Well, really, the only stat we're concerned with is the mob's level. As I don't think any stat but lvl, skill, and shield size were ever shown to affect block rate.
But even so, GC are only lvl 82~83, IIRC, and I strongly suspect that Ochain will have a capped block rate at base skill for anything below even match.
Ochain
|
Ochain
well, i didnt read any of this thread but ill throw in some input, our pld that has one can easily solo Dark Ixion outside abyssea w/o any trouble whatsoever. Absolute virtue didnt even hit him for more then 20 damage only magic hurt him.
As for inside abyssea, if you have almace and this then your basiclly not gonna die to anything besides maybe meteor spam. Ragnarok.Martel said: Hmmm. At the moment, the only parse sample of that size I have lying around would probly only be good for a block rate cap test. It would be nice to find out if Ochain's block cap is 95%, like alot of thigs in FFXI. Its 5k+ hits sample of EP murrain chigoes. No PDT, no buffs, 369 shield skill(base+merits). Originally I wanted to use it to figure out Ochain's DMG reduction on block. But I realized afterwards that since they hit for so little per hit it wouldn't be very accurate. Too many hits getting capped at 1 dmg. I took a twin sample with creed hands +2 on in an attempt to figure out the shield defense bonus increase, but it'd have the same accuracy issues as the other test. Hmmm, I also have a rather huge sample from a very painful Hydra solo attempt. It probly wouldn't help for a shield kill test since I was swapping in skill for casts, but it might be usable as a dmg- test. Or just as a Basic Ochain vs lvl 80 mob test. Well, if you think the data would still be useful/usable I can post or PM it. For future testing I think my main issue would be finding mobs of a static, known lvl that aren't nasty NMs of some kind. I could probably pull out some information from the chigoe tests, but the differing lvl ranges of the chigoes and low base damage would make it hard to test damage taken (as you've already pointed out) without a -very- large sample size. If you get a wind to dispel the enspells, you can probably use a Sky god for some long testing, since they have known lvls (could probably pop -> go afk with a curebot p2ing and curing lol). Would piss off someone that wanted to do that fight, but who cares! I don't remember if the Scars of abyssea NMs pop with lvl ranges or static lvls (should be easy to test), but if you had someone building time, you could do the same afk test with the Mandy NM for a lvl 80~82 mob. The rabbit NM in Vunk would probably be decent for lvl 85ish afk testing. Tiger NM in the range would be decent for lvl 90+. Though I'd think you would need 95+ to actually start seeing the shield skill curve for Ochain.. When I finish with lvl 90 trials for the wife and my empyreans, I had planned on doing the same shield block rate tests with lvl 90 Aegis, but I may just wait till 95/99, we'll see. edit: And yes, as Martel said: It's lvl and shield skill. I've seen no difference in things like Agi/Dex differences, though my sample sizes for those tests were too small to rule out a correlation (I'd say it's a safe assumption, but won't argue against it being possible). Ochain: O :D:D:D:D
Ok I was gimping my shield skill just for a fun to capping it with Ochain. It is beyond what I expected 10+ VT mandragoras hitting for 0 went from 282 to 330 in 15 min :D
[+]
Well, perhaps we can establish a value for the Ochain block rate cap with this data. I mean, EP mobs, I damn well should be capped. If there's a need I could retest in full shield skill gear and see if there's an increase.
Once we have a defined cap, we can jump straight up to a lvl 90 or 95 mob test. If 90's still capped, move up in mob lvl. If not Then we can start adding skill to observe the difference. I Anyway, data Grat's Atoreis!
You're making me want to do one, if I didn't LOATHE VNMs. I'll take a look at that data tomorrow Martell, TY. No, man. Thank you! Doesn't matter how much data I gather, means *** all if I can't analyze it.
Grats Atoreis. Ochain is crazy fun. For some reason they don't really let you do spacing here. For example I hit space a bazzillion times between spacing and here and yet look at it. You could however make a table or if you send me the actual pic of it I could for you so it's easier to digest. Cause honestly kinda looks like gibberish like that
Honestly, I've no idea how to make a table on these forums. Or on any forums really. I certainly would have liked it to look cleaner and more readable, but the one way I know how to do that doesn't work here.
It should still be readable by anyone familiar with KParser, but its certainly harder on the eyes. By pic, are you asking me to take a screen shot of the parser log? Not sure I really get it. I just copied all of that right outta the parser then pasted it. Sure or a SS of what you posted on BG. Also if you click the "reference page" link down below with other example BBcodes it shows you how to do all kinds of cool coding like making tables and more
I looked at the reference page before posting the data, but it made very little sense to me. I tried messing with some of the code, but couldn't get it to do what I needed.
As for a SS.. got an error when tried to upload one. Going to bed. Might try again tomorrow. Yeah I find it easier just to take a table from there and modify it. Could PM it to me too. Anyways good night
One thing though while I'm looking at this:
Where did the 800 damage hit come from lol? And also I'm not seeing any 0 damage hits.. Did none of the chigoes hit for 0, I don't remember if KParser doesn't log 0 damage hits, which would skew the data curve for block rate =/.  The block rate distributions for both sets of data were nearly identical (I set the bins to 100, even though the sample size was about 1000 hits each, but it shows most hits fell within the same bins for both gloves on and gloves off). I could only extract one Mode from the data, and that was the Shield Block mode, and with the estimated variance, I determined anything above 34 damage in each chart to be a "non block" with a confidence of 97%. That leaves 190/1015 hits unblocked, or a net Shield Block Rate of 81%, though as I said only with a confidence of 97%. This however can be imprecise if: 1. 0 damage hits were made but not logged. 2. The chigoes were not always in front of you for the hits. 3. The parser missed data or mislogged data. 4. Shield skill/lvl changes on mobs throughout the tests. Unfortunately, a "non block rate mode" is impossible to extract from the 190 samples, due to the large variance in damage, so I can't even begin to *** a damage reduction from this data set.
[+]
That wasn't an 803 dmg hit. That was 803 hits for 1 dmg.
With data in vertical columns the number on the left is the damage taken, the number on the right is how many times a hit for that damage occurred. Each of those samples is over 5,000 hits. I apologize for not making the data clearer. I think I'm just too used to KParser, and most ppl recognize KParser's data format, so I didn't think to explain it. And I really wish these forums wouldn't squish the data like that.. But I don't really know how to fix that short of posting it n BG, or something. There were no 0 dmg hits in the first test. I thought it was pretty interesting that creed hands +2 pushed the dmg down to 0. Kinda makes me wonder exactly how shield def bonus works... a percent based reduction shouldn't be able to reduce the whole number to 0 without being 100%, right? Unless it got it below 1 dmg decimal wise then floored it...? As for test precision. These were controlled tests. Mobs were maneuvered into position in front of me, all gear was locked, then the parse was started. The mobs were neatly clustered in front of me, so there's no chance that any of them were outside of my block radius. As I stated, 0 dmg hits were made, and logged, but only in the creed hands test. Again, no Buffs aside from regen atma were present. Also, the same pull of mobs was used for both tests. I was wearing other gear. I'd have to go find my screen shot to say exactly what. The main point was that I was careful to wear no PDT or shield skill gear.
However, this does present another possible reason for the minimum dmg taken dropping from 1 to 0 between the 2 tests. Creed hands +2 have quite a lot of def/vit. Perhaps that ticked my def over some critical value that allowed them to hit me 0. Or some combination of that and the shield defense bonus enhancement. Also, both tests were done with Ochain. Aegis was never in the picture. Ah yes I was reading the data backwards, give me a sec to replot and determine the new modes (which should up the block rate considerably and perhaps pull out both modes).
No, I wasn't wearing Creed body. It has damage taken- on it. I was trying to avoid any PDT/DT gear so that all the DMG reduction I was seeing would be from Ochain's Block dmg-.
With Failanx up and -pdt gear you could be a pretty tough guy.
With very few exceptions, if it's physical DMG Ochain laughs at it. I rarely bother with PDT recently, cept for when I see a ws coming. Hurray for bloodtanking in DD gear
Ragnarok.Martel said: With very few exceptions, if it's physical DMG Ochain laughs at it. I rarely bother with PDT recently, cept for when I see a ws coming. Hurray for bloodtanking in DD gear Allmost makes me want to do ochain and pld :P  New distributions, since I was reading the first set backwards (I knew better, but it's been a while since I used KP). Tests with gloves are on the right, tests without gloves on the left. I examined the distributions with and without crits to determine variance and mean for each of the distributions, without being effected by the crit hit pdif shift. Unfortunately as you can see, the normal distribution for non blocked hits has a stupidly high variance in comparison to it's mean, which overlaps heavily to the Mode of the blocked hits, making it impossible to determine an accurate block rate or damage reduction from this set of data. The one thing we can pull from the data however is the observation Martel pointed out about the Shield Def bonus from the gloves shifting the blocks over towards 0 damage, which I would assume is a rounding/flooring difference... Just seems unlikely that you would be "right on the edge" of 0 damage blocks and only get pushed over with a ~5% PDT boost on shield block (or ~5% defense boost if shield blocks just change your pdif calc to some massive defense rating). I'll need to mess around with the math and other damage equations to see what would cause that.. Phoenix.Neosutra said: Unfortunately as you can see, the normal distribution for non blocked hits has a stupidly high variance in comparison to it's mean, which overlaps heavily to the Mode of the blocked hits, making it impossible to determine an accurate block rate or damage reduction from this set of data. Yep, that's what I was worried about. Still, it was a pretty fast test for its size, and it was hella fun. Unfortunately, anything that likely to give us more useful data, is going to be much slower going. Phoenix.Neosutra said: The one thing we can pull from the data however is the observation Martel pointed out about the Shield Def bonus from the gloves shifting the blocks over towards 0 damage, which I would assume is a rounding/flooring difference... Just seems unlikely that you would be "right on the edge" of 0 damage blocks and only get pushed over with a ~5% PDT boost on shield block (or ~5% defense boost if shield blocks just change your pdif calc to some massive defense rating). I'd really like to see some base tests on Shield def. bonus. Did anyone ever do any after it was released? Then compared to older tests with the same tsize/def shield? Huh.. I'd always thought of blocks as a form of PDT in a different term from normal PDT. But thinking of it as a change to pDIF is pretty interesting. I wouldn't call it evidence, but there is something that makes me think that could be right. Damage absorbtion, via creed set, or anything else, ignores pretty much every form of dmg reduction. PDT, MDT, MDB, SS, phalanx, whatever. You'll still get cured for the full amount, assuming you're missing enough hp. As far as i can tell, the only things that affect it are the base stats for determining dmg. Def, vit etc. Except shield blocks. When a blocked attack is absorbed, you get the reduced amount. Again, I'd hardly call it hard evidence, but it is an indication that blocks work differently than PDT, and it suggests that they'd be in the same category as def/pdif. Kinda similar to guard I suppose? Hmmm, someone do an absorb test on guarded hits, lol. Now I have no freaking clue how you'd go about trying to prove this. But I thought it was an interesting idea. Oddly enough, the "With Gloves" charts look like you just shifted all the damage taken over by 2 points.. That's so odd, it doesn't follow a standard PDT equation.
It occurred to me that Shield Def Bonus may act like a form of Phalanx, in that it cuts off a static amount of damage per level. I'll see if there is a good way to test this later.
Phoenix.Neosutra said: It occurred to me that Shield Def Bonus may act like a form of Phalanx, in that it cuts off a static amount of damage per level. I'll see if there is a good way to test this later. Well, I'm working on a new data sample. I decided to use Hydra. He's only lvl 80, but his lvl is static, and he's always up. He also hits decently hard.
What would be a reasonable minimum sample size? This is a lot slower than getting hit by chigoes. A 5k sample size is a long long ways away. Also, would having a sample of purely unblocked(turn away from mob) hits be helpful for comparison? It'd give us a base dmg that we know doesn't have any blocked hits mixed in. Should be useful for finding the block dmg reduction as well, I'd think. That, and after a 1k+ hits thus far, I haven't seen a single hit that I was certain wasn't blocked. <,<; How large would an unblocked sample need to be, assuming it'd actually help? It's gonna be pretty hard to gather so I hope it doesn't have to be huge. Oh, right. About posting the data. If I can't find a way to maintain the spacing on these forums I might post it on BG then link it. It's roundabout and annoying, but I'd like my test data to be easy to read. Especially if we actually get some good results this time. Hydra base block sample is at 2560 hits
Unblocked sample is only at 257 hits. It's a *** to stay alive with my back to hydra. It's amazing how fast you can run outta mp outside abyssea. And I'm losing not only Ochain dmg-, but the MP conversion and the spell interruption. Gonna put the unblocked sample on hold till I know how big it needs to be. Still, even with a small sample, the difference between the dmg ranges is quite noticeable. There's actually quite a large gap between the dmg ranges of the two. Either that's because of the small sample on the unblocked test, or the blocked test had no unblocked hits at all... In any case, I', going to hold off on more for those samples till I know how much I need. I'm going to start on a Creed hands test next. Perhaps it'll shed a bit more light on shield def bonus mechanics. |
All FFXIV and FFXI content and images © 2002-2024 SQUARE ENIX CO., LTD.
FINAL FANTASY is a registered trademark of Square Enix Co., Ltd.