Poetic (Ironworks) Gear

[Donjo]

that's not really true (though it may be arguing meaningless difference)

in your example, you have 10% chance to block (90% hits resulting unmitigated) then 5% chance to parry the remaining hits, which results in 4.5% effective parry rate. While it doesn't matter for low numbers, it can result in a noticeable difference later on (30% block and 35% parry results in only 24.5% effective parry rate, 54% mitigation on hits when you could expect 65%). That's the main reason why a parry build is more effective on a WAR than a PLD considering raw numbers.

Now does it really matters ? Well, not that much. After all, you're gaining more from blocking than you're losing on the alternative (parrying) so overall your mitigation improves, and is better than the one of a WAR.

general consensus is "Don't stack it but don't refuse it either".

Parry > skill speed

I would personally argue that it's more effective to stack Parry on a PLD than a WAR. We can ask one of two main questions when it comes to Parry, however.

1. How much benefit will I gain per individual point of Parry? WARs get the full 0.075%(or so) Parry rate per point. PLDs get somewhat less depending on the block rate of their shield, as you've explained. So, WARs do win here.

However, this doesn't necessarily mean that Parry is less effective on a Paladin than a Warrior, because there's another way to look at it.

2. If I stack parry, how often will I mitigate a physical attack? Let's pretend we have 600 Parry. This conveys a 31-32% Parry rate. A WAR with 600 Parry will mitigate physical attacks exactly that often. A PLD, however, has their Shield.

I'm going to lump Block and Parry into the term "mitigate a physical attack" because they both literally do the same thing. It's been well theorized that Block is checked before Parry. Thus, a PLD's physical mitigation rate is their Block Rate plus the adjusted Parry Rate of attacks which are not blocked. Their mitigation rate, to put it in simpler terms, is the chance that at least one of them will proc, provided that both cannot occur at the same time.

This means that, for example, a PLD with a block and parry both around 32% will have a physical mitigation rate of about 42-43%, which is greater than a WAR can ever reach. That was with the logic of a binomial distribution in which there are two trials that each have a 32% chance of succeeding. The alternate calculation is to take the 32% block rate and then 32% of the 68% of hits that aren't blocked for a ultimately higher number of 53.7%, but I can't say for certain which probabilities are actually functioning with respect to the in-game mechanics. I can certainly hope it's the second

I'll try calculating the other numbers with respect to other i110 shield types(using the second calculation type. Binomial distributions don't work when trial 1 and trial 2 have different odds.)

Noct Hoplon: 22% block rate + 32% of 78% = 46.96% total mitigation rate, with Block being significantly more powerful than Parry
True Ice Shield: 41% block rate + 32% of 59% = 59.88% total mitigation rate, with Block being less powerful than Parry(less overall damage reduction)

So, no matter what shield one is rocking or what math is correct, the fact of the matter is that the PLD who stacks Parry will mitigate attacks more often than the WAR who stacks Parry. In most cases, a PLD would mitigate attacks significantly more often. In my eyes, that means that it's more effective to stack Parry on PLD than WAR. The other caveat though is that the presence of the Shield makes it less necessary for the PLD to stack Parry in order to attain a reasonable mitigation rate. Thus, having less Parry won't hurt a PLD as much as it can hurt a WAR.