Why not just make everyone a DPS?

[Phoenicia]

There is so much wrong with this. Crit has diminishing returns and isn't the most optimal off stat for all jobs.

I don't disagree with the general spirit of your post but do be careful with the facts you lay down.

There is no actual diminishing return as in each point of crit gives exactly the same % of crit chance, this doesn't change no matter how high you go in crit. The stat value doesn't change and each point is as good as the previous one. Crit is as linear as any other stat in this game.

The "diminishing return" people talk about is due to forced crits from things like Bootshine/Life Surge and crit buffs like Internal Release and Abandon, etc. This affects crit stat weight, not crit's % chance per point.

[Reynhart]

There is no actual diminishing return as in each point of crit gives exactly the same % of crit chance, this doesn't change no matter how high you go in crit.

That's exatcly why it's considered a "diminishing return".
If you need 100 for 5%, going from 100 to 200 actually doubles your ctitical hit rate while going from 200 to 300 only increase you chance by 50%.
It's also why Skillspeed/Spellspeed have an increasing returns since 0.1s weights "more" when you have 2.0 GCD than when you have 2.5.

[Phoenicia]

That's exatcly why it's considered a "diminishing return".
If you need 100 for 5%, going from 100 to 200 actually doubles your ctitical hit rate while going from 200 to 300 only increase you chance by 50%.
It's also why Skillspeed/Spellspeed have an increasing returns since 0.1s weights "more" when you have 2.0 GCD than when you have 2.5.

Yeah, that doesn't change the linear nature of how stats increase per point. It just means the increase is linear but the benefit is "diminishing".

Every stat increase increases damage by a less % of the point before. (When you have 100, increasing it by 1 is 1%, but when you have 200, increasing it by 1 is only 0.5%). This is true for all stats with the exception of Skill/Spell Speed.

The "diminishing returns" I was thinking of was that stats increase with a formula with a denominator. For example in League of Legends, armor decreases the damage at a rate of "armor / (100+armor)". Each point reduces damage less than the previous. You get a non-linear (it's a curve opening upwards) graph that closes to, but never reaching, 0. Though the benefit is linear (each point increases eHP by 1%), the effect of each point is "diminishing". At 1 you reduce damage by 1%, at 2 you reduce damage by 1.96% (each point averaged 0.98) and at 3 you reduce damage by 2.91% (average 0.97) and so on. At 100 armor you reduce damage by 50% (each point averaged 0.5%) increasing eHP by 100%. This made it so it is equally important to build HP as well as armor/magic resist to efficiently increase survivability as you had only 6 slots for items.

With disregard of semantics. My whole point is, in here, you end up with a linear approach of "More x stat is always better" and you have no point where you should "stop" building that stat in favor of another stat for added efficiency. Or if that point exists, it is far beyond our gear's reach (Current and future patches').

[Reynhart]

Every stat increase increases damage by a less % of the point before.

Yet, but only critical rate is an actual percentage.

Even in case of "diminishing returns" more stats is still better. It's just that, at low values, Crit will probably have more impact than DET and at higher value, the gap will only increases more and more and favoring DET.

It takes roughly 430 additional crit in gear to increase that rate by 0.1%.

Really ?!
Suddendly, it sounds almost as useless as focusing on parry...

[Phoenicia]

Yet, but only critical rate is an actual percentage.

Even in case of "diminishing returns" more stats is still better. It's just that, at low values, Crit will probably have more impact than DET and at higher value, the gap will only increases more and more and favoring DET.

I think we're really trying to explain the same thing. My whole point is we do not reach a point in this game where it is more beneficial to stop building a stat in favor of another. Stats are mainly linear. For example (hypothetical numbers that hold no truth): 1 WD = 8STR, 1STR = 15 Crit = 17 Det = 22 SS. No matter how high or low we go this stat weight is fixed as the value of each point of a stat is always the same compared to the value of another stat.

Take the LoL example from my previous post. Assume you have 1,700 HP and 200 armor (3 item slots) ending up with 5,100 eHP. At that point it is better and cheaper to build 600 HP (1 item slot instead of 3) than building another 100 armor. Since the new HP will be tripled in eHP because of the old armor, You end up with 6,900 eHP by building straight HP instead of 6,800 by building 100 armor.

Mathematically such choices can happen in FFXIV, but we cannot reach those points with the current or future itemization.

Yeah I was mad about crit too till I saw the new damage boost tied in now.

I think you misread the data. Looking at the document you linked:

After 354 (base) every 1 point of crit is 0.0232558% chance, or 43 points per 1% chance. Stat is linear and every additional 43 points will add 1% crit chance. At 430 additional points (784 total) of crit you have 10% additional chance (14.95% total). Your multiplier will be x1.5485.

The damage scales at exactly the same rate. Every 43 points adds 1% damage (1.01 multiplier). At 430 additional points (784 total) of crit Your multiplier will be at x1.5485.

This is more in line with the parry rate tests done on another thread where we reached to the formula in http://forum.square-enix.com/ffxiv/t...269"]this post. Quoting it with fixes:

Also enemy crit chance should be factored in for an absolute per-enemy chance. Making the function more like:
(((Parry-354)/(858*4))+0.05)*(100-dodge%)*(100-Shield block rate)*(100-enemy crit%)

1/(858*4) = 0.0029137. So the functions are close.