250 dex cap and parry - Test results!

[Gidget]

nice ending really, but i still want to know, you know!

so how to figure what made the block and parry in caelis test go up that significantly above 250 dex? :touch:

 

[Zoia]

nice ending really, but i still want to know, you know!

so how to figure what made the block and parry in caelis test go up that significantly above 250 dex? :touch:

Aye, that's what i'm wondering about too.
So far, my theory is that there's some sort of a soft cap at 250 dex, but after that it's still somehow affected by your spec. Though who knows? I certainly don't.
If only we had access to a test server, so i wouldn't have to buy a bunch of RA stones every time!

 

[SethNaket]

Well the reson evade is showing smaller difference is probably just because it's not pure dex but (dex+qui)/2 that determines it. When you raise dex from 150ish to 250ish, evade only "gets" a 50 increase. Then when you put on a spec buff dex only "gets" a 45 increase from qui since dex is already at the assumed cap.

That said, there's no inherent problem in looking at evade just because it's lower in %, the results are actually more accurate than if you look at something close to 50%. You can see this directly in the formula for standard deviation for a proportion: sd = sqrt( p*(1-p)/n ). p*(1-p) becomes smaller and smaller the farther you go from 50%. For instance if you test 1200 swings, the standard deviation for a 5% evade rate is only 0.44% while for a 50% block rate it's 1.44%.

Btw, 42% parry with only 17+16 parry? That's pretty high, isn't the cap 50% on parry?

 

[Zoia]

Well the reson evade is showing smaller difference is probably just because it's not pure dex but (dex+qui)/2 that determines it. When you raise dex from 150ish to 250ish, evade only "gets" a 50 increase. Then when you put on a spec buff dex only "gets" a 45 increase from qui since dex is already at the assumed cap.

That said, there's no inherent problem in looking at evade just because it's lower in %, the results are actually more accurate than if you look at something close to 50%. You can see this directly in the formula for standard deviation for a proportion: sd = sqrt( p*(1-p)/n ). p*(1-p) becomes smaller and smaller the farther you go from 50%. For instance if you test 1200 swings, the standard deviation for a 5% evade rate is only 0.44% while for a 50% block rate it's 1.44%.

Btw, 42% parry with only 17+16 parry? That's pretty high, isn't the cap 50% on parry?

Heh, yea, it is 50%. That's one of the things that made me wonder a bit too. 17+16=33. that's 33/2=16,5% parry. Add 5% base parry to that and MoParry1 and you have 23,5%.
But it doesn't seem too unreal that 150 dex adds 7.2%(=30,7).

What's strange is that the other 99 dex adds 9.8% for a total of 40,5%. However, i've read that quickness only increases your swing speed from 60+. It could be the same thing for parry and dex. Anything below 60 dex doesn't help at all.
That would make the first 90 dex increase it by 7.2% then the next 99 dex increase it by 9.8%. With some variance, that seems ok, almost.

Then you have the last 94 dex increasing it further by 2%. Perhaps even with a 1200 swing test you have 2%+ variance?
At least that could explain the 2% increase in the last 345 dex test and prove you don't get much of an increase with over 250 dex.

 

[aika]

I think that understanding Quantum Entanglement is easier than those formulas tbh!

 

[Tuthmes]

At some point i'm gonne wanne say; "hate to say i told you so." but not just yet.

Ill be off to beeing stubborn somewhere else, but closely following this thread though!

 

[MegaMaejter]

Zoia, can't you run a test on ur valk ? Shaman hitting valk that is..even though by logic ( after reading the grab bag quotes ) parry, block and evade should work under similar code, you never know, and a simple test would quite easily show either way.

Seems like whenever you answer one question/theory two knew questions rise in this matter