Avoidance Parsing

[Ughbash]

I am curious.

Could someone post what the avoidance of an 85 paladin is on a level 85 test dummy?

Curious how an 85 paladin with shield block maxxed, CA maxxed, avoidance+ 100 and Improved parry 5, Improved Dodge 5 would compare to the following classes.

Monk 85 with CA maxxed, staff blokc maxxed, avoidance 100+, Improved block 5, Improved Dodge 5.

And a ranger 85, CA maxxed, Shield block maxxed, aviodance 100+, Improved parry 5, Improved Dodge 5, running their defensive proc spell.


Most people would say the monk will AVOID more, but I am curious how true that is, my money is on the ranger due to Shield block > Staff block and their defensive proc spell.

[varutia]

Monk's block skill is huge, I remember it blocks something in order of two to three times hits compare to parry.

[frocus]

You talking the defensive proc that fires a 6 second I think it is weaponshield? If its the ranger WS proccing thing I'd assume ranger would have the highest avoidance but monk would likely be close and the pally way behind.

[riou]

you would have to also figure in heroic dexterity and heroic agility, as these can mess up the parses you are for(if there's large differences in either heroic stat).

[Ughbash]

OK, Here are some stats from a monk perspective.

This was not in test, not honestly sure how to get to test server *blush*

This was in bottom of library wher I was campign the librarian for the AC aug as people who have done this camp know mobs are PRIMARILY 85 (though the golems are dark blue).

5695 Hit attempts
1995 hits

So I get hit 35% of the time. (1995/5695).

If anyone cares magelo is

Important stuff is Avoidance 100+.
Max Combat Agility
Improved Dodge 5 (with +8% dodge mod)
Improved Block 5 (with +8% dodge mod)
Staff Block Maxxed.
EDIT: Had Unity 2 the entire time.

Heroic Agility +32.
I did not think Heroic Dex mattered for avoidance but for full disclousre +17.

Now could a paladin and ranger with similar gear and AA post?

[frocus]

Just type /testcopy while ugbash is logged in on normal server should get a message saying char being copied. Make sure you have 2 instances of eq installed 1 for test 1 for normal server to avoid a patch every time you go from normal to test or vice versa.

[frocus]

1. Block check - No Block Skill
2. Parry check - 9.98%
3. Riposte check - 9.01%
4. Dodge check - 11.25%
5. Shield Block check - 5.61% (which is 8% of attacks that make it through 1-4)
6. Avoidance check - 42.9%

For a total of 78.77% defended. This was off of a parse of ~164k attacks so it should be fairly accurate.
Stolen from SK boards. Only dodge 3 no parry mod. Pally at similar level to gregur should have similar dodge, parry, shield block etc.

[frocus]

Also shamelessly stolen from EGN

Heroic Agi - Adds to Dodge, and supposedly Block (not shield block) fire rates and allows for the limit on fall damage before auto death to increase at an unknown rate. The formula on Dodge (and Block if that's true) is +.1% rate per point The Agi increase does very little. While AC softcaps are determined by AGI, beginning at 200 AGI players experience a 1/3 return, overcapping a player at 235 real AGI which works out to 305 displayed. Any increase to agility itself past this point, does absolutely nothing other than to increase your displayed (fake) AC by a rate of 18:1.

Heroic Dex - The boost is +.1% to Parry and Riposte rates (Block too most likely, but not shield block), also the stat increase actually does something which is just above doing nothing. Each point provides an increase to your crit chance, I forget the exact number but it's something on the order of 0.0000000000023429% higher chance to crit per point, and that's not an exaggeration.

[Ughbash]

[Uhaen]

There are several things missing from both parses in order to create an actual tangible display of results; not discounting your efforts at all and they are appreciated, but in order to to put them in a plaque a few things are necessary. Included:

The level of the mob hitting you (Frogs in OS will not hit you as often as trash in Korafax)
The type of mob hitting you (kyvs have a higher accuracy than norm, etc)
The percentage values for each respective defensive ability (block, parry, riposte, dodge, shieldblock, misses, and hits)
And of course, all parses must have equivalent parry/block and dodge modifiers in order to show us if one class has a greater advantage in any defensive ability than others.

Scientific method baby, yeah! You have to nullify all the variables except one. Ideally, we want all parses against a test dummy of the same body type and the same level, and for the same duration (preferably several hours so that [probably!] probability fubars won't show nearly as badly; at most, a few tenths of a percentage point should be our margin of error).

Unfortunately it appears as if TestEQ is taking too long to patch and I won't get a chance to have all the information recorded by the time I get home since I will have to leave soon, so I won't be able to provide my own parses. But for the purposes of ease, I will be parsing:

85 paladin
Max defensive AA
100 Avoid, 77 hAGI, 46 hDEX
Improved Parry V, Improved Dodge V
+8% dodge modifier
Sexy hairdo/dye combination (yes, this does increase your defensive capabilities.)

Ta-ta for now!

Edit: I misspelled plaque. Silly me!

[frocus]

Yeah the Sk's parse surprised the hell out of me when I saw your numbers too is why I posted it. Especially given that it was versus 164k total attacks. Maybe RNG was hating on you for the 6k or so attacks you took. Was vs 85 test dummy forgot to post that. I may patch test and run a test with group geared pally only 64 avoidance(gotta love that dain aug) but 34 Hagi and 35 Hdex.

[Ughbash]

[frocus]

To get into test in the copied folder of eq file you made make a hotkey on desktop of the test eq button named test then log in as normal after the test patch might be pretty big.

[Uhaen]

[Goldax1]

Had a short run in Kaesora library. Same spot, so figured i would post some numbers for you. Its not a long parse but i guess it will do for now. I had a longer parse from yesterday, but that was messed up by the fact that i had alternated a bit between 2hander and 1hander because of a few grelik named that popped, so there had sneaked in some blocks.

Well few stats before i post the parsed numbers:

Magelo is on the right.

Avoidance 100
Max Combat Agility
Improved Dodge 5 (with +8% dodge mod)
Improved parry 5
Hero dex 21
Hero agi 38

Had Unity 2 the entire time.

Tanking summary for: Goldax --- Total damage: 3046271 --- Avg hit: 1427 --- Swings: 4893 --- Defended: 1320 (27%) --- Hit: 2134 (43,6%) --- Missed: 1439 (29,4%) --- Accuracy: 59,7% --- Dodged: 502 (12,3%) --- Parried: 521 (10,6%) --- Blocked: 0 (0%) --- Riposted: 297 (6,8%)

Hitrate 43,6 %

Gonna post the parse of yesterdays encounters also, but take into consideration that its with a mix of 1hander and 2 hander, mostly 2handed thou. Named fights are removed from the parse, so the blocks comes from the few fights where i didnt change back to 2hander.


Tanking summary for: Goldax --- Total damage: 7172278 --- Avg hit: 1424 --- Swings: 11878 --- Defended: 3443 (29%) --- Hit: 5034 (42,4%) --- Missed: 3401 (28,6%) --- Accuracy: 59,7% --- Dodged: 1226 (12,7%) --- Parried: 1336 (11,4%) --- Blocked: 130 (1,1%) --- Riposted: 751 (7,2%)

Hitrate 42,4 %

If we add the 1.1 percent blocked and add to the 42,4 we get 43,5 % which is close to todays parse.

Its still fairly small parses but with the 2 parses being so close i think its fair to say the hirate on this paladin in library on mostly even cons is 43.5ish %.

I can make a full parse with shield equipped all the time if you want.

Goldax

[Goldax1]

Well had a lil go again today, but this is the last time cause i got the aug. Can do some paring on test if needed.

Well same condition as the post above, and its a rather short parse. 1hander and shield equipped all the time. Named mobs not included.

Tanking summary for: Goldax --- Total damage: 2038237 --- Avg hit: 1323 --- Swings: 3747 --- Defended: 1199 (32%) --- Hit: 1540 (41,1%) --- Missed: 1008 (26,9%) --- Accuracy: 60,4% --- Dodged: 418 (14,1%) --- Parried: 410 (11,3%) --- Blocked: 129 (3,4%) --- Riposted: 242 (7,5%)

Hitrate 41.1 %

Goldax

[Ughbash]

Thanks Goldax,

It does appear that the library mobs hit a lot more often than the test 85.

So your last parse admitedly short was about 41%, mine was 35% (also relatively short). So it appears 6% additional avoidance between monk and Paladin (roughly, very roughly) on mobs in the library. I will try to get a good parse on test 85 to compare it to also. Am curious if the 6% will stay on mobs that do not hit as often and also on mobs that hit a lot more (like Kyvs).

Now if we could find out how rangers and warriors parse on these.

[Goldax1]

My normal blockrate seems to be around 4.7% so i got a low number on this short parse. If i should come with an educated guestimate about what the hitrate would be on these mobs on a long parse it would be just around 40 %, leaving a difference between monk and paladin at 5 % at an almost equal gear level. But i will see if i cant do a dummy parse tonight with some lenght on it.

Goldax

[Ughbash]

Just wanted to comment on one more thing I noticed, even if it is not techincally Avoidance.

If anyone doubts the value of a shield, while wearing a shield Goldax had an average hit of 100 less then when using 2hs.

Pre-softcap AC ROCKS!!!!!!

[Goldax1]

Yeah, i wanted to comment on that also in a later post, just didnt have time. Will comment on it later when i got more time.

Well back to avoidance. I moved Goldax to test this evening and had a semi long parse, not sure exactly how long cause when i got back home i was kicked to serverselect, but it is still 34k hits.

Target was a level 85 dummy that i increase min hits on and increased base damage to avoid rune proc removing a hit. I am not entirely sure how gamparse works with hits that are completely absorbed by rune. And i am positive that the minimum damage was not raised high enough to not encounter for entire hits being absorbed. Before i post the parse i will just say that the dummy is weak as hell and the result is really 1-900-MixALot up according to what we see out in the real environment.

Tanking summary for: Goldax --- Total damage: 2561917 --- Avg hit: 391 --- Swings: 34423 --- Defended: 13144 (38,2%) --- Hit: 6545 (19%) --- Missed: 14734 (42,8%) --- Accuracy: 30,8% --- Dodged: 4303 (16,8%) --- Parried: 4380 (13,4%) --- Blocked: 1795 (5,2%) --- Riposted: 2666 (9,4%)

The pure misses are just so big a part of the avoidance now that i dont know what to make of it. (a new test cuold be done while raising the min hits to be above my rune proc, or simply by not having autoattack on myself, or removing the rune aug before /testcopy) Changing some accuracy settings on the dummy might be needed to simulate SoD mobs. Parry, dodge and blocks are also bigger in percent then any of the other parses. This might have to do with the heroric striketrough that SoD mobs apparently have.

Did you have a chance to go one on one with a dummy Ughbash?

Goldax

[Ughbash]

[frocus]

So the sk's numbers were inflated due to the gimpness of test mob it looks like?

[Uhaen]

Removing the false-positives of Rune:

Block the spell. It can cause your graph to display 20 more values than actually exist, and even more values if the DI1 hit is less than 700 (for rune5).

[Ughbash]

Attempts 22552, defended 10932, Missed 7813, hit 3807.

So hit 3807 times out of 22552 or Hit 16.9 precent of the time.

Defintley test 85 does not give results similar to real game.

I will try to run a test on test 95 tomorrow.


Hmm also going to post some other numbers from today in two camps both of which are too small of a parse to give good data but are closer to real data that we see in the game.

Bloodfields variuos mobs Nocs, Ukuns, Ikav, and one Thug.
Attempts 4732, Defended 2047, Missed 981, Hit 1704.
So hit 1704 out of 4732 or 36% of the time.


Korascian Warrens, Instance and then at th thralls camp.
Attempts 7116, Defended 3066, missed 1542, Hit 2508.
So hit 2508 out of 7116 or 35.2% of the time.

I really wish the test 85 gave realistic numbers hopefully test 95 will be closer ot what we see in game.

[Ughbash]

As promsied a parse form test 95.

29563 ATTEMPTS, 14166 DEFENDED, 10310 MISSED, 5087 hITS

So hit in total for 17.2 percent of the time.

If only mobs in the real game hit like this I could tank, BUT I see now why SOE thinks monks are great tanks, we do awsome on the gimp things they have set for parsing, it just is not reflected in the real game. It woud appear that the parser Hits about half as much as real mobs do.

[Brohg]

Developers do not rely on Test dummies for data.

[Khauruk]

[Ughbash]

I suspected they did not use test dummies.

I suspect they have their mathematical formuals which portray how they think stuff works, but doesn't.

I know a few years ago they just used formulas rather then parsing, perhaps it has changed.

[Khauruk]

They have a simulation environment where they can take a character ("yours" even, from my understanding) and run through a million hits as fast as the computer can process those, rather than wait for it to happen "irl."

[Goldax1]

I am still not totally familiar with the test dummies, so dont know if we can change their accuracy to be in line with ingame values, and wont have time to play around with them, as my time will be spend hunting for the new random augs in tier 4 and 5

I have however been spending some time in Toskirakk and in Warrens the last week and have a bigger parse then the semi short ones from Library.

Same conditions as before, shield equipped all the time.

Tanking summary for: Goldax --- Total damage: 20045324 --- Avg hit: 1558 --- Swings: 31626 --- Defended: 10433 (33%) --- Hit: 12859 (40,7%) --- Missed: 8334 (26,4%) --- Accuracy: 60,7% --- Dodged: 3434 (13,9%) --- Parried: 3519 (11,6%) --- Blocked: 1340 (4,2%) --- Riposted: 2140 (8%)

I belive the mobs are level 83-85.

In tier 5 on shorter parses i seem to see a hitrate above 43 %. I guess i need to work on my heroric dex and agi.

About the AC difference in the 2 Library parses where one is with 2hander and one is with 1hander. My 2hander is a chainrender's mace with 42 AC and i am using a Pneumatic Powerblade with 58 AC as 1hander. My shield is Buckler of Corruption with 203 AC. So other then the shield AC the Ac between the 2 setups are for all intent and purpose the same. In this case the Shield AC reduces the avarage damage by 7.29 % and removes 3.2 % of all incomming damage by blocking them (this number is over time bigger as that particular parse had a very low block number). I know this is all known info, but here it is again.

Gotten any luck with the warriors or rangers to do similar avoidance parses?

Goldax

[Ughbash]

[Brohg]

What happens in game that doesn't follow the strictures of those formulas?

[Khauruk]

[Ughbash]

[Khauruk]

They don't consider the Test dummies to be accurate representations of true mob dps, they're just not going to prioritize fixing them over other game issues. They've been unreal representations since at least back to Prophecy of Ro, and I'm sure before that too. They're still very useful for many purposes however...just not all.