I'm a lvl 70 bishop with one tattoo slot left (have +5 WIT) and I can't decide whether the ~300 HP and ~150 CP (plus slightly higher stun resist) or +5 run speed (plus slightly higher evasion (useless)) would be more beneficial to me.
Thing is I'm usually dead meat if anyone attacks me, and usually it's a nuker, archer, or dagger who will do 1500-4000 damage -- in most cases, an extra 450 HP/CP won't help me live any longer (or if it does, only for 1 more second).
In frequent PVP cases, my party will wind up chasing/ running after/towards our target, and we usually get spread out really far and I'm way in the back because of my low run speed. Thus, when the people in the front start taking damage, I'm not close enough to heal them and they're dead by the time I get there. In some cases if I chase after the front-runners and wind up in the middle of our enemies and then drop in an instant. I leveled my own wind strider, which I use to run long distances, but I can't use it in PVP effectively. +4 DEX would give me like +5 run speed, which I don't think would make any difference (and I can't afford DC robes for quite some time).
Which, allbeit little, difference do you think would be better?
Well just from the way you worded it, seems like you would rather go with the dex increase. Faster run speed is always good, plus you get higher shield block, evasion(you might get lucky) accuracy(yeah useless for mages) and crit rate(meh). If you die really fast with our without the con tat, then I think you should go with dex.
At best, you can then outrun a -dex armor set human fighter? Unless you af over long distances a lot, take the extra cp/hp and pray you end up with 100 left now and then. Oh, stun resistance too.
Mystics have low physical stats, so the small increase won't give you enough of an edge. I went +dex -str on my EE at an early level to give me an accuracy bonus (which was actually fairly helpful). +con is much more likely to help you survive than +dex (especially since you can't evade magic or dagger/archer specials).
You can block some magic, archer hits and dagger specials. And +Dex also gives you shield block bonus. But without shield buffs, the bonus might be too low to matter.
Yes, I tested this on the l2 calculator actually, since it now has stats for shield block. For a level 68 EE, with bless shield level 6, +4 dex gives.......+1% shield block!
1.5%, actually (if we disregard that the shield stuff there is pure speculation). It still means 10% more arrow attacks blocked (if you face your opponent), compared to just extra 3% of arrows blocked by tattooing an unbuffed Bishop.
1.5%, actually (if we disregard that the shield stuff there is pure speculation). It still means 10% more arrow attacks blocked (if you face your opponent), compared to just extra 3% of arrows blocked by tattooing an unbuffed Bishop.
The shield block rate given on L2Calc isn't pure speculation. I just haven't bothered to update l2p.bravehost.com yet. All of the testing I and others have done thus far have shown the shield block rate to be accurate. The shield P.Def. in v0.31 however is incorrect. It will be updated relatively soon along with a number of new features that I have tested personally with a good number of wasted hours!
Hey Nyo. :) Any chance for a crit damage section? ^^
</hijack>
I have +DEX and am loving the runspeed, but that's for PvE. I'd say at your level the dyes are cheap enough to get them both (the one you think you will like less first) and experiment a bit.
Speaking about PvE: I've made a lot of death runs on Aden siege on PTS (through the hole in the wall with the archer guards still on top). And I found I had much more chances to survive if I had Zerk on :p
Hey Nyo. :) Any chance for a crit damage section? ^^
Hum, I guess I shouldn't give away too much but the answer is absolutely. :P
Btw, L2Calc does NOT show defense rate against arrows as I'm sure many of you realize. Your block rate against arrows is much higher than against any other kind of attack. I have one test done of a human prophet with no dex mod -
block rate versus sword - 17/100 (17%)
block rate versus bow - 60/101 (59.4%)
I still need to test arrow block rate more but it's likely either +40% chance or *3 chance.
It could be 15% or 20% as well, 100 is too few samples to tell. If you need your estimaton to have better than 1% of absolute precision, you need thousands of samples here.
No kidding. Why do you think I said the arrow block rate "needs more testing"? I only gave the example that applied to the discussion being involved. I've been working with statistics for quite a while now and am very familiar with margin of error.
First of all, I have a formula that is linear because it neatly applies all the stats of each factor being involved. Then, I set out to prove it with various statistical applications - different characters, levels, buffs, etc. When you already have a formula and you have a variety of different applications that are able to display a trend, you do not need thousands of applications. For example, here are some of the tests I did last week:
w/BS3 on TK - 73/100
w/BS3 doom on TK - 89/100
w/nothing on TK from behind - 48/100
w/nothing on TK from front - 52/100
prophet - 17/100
prophet w/BS3 - 54/200=27%
prophet unbuffed vs arrow - 60/101
prophet w/BS3 vs arrow - 95/100
What I do is if it fits within the margin of error, the test passes. If it doesn't, I make the test longer in order to make it as accurate as I can stand to. For example "prophet w/bs3" started out at something like 22/100, which was pretty far off from the predicted 30% so I tested some more and found the % rose by the 200th try.
No kidding. Why do you think I said the arrow block rate "needs more testing"? I only gave the example that applied to the discussion being involved. I've been working with statistics for quite a while now and am very familiar with margin of error.
Well, if you show confidence intervals instead of point estimations of the parameters and end results you provide, everyone will see that much more testing is needed.
What I do is if it fits within the margin of error, the test passes. If it doesn't, I make the test longer in order to make it as accurate as I can stand to.
Which means that your sampling is not actually random. It's a very dangerous approach.
Well, let's give you the benefit of the doubt here. We can talk about it all we want. Can you actually prove the theory to be incorrect? With real data as opposed to theory.
With your data, I can show that it's unproven. Unfortunately, to prove if it's correct (or incorrect) within 1% margin of error, I would need to break the EULA. More than 6000 tries just to check that Human Mystic's base block rate is 20%, but neither 19% nor 21%, are not that I'd like to do manually.
You wouldn't really need to do 6000 tries if only you understood a simpler way to go about it. Take the critical point where shield block rate goes from 99%->100% and begin the testing around there. At just barely 100%, even 1 failure to shield block would prove the system wrong. Or at 99%, if after several hundred blocks there was still no failed block, it would be a representation of something being wrong.
With regards to this discussion, you have shown that you are not willing to do more than nay-say because it doesn't necessarily agree with your idea of how it should work. Maybe you do think that way, maybe you don't. You are not the first person to react this way and you certainly won't be the last since l2calc is now delving into things that can't simply be displayed in the status window. I was hoping you would be able to provide a more intuitive response than you did. It isn't just the fact that we've been going back and forth, but also that you missed the most obvious hole I left in my argument because you didn't even bother to look at the data for what it was.
I left holes because I am not interested in saying "I am right and you are wrong", but rather because I desire to see l2calc become as accurate as possible. As I said, maybe you were trying to be helpful or maybe you just don't like the results it gives. It isn't clear one way or the other. But to have you say "no it isn't accurate!" and then make no effort whatsoever to assist makes discussing this with you a waste of time. Yes, there is a very real possibility it could be wrong, but all indications (and no, what I posted here isn't the only data I have) show that this is the most likely formula for shield block rate.
Look, we were speaking there about possible benefits of +4 Dex for Bishop. Which, frankly, has nothing to do with how good your formula works at 99%-100% Shiled Block (as even if you specially fitted it for that range, which indeed is not that hard, it still tells nothing about 20%-50% range).
And if you want some help with proving or disproving your formulas, why don't you create a separate thread for it, and take some more constructive approach there?
You called into question the accuracy, and I responded in kind. There is nothing more to it than that. The question asked at the beginning of the thread was already answered. But yes, I will leave well enough alone now that I realize what you intended.