Main Menu

Beastlord offensive AA skill comparison

Started by Coprolith, January 06, 2004, 09:13:37 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Tastian

Testing defensive skills is tons more annoying and much harder imo.  I've done a lot of parsing in regards to LR/CS/ID, but the damage reduction will vary a lot from mob to mob.  On the old boards I posted what my avoidance #'s were (dodge/block/miss/riposte).  However, I didn't have solid baseline #'s for those.  If any 65 bst without CA/LR that wanted to get wacked for a couple hours and put up some results I'd do the same.

Coprolith

amen tastian

While parsing out defensive skills is easy in theory, in practice it's a lot harder to do then parsing out offensive skills. There's a number of mobs in-game that don't fight back. When testing a offensive skill you can just stand behind them, hit auto-attack, turn off the monitor and go to bed. In the morning you'll have a nice large samplesize. And you've got very good control over the conditions (haste, str etc). If you're testing an activated ability or the effect of some spell that requires more user interaction, but you can still do those parses with only half an eye on the monitor. I usually combine them with some household work. And sometimes, when you have a good baseline parse, you can straight out calculate what an ability or spell will do, no parsing necessary.

Not so with defensive skills parsing, they require constant user attention.  Parsing during normal xp sessions doesn't work very well, because the conditions may change midfight (buffs fading for instance), xp mobs usually vary in level, etc. All that adds to the statistical variance of the parse and as a result you need a huge samplesize. I tried to do this myself, but the results were so depressing i just gave it up.
The best way to do it is to pick a single mob and let it hammer on you, but without fighting back yourself cuz you dont want to kill the mob. So you'll need constant healing (and making sure you never drop below 20% health) and still stay on top of the mob's agro list. And while there are blue con mobs that, once slowed, can barely scratch my natural regen, slowing the mob is counterproductive because it also slows the sampling rate. You have to realize that to parse out the relatively small effects of the defensive skills you need samplesizes of the order of 10000 swings. Even for unslowed mobs, that's several hours of per parse. You can do with less, but that can make the variance in your results so large that you often can't draw hard conclusions from them. So its hours and hours of parsing in which you have to pay constant attention to your hp bar and don't get anything in return but your parse results. As Tastian said, it's annoying not to mention boring. You have to sacrifice a lot of time that could be spent otherwise which is why good defensive parses are sparse.

And you have to get it right from the start, there's no going back once you've acquired a defensive skill as happened to me. Once you've got the baseline, you'll have to do the parses of the defensive skills under the same conditions. But it could take months to get all the necesssary AA points, during which a lot can change, your own gear, changes to the game engine, even revamps of zones so that the mob you started on is no longer there (this happened to my paladin's fav parse mob).

In short, it takes a lot of hard work and deducation to parse out defensive skills, and people that are willing to do so are to be highly commended. Even if their samplesizes are limited because of time and don't provide hard numbers, it still adds to the overall qualitative picture of the defensive skills.

To make matters worse, no one really knows for sure how SoE exactly defines avoidance and mitigation. They sure aren't using the definition the rest of the world uses. And because of that, we can't simply calculate what the listed effects will do from a parsed baseline either. When it says that CA3 will increase your avoidance by 10% that doesn't mean you'll take 10% less damage. As far as people have been able to determine from CA parses the mob will miss 10% more often, but total avoidance also includes ripostes, blocks and dodges. So the effective damage reduction is less then 10%. (note: and keep in mind the parsed numbers have considerably uncertainty to them).
With CS its even less clear what SOE means with a 10% mitigation bonus.

So the best we can do is make educated guesses based on numbers with relatively large uncertainty. And while i have made such educated guesses in the past (i like to think im good at educated guesses, i do alot of that professionally), i'm not going to post them here because i don't want to risk them being considered as hard numbers

Hmmm, got a bit long-winded again did I?  :D

/hugs
Elder Coprolith III
Trollie ferrul lawd of 65 levels (retired)

Tastian

Yeah cop the parses I did on ID were some of the most boring bleh times I've ever spent in EQ.  It took me hours to get the samples I did and they still were only about half the size I'd have really liked.  What we need is a 65 bst with ~1k AC to get pounded on by sun revs or something for a few hours and give us a really solid baseline.  My parses of CA never showed much parse by parse, but there's definetly a difference.  If we could get a decent size sample of 65 bst no CA/ no CS then we could run them againist 65 bst LR5 ID5.  I was actually going to go back and try to get more samples againist sun revs and possibly figure in PE, but the effects are so small from what I've tested so far it'd take me about 5(ish) hours of getting beat on to get the effects of ID to be less than the error margin lol.  This is definetly something I would like to do though.  Since I have my 30 AA banked for GoD I might put in a few hours this weekend and see what I can come up with.  *shrugs*  It's never going to be nearly as nice as how offensive skills break down.  For avoidance you can look at mobs dps and factor in increase in miss%, and for mitigation maybe look at reduction in damage based on mobs DI but *shrugs* more data can't hurt (unless it's flawed lol), but I'll see what I can do.

Coprolith

What i've been thinking about is to have to my druid bot beat on me for a while. Of course the mechanics of PC damage and NPC damage are not quite the same, but im hoping my natural regen is sufficient to counter the damage done by the druid ad infinitum. If that's the case, i can at least do some overnight parses and get really big sample sizes. Just gotta figure out what part of the results are portable to NPC damage.

/hugs
Elder Coprolith III
Trollie ferrul lawd of 65 levels (retired)

Tastian

I can off-set mist panthers regen, but his damage is so small it's hard to see much of anything.  At that point you simply see the shift in hit distribution.  That's mainly what mitigation is and at this point I think the best my results can be will be increase of min hit by ~x% and decrease in max hit by ~y%.  Was hoping to be able to draw a better conclussion based on DI of mob though.  *shrugs*  I'll throw up what I have this weekend and if nothing else it's some data that maybe others can build on or use.  I wish I had done more parses before I got CS/ID.  8(

Ukator Iceblood

Had 9 aa's saved and banked for GoD then twitched and bought Ambi. Oh well:)
Ukator Iceblood and Wolf
70 Beastlord

Kator Kerrath
Berserker

Tunare's Benevolence
Xegony

Lacerate

My most appreciative thanks to all you number crunchers. You've saved me a lot of heartache.