Anyone Have Data on Block and Parry Percentage?

[Ultimatecalibur]

It seems that monsters have a hidden accuracy stat that in some way influences dodge, block and parry rates. This could explain how block rates seem to stay roughly the same as you level up and equip better shields even though GLA/PLD shield "Block Rate" stats increase from 35/10 to 181/181 & 295/92.

If Bulwark is actually modifying the "Block Rate" stat this may explain how a +60% increase seems to lead to a 100% block rate when used in overleveled content. Holy Shield's "Block Rate" under Bulwark would be 289.2 and Onion Shield's would be 472.

[WingsofWar]

snip

Don't think its necessarily hidden, otherwise Mob levels are just for aesthetics. Wouldn't be be pretty clear that if we have stats based on levels, so would mobs have similar stats as the game progresses? I thought this was the case in t5 Twintania actually has some high DEX/PARRY values that would make us want more accuracy to make up for it. Even some skills lower stats of mobs, like STR, DEX. Flash's blind effect would lower a mobs ACC would it not?

[Ultimatecalibur]

Don't think its necessarily hidden, otherwise Mob levels are just for aesthetics. Wouldn't be be pretty clear that if we have stats based on levels, so would mobs have similar stats as the game progresses? I thought this was the case in t5 Twintania actually has some high DEX/PARRY values that would make us want more accuracy to make up for it. Even some skills lower stats of mobs, like STR, DEX. Flash's blind effect would lower a mobs ACC would it not?

True, but I meant hidden in they way that we can not directly see them and do not know the exact formulas for how it influences things. We don't know how +108.2 and +177 bonuses actually influence block rates.

[Gamemako]

The most reliable source of data for Block/Parry RATE testing that I've uncovered so far comes from a Guy on Reddit called Kestiel. The thread linked has stats on the Onion and Allagan/Holy shields, but he's also posted some other numbers in the past which largely tie in with what was previously posted by Ninjiitstu and Ruckus.

I will have to do some testing to corroborate this, because I distrust his posting in general. He has repeatedly stated things like "dex follows clear 1% tiers of x", where x was 40-41, then 30-31, now 20-21 -- and the data is always in need of cleanup and can't be posted. The Bulwark testing makes me think he just made it up entirely. You would need to test for 29 straight hours to get the number of hits he has reported under Bulwark. Either way, I put no stock whatsoever in his posts.

[Maelwys]

I will have to do some testing to corroborate this, because I distrust his posting in general. He has repeatedly stated things like "dex follows clear 1% tiers of x", where x was 40-41, then 30-31, now 20-21 -- and the data is always in need of cleanup and can't be posted. The Bulwark testing makes me think he just made it up entirely. You would need to test for 29 straight hours to get the number of hits he has reported under Bulwark. Either way, I put no stock whatsoever in his posts.

If the past posts are to be believed there was at least three weeks worth of testing that went into it.

Frankly I'd like to see some of the raw data too... but there's a few mitigating (if circumstantial) factors that make me more inclined to take his posts at face value.

He's a moderator on the subreddit group in question (link - bottom right), the general tone of his past forum messages and More Detailed Reddit posts don't really suggest themselves to me as trollish. And he's at least been consistent in the claims about gathering data for testing.

I'm not sure where you're getting that "inconsistent Dex tier" thing from though.... All the past posts I've seen have been suggesting that there is a gain of ~1% chance when going from ~210 Dex to ~241 Dex - which would be 31 points of DEX. Indeed, I recall seeing him post that he "couldn't test lower" than 210 Dex without skewing the results by using inferior equipment. I'm willing to give him the benefit of the doubt for now.

I'd not put too much faith in his "Parry formula" being exact, but ~0.076-0.078% rate increase per point is a decent ballpark figure, and the more testing we have that ties everything together the better. (So by all means if you're up for more testing I'd love to see the results!!)

That said, I'm also not a great fan of the Str/Dex tiering idea - I'm pretty much convinced that the Red Numbers you see popping up in the game are only integers due to rounding... (More STR + More Dex + More Parry = always better!) but it'd be nice to prove this once and for all. I wonder how many attacks there are in the game that do a 100% predictable amount of damage... "1000 needles" anyone? Maybe the next time I get a free weekend I'll try and do a test or five myself...