. Or did you miss where I was talking about hands and feet and neck etc?

The agi reduction to TP gain is moot if you're using gnarled horn anyway. TP gain is the only thing I can't account for because A) I'm not doing the math on overflow probability for 10 rounds->100TP + a bunch of DA and TA (and I doubt they did either) and B) as far as I know, *no one* knows what the difference between say, a 110 TP evisercation and a 100 TP evis is.

The last two parts I direct towards you. It's just easier to lump the first with them. Anyways, what it sounds like is you think the sTP and AGI are useless. It might be good to hear from someone else more versed in this department, or even better, the people who suggested Loki's is better.

I mean, I personally don't care what the WAR and NIN who are just there to proc red or the mnk who's just there to proc BLU does, but I don't tell them not to try to improve their gear

Joke + showing I don't have a stake in the outcome (i.e. not biased).

I mean, it's not like THF is SAM or anything, they can actually DD.

DRK works too you know. In fact, I recommend you use it since DRGs oppose DRKs.

Edit: Niiro, I actually think it is diminishing returns. If it's based on the entire damage and you do 100 DMG, going from 100 to 101 is greater than 199 to 200 if bonus is 1%.

[Foldypaws]

The last two parts I direct towards you. It's just easier to lump the first with them. Anyways, what it sounds like is you think the sTP and AGI are useless. It might be good to hear from someone else more versed in this department, or even better, the people who suggested Loki's is better.

I don't think it's useless, I'm aware I'm undervaluing Loki's by ignoring it. I just don't think anyone has *accurate* numbers for how much it's worth.

Edit: The sTP I mean. I do think the extra AGI is useless except as EVA.

Double Edit: I thought DRG opposed SAM, isn't PLD the job that opposes DRK? (Demon/Dragon Ecosystem vs Undead/Arcana ecosystem)

[Yugl]

Ok, that's fair.

[Foldypaws]

Actually, lemmie take a stab at this. Mandau/Kila+2 (I believe the OPs combo) is 4.5 TP a hand before sTP. 4.7 w/ Rajas only, 5 with Rajas+Loki's.

11 attack rounds vs 10 attack rounds. My last parse on THF put me at something like 60% melee to 40% WS damage, so absolute best case scenario for Loki's puts it at something like 4% extra damage from sTP.

Assuming a 77% crit rate (GH + RR ), the 3.4% extra crit damage from loki's is 2.6% additional damage. The attack is roughly a 5% increase in total damage (all melee + TP). I've got a loki's so at some point I'll parse rounds/WS with it and without (given that I don't have 26% with loki's right now, that's actually a better case scenario for loki's than otherwise, as there's less chance of an extra attack round sneaking in before I can WS.)

If my (or your) swings/WS ratio goes down by at least 6%, loki's will be better. (I don't have a mandau, but I have a triplus which is the same delay).

[Niiro]

You can't consider gear in a vaccuum. It absolutely *does* matter how much you already had. You have to consider the *marginal* damage increase.

+5% crit damage is only worth +5% damage to crits if you had 0% crit damage before.

If you can add 4% damage to your crits by adding attack, that attack is *better* than 5% damage to crit if you have more than 24% crit damage already, and *worse* if you have less.

That's true but you can only quantify it by knowing what you're trading the 5% extra damage for and how it will effect your crits.

Saying 38->43 is a 3.6% increase is still meaningless unless you have an alternative value to compare it to.

The gains from increased critical hit damage are completely linear, they're based off of your base crit damage not off of how much you had previously.

Edit: Niiro, I actually think it is diminishing returns. If it's based on the entire damage and you do 100 DMG, going from 100 to 101 is greater than 199 to 200 if bonus is 1%.

If your base crit damage is 100 you're going to get an extra 1 damage for every 1% of increased crit damage whether you start with +0% or +38%.

That's a linear increase, not a diminishing one.

[Foldypaws]

If your base crit damage is 100 you're going to get an extra 1 damage for every 1% of increased crit damage whether you start with +0% or +38%.

But the extra damage added to crits from extra attack isn't based on your "base crit damage", it's based on your *actual* crit damage.

If you increase your pDif by 4%, that may be worth more or less than +5% crit damage, depending on how much +crit damage you already have.

Which is the entire point.

If you claim "5% crit damage is always worth more than 4% pDIF, because 5 is bigger than 4" you're flat out wrong.

a 5% marginal damage increase is *always* worth more than 4% pDIF though, because 4% pDIF is a 4% marginal damage increase.

Likewise, putting loki's kaftan on a naked thf (w/ RR) is always worth a 3.6% pDIF increase.

[zagex]

You guys are discussing 2 different things. 1) the definition of diminishing returns. 2) how to compare what's better. One person is correct in stating what the definition of diminishing returns is. The other is correct in how to compare what's a better stat to put on.

For the record...
Static return
25% double attack = 1.25 x original number of attacks
50% double attack = 1.50 x original number of attacks
This is static because, compared to the base value with 0% double attack, the increase in attacks is the same % as the increase in double attack stat. 25% DA gives you exactly 25% more attacks.


Diminishing returns
25 MDB = 20% damage reduction from spells
50 MDB = 33.3% damage reduction from spells
This is diminishing because, compared to the base value with 0 MDB, the decrease in damage reduction is less than the amount of MDB you add. 25 MDB gives you less than 25% damage reduction.


Increasing returns
25% haste = 1.33x original number of attacks
50% haste = 2x original number of attacks
This is increasing because, compared to the base value with 0% haste, the increase in number of attacks is more than the increase in haste stat. 25% haste gives you more than 25% more attacks.

[Niiro]

You guys are discussing 2 different things. 1) the definition of diminishing returns. 2) how to compare what's better. One person is correct in stating what the definition of diminishing returns is. The other is correct in how to compare what's a better stat to put on.

This is our problem I think.

I'm just saying you can't quantify an increase crit damage amount VS the effect an ATK increase would have unless you know beforehand how much that attack will effect your pDif, which varies depending on what you're targeting.

[Yugl]

Last I checked, most traditional posters calculated DA as diminishing returns. Hence, why they placed 1.DAnew/1.DAold when looking at the output. It changes when you include other shit I believe, but it should still imply diminished returns.

Edit: Ya Zage, you're looking at it differently. When you compare DA, you compare using what you had beforehand. If you're comparing a 50% DA piece of gear with a 25% DA piece of gear for the same slot and have no DA, you're correct. However, equipment comparison is often X has R% DA and I already have Y% DA. You then calculate based on that because you cannot DA your DA.

[zagex]

I just got annoyed that people continually misstate what diminishing returns are.

[Foldypaws]

This is our problem I think.

I'm just saying you can't quantify an increase crit damage amount VS the effect an ATK increase would have unless you know beforehand how much that attack will effect your pDif, which varies depending on what you're targeting.

The higher the enemies level, the bigger difference attack is going to make due to level correction from pDIF.

Worst case scenario for the attack build (level 90 target, 450 defense), you gain 5% pDIF (not .05 pDIF, but 5%.) As you go up to a level 100 target (500 def) like the IT mobs in the new zones, the extra attack is worth up to 8% pDIF. (at which point it beats even best case for loki's with TP included)

[Mojo]

For the record...
There's no real point in analyzing Double Attack the way you did because what truly matters is your relative attack rate increase. Thus Double Attack has diminishing returns also. For instance, let's say you start with 10% Double Attack and equip a Brutal Earring. You go from 1.10 swings per round to 1.15 swings per round, which is only a 4.55% increase in DoT compared to adding Brutal Earring with an initial 0% Double Attack rate (which would yield a 5% increase in DoT.) Nothing really has static returns unless you're looking at it in an irrelevant way.

[Foldypaws]

I just got annoyed that people continually misstate what diminishing returns are.

Personally I get annoyed when people talk about returns at all, when what's important is marginal return.

Mojo: You're right that nothing in FFXI has static marginal returns, as nothing in FFXI is an exponential function. A bunch of stuff *does* have static return though, just not static *marginal* return.

(To head off people claiming otherwise, haste isn't exponential. It's a negative power law. An exponential wouldn't go to infinity at 100. Also, an exponential would have static marginal returns, haste has increasing marginal returns, just like Damage taken down.)

[Niiro]

Last I checked, most traditional posters calculated DA as diminishing returns. Hence, why they placed 1.DAnew/1.DAold when looking at the output. It changes when you include other shit I believe, but it should still imply diminished returns.

DA is a static (linear) return on paper it's only considered a diminishing return in practice relative to other stats like haste because you can only ever DA once per attack, compared to something like haste which compounds upon itself.

[zagex]

Last I checked, most traditional posters calculated DA as diminishing returns. Hence, why they placed 1.DAnew/1.DAold when looking at the output. It changes when you include other shit I believe, but it should still imply diminished returns.

I'm not arguing that the relative change is less when you have a high DA compared to a low DA. But DA, by no means, fits under the definition of diminishing returns. If you only compared new/old, there's nothing in this game that gives static returns. However, a lot of stats fit under the definition of static returns. I guess it's nitpicking a definition, but for some reason it's really annoying today.

[Foldypaws]

DA is a static (linear) return on paper it's only considered a diminishing return in practice relative to other stats like haste because you can only ever DA once per attack, compared to something like haste which compounds upon itself.

If you could DA on a DA, DA would have increasing returns, both regular and marginal, not static for either. (Number of attacks/round would approach infinity as DA-> 100) (subject to 8 hit limit, but it'd still be increasing.)

[Yugl]

I'm not arguing that the relative change is less when you have a high DA compared to a low DA. But DA, by no means, fits under the definition of diminishing returns. If you only compared new/old, there's nothing in this game that gives static returns. However, a lot of stats fit under the definition of static returns. I guess it's nitpicking a definition, but for some reason it's really annoying today.

Well w/e the definition you want to implement, we're using it for comparative purposes. Also, Foldy is right about the DA on DA. If you had % DA and the ability to DA on a DA? Oh shit.

[Kaegos]

Also, Foldy is right about the DA on DA. If you had % DA and the ability to DA on a DA? Oh shit.

Yo dawg.