AC Seeming to have little to no impact

[heartbrand]

Been doing some HoT / Master of the Guard and I am seeing little to no statistical difference in tanking between 1200ac and drastically lower AC. Do we have any idea on the formula being used at all?

[Smedy]

i think ac scales down above 1000 (or whatever number it's set at) drastically, much like int and wisdom there's no point in getting more than a certain number (around 210?). confirm / deny?

[Throndor]

not true. AC is a stat which currently on live, does, and always has operated on a bell curve. It was revanped in the SOF era after nearly every class was capable of reaching ac values that trivialized group content when raid geared, while simultaneously causing group geared characters to get destroyed by group mobs. This anomaly stemmed from developer attempts toward providing challenging group content to raiders, and in the end they found the only way they could fix the disparity was to revamp the tables.

With that said, when a raid tank was sitting in top end raid gear with maximized ac group content was classicly and still is currently, rarely if ever hitting u at max value. Tbf, there is a point at which u are seeing minimum hit so often where more ac is not really an effective gear decision on LIVE, but as it currently stands mob attack values throughout p99 are substantially higher than they were in classic. Either its outright coded incorrectly, or its an intentional sleight to add "epicness" to the group content on p99.
In maximized ac from raid gear, P99 kunark era tanks are seeing little if any difference from group geared tanks in parsing toward the low end of the DI table and are forced to counter this primarily with larger hp pools which does little to improve the efficacy of the healers responsible for sustaining the damage, aside from buying an extra round of combat or 3 per ch. It's not surprising though. When it takes 5 slows in a row to slow a mob 4 to 6 levels blue to you at a camp like fungi king despite it being tashed and maloed, its pretty telling that things like mobattack and MR are drastically exaggerated on P99.

[dbouya]

Back during live velious there were rumors about ac being broken, but the main tank in the raid still wanted the druid hp buff that resulted in like 5more total ac than the cleric one.

Trouble slowing doesn't sound that odd too the golem and deeper areas of seb used to be pretty hardcore luck required to land spells in 2000 too from what I remember.

[Schortt]

Hmm. Mitigation seems to be scaling pretty well within my means of testing it.

All data is a 60 Iks Monk getting hit by the same level 59 Crystal Destroyer in Velks (max hit 350, min hit 81) for roughly 10 minutes.

All AC values were with identical buffs (aego) and achieved by removing gear down to 877. To get below that I had to load up on plat to get some AC penalty going on.

1171 AC
--- Total damage: 32455 --- Avg hit: 186 --- Swings: 399 --- Defended: 100 (25.1%) --- Hit: 174 (43.6%) --- Missed: 125 (31.3%) --- Accuracy: 58.2% --- Dodged: 18 (5.7%) --- Parried: 0 (0%) --- Blocked: 67 (16.8%) --- Riposted: 15 (4.5%) --- Absorbed: 0 (0%)

1070 AC
--- Total damage: 44798 --- Avg hit: 199 --- Swings: 481 --- Defended: 112 (23.3%) --- Hit: 225 (46.8%) --- Missed: 144 (29.9%) --- Accuracy: 61% --- Dodged: 17 (4.4%) --- Parried: 0 (0%) --- Blocked: 75 (15.6%) --- Riposted: 20 (4.9%) --- Absorbed: 0 (0%)

969 AC
--- Total damage: 40925 --- Avg hit: 210 --- Swings: 447 --- Defended: 115 (25.7%) --- Hit: 194 (43.4%) --- Missed: 138 (30.9%) --- Accuracy: 58.4% --- Dodged: 28 (7.8%) --- Parried: 0 (0%) --- Blocked: 68 (15.2%) --- Riposted: 19 (5%) --- Absorbed: 0 (0%)

877 AC (naked + aego)
--- Total damage: 45533 --- Avg hit: 229 --- Swings: 498 --- Defended: 142 (28.5%) --- Hit: 198 (39.8%) --- Missed: 158 (31.7%) --- Accuracy: 55.6% --- Dodged: 28 (7.3%) --- Parried: 0 (0%) --- Blocked: 92 (18.5%) --- Riposted: 22 (5.4%) --- Absorbed: 0 (0%)

770 AC (naked + aego + enough weight to drop my AC that far)
--- Total damage: 45136 --- Avg hit: 246 --- Swings: 436 --- Defended: 103 (23.6%) --- Hit: 183 (42%) --- Missed: 150 (34.4%) --- Accuracy: 55% --- Dodged: 19 (5.4%) --- Parried: 0 (0%) --- Blocked: 66 (15.1%) --- Riposted: 18 (4.9%) --- Absorbed: 0 (0%)

I will try out MotG or some other much higher level mob when I have the time.

[Raev]

95% of the time, the people complaining about these things are just statistically ignorant. You can't expect to eyeball AC.

[cormag]

On the other hand: the iksar AC curve seems to be a little more even compared to the human data here http://www.project1999.com/forums/showpost.php?p=1316375&postcount=92

[Schortt]

I'll throw a betabuffed human monk into the mix next time I do some testing to see if it matters (probably a warrior too). Cliff Golems are handy for testing on Blue because you can convince your cleric buddy to hammer over and spam you with heals for awhile, but their relatively low max hit means it's more difficult to show significant differences. Perun's test is great work and shows that AC isn't doing as much as some people would like, but the closeness of the numbers makes it really tough to tell anything other than AC values over 800ish are having less of an effect than the RNG for test runs that are short enough to preserve your sanity.

A lot of the concern about AC seems to stem from this idea that we need to make sure it's working for Velious when numbers get bigger and fights last longer so people are making informed choices about gearing - just another reason to be testing on hard hitting Velious mobs on Beta rather than sissy stuff on Blue.

[wycca]

Thanks so much for testing, the # of swings seems good. It's obvious that AC scaling does work on p99....likely in group content at least (where I've always thought it had noticable returns). Not so sure about raid content sometimes =P.


Softcap AC & returns over softcap testing on p99 does require eliminating alot of variables imo.

- Monk low-weight bonus ac, a mitigation ac bonus on top of a softcap?
- Iksar bonus ac - a mitigation ac bonus on top of a softcap?

Either one would raise the apparent softcap and cover some AC scaling/returns issues. Heck, it could be possible that an iksar monk is tanking better than warriors because of a softcap and returns over cap issue.

- Need to know raw ac from a gearset you're parsing really.
- Ideally would eliminate iksar ac and monk bonus ac, and conduct testing soley through scaling of raw (ie worn) ac to discover a softcap. You could then go back and repeat the test with monk ac and iksar ac if you wanted to know what those are worth on p99 btw.
- Should aim to discover softcap, then returns over cap. Would be interesting to conduct test with warrior to compare.

Boring for sure, I'd be game to help conduct the tests though, would finally give us some insight into the softcap/returns system on p99 and help people evaluate if AC is truly worth anything.

[deadlycupcakez]

I did some testing; here's my conclusions after tanking an ancient sentry in ST twice, once with 1166 ac on a warrior, and the other 847.

1166 ac data:
285 hits
253.3894737 avg
316 max
115 min
140 18 14 29
miss riposte dodge parry

fight lasted 20 minutes


847 ac data:

hits miss rip dodge parry
331 180 33 29 34

276.3927492 avg
122 min
316 max

fight lasted 20 minutes




A reason there were more hits is partially due to, on the first attempt, I messed up and died with the mob at ~30%, however the data's extremely comparable.

My conclusion? AC's effectiveness is close to 0.

Except that you took approx 7,600 more damage in the second fight over the same amount of time. Which means over 20 minutes of fighting those mobs you will need at least a few more CH's

[deadlycupcakez]

Even with the same amount of hits you'll take 6500 more damage. I'm just pointing out from a clerics viewpoint in a longer term group setting, the mitigation of that extra 6,500 damage seems very valuable to my mana pool/appears to be working properly.

[Ciroco]

Do DB and DI work the same way here as they do on Live? If so, parsing against one type of mob isn't enough to prove a lack of effectiveness.

[Bodybagger]

Hmm. Mitigation seems to be scaling pretty well within my means of testing it.

All data is a 60 Iks Monk getting hit by the same level 59 Crystal Destroyer in Velks (max hit 350) for roughly 10 minutes.

All AC values were with identical buffs (aego) and achieved by removing gear down to 877. To get below that I had to load up on plat to get some AC penalty going on.

1171 AC
--- Total damage: 32455 --- Avg hit: 186 --- Swings: 399 --- Defended: 100 (25.1%) --- Hit: 174 (43.6%) --- Missed: 125 (31.3%) --- Accuracy: 58.2% --- Dodged: 18 (5.7%) --- Parried: 0 (0%) --- Blocked: 67 (16.8%) --- Riposted: 15 (4.5%) --- Absorbed: 0 (0%)

1070 AC
--- Total damage: 44798 --- Avg hit: 199 --- Swings: 481 --- Defended: 112 (23.3%) --- Hit: 225 (46.8%) --- Missed: 144 (29.9%) --- Accuracy: 61% --- Dodged: 17 (4.4%) --- Parried: 0 (0%) --- Blocked: 75 (15.6%) --- Riposted: 20 (4.9%) --- Absorbed: 0 (0%)

969 AC
--- Total damage: 40925 --- Avg hit: 210 --- Swings: 447 --- Defended: 115 (25.7%) --- Hit: 194 (43.4%) --- Missed: 138 (30.9%) --- Accuracy: 58.4% --- Dodged: 28 (7.8%) --- Parried: 0 (0%) --- Blocked: 68 (15.2%) --- Riposted: 19 (5%) --- Absorbed: 0 (0%)

877 AC (naked + aego)
--- Total damage: 45533 --- Avg hit: 229 --- Swings: 498 --- Defended: 142 (28.5%) --- Hit: 198 (39.8%) --- Missed: 158 (31.7%) --- Accuracy: 55.6% --- Dodged: 28 (7.3%) --- Parried: 0 (0%) --- Blocked: 92 (18.5%) --- Riposted: 22 (5.4%) --- Absorbed: 0 (0%)

770 AC (naked + aego + enough weight to drop my AC that far)
--- Total damage: 45136 --- Avg hit: 246 --- Swings: 436 --- Defended: 103 (23.6%) --- Hit: 183 (42%) --- Missed: 150 (34.4%) --- Accuracy: 55% --- Dodged: 19 (5.4%) --- Parried: 0 (0%) --- Blocked: 66 (15.1%) --- Riposted: 18 (4.9%) --- Absorbed: 0 (0%)

I will try out MotG or some other much higher level mob when I have the time.

To clarify, when you say you used plat weight to drop your AC further, does that mean your dropped you AGI to drop AC and did you drop below agi cap at 80 or whatever it is? I was surprised not to see even more dmg here

As far as this data shows you're taking roughly 10% less dmg per hit per 100ac increase below 1k AC and about 6% less per 100 AC beyond there? 400+ AC getting you 25% less incoming dmg based solely on average damage per hit (what AC directly affects, right?)

And per Youlaths Data, shows you took roughly 10% less dmg ~850AC vs ~1150AC which isn't in line with the monk parse shown...

Does anyone have parsing showing # of hits for Max dmg as one of the figures? RNG will play a factor always but I feel # of hits for max dmg being diminished, # of min dmg hits increased, as well as overall impact of of RNG average hits would be what AC affects, but I know nothing about any of it, purely speculation as to how I'd imagine it working for dice roll dps/mitigation system

[koros]

One important thing to keep in mind is that mobs have a 1-20 interval they can hit on + their damage bonus. If you have a mob that has a min hit of 201 and a max hit of 220 (200 damage bonus), it's not going to seem like ac is doing as much.

To really calculate the effects of AC we need to determine how it changes the average interval hit. Maybe going from 600 ac to 1100 ac changes the average hit on a mobs interval from 17 to 6. This can be big, especially for mobs with low damage bonuses and high DIs.

Correct method to parse AC:

Determine mobs minimum hit and damage interval. Subtract damage interval from min hit to determine damage bonus. Now, figure out the average and median interval hits, not just the average overall hit.

Here's an example. A mob has a damage bonus of 120, and an interval of 9. This leaves us with a a minimum hit of 129 (120 + 1*9), and a max hit of 300 (120 + 20*9). The mob can therefore hit anywhere in that range in intervals of 9: 129, 138, 147, 156, etc.

If parsing at 600 ac, you got an average hit of 255 you know that the mobs average interval hit is 17.2 (255-100 = 155/9 ~= 17.2).

If you then parse at 1100 ac and get an average hit of 200 you know the mobs average interval is (200-100 = 100/9 ~= 11.1).

That's a huge change in theoretically removable damage (damage bonus damage can't be removed)

Sorry to be redundant on that, but it's important to make that clear if you want AC parses to be meaningful. What mob you parse with matters. Ultimately I bet we're going to see a pretty linear change in average interval hit, on a per mob basis, as AC varies.

If anyone wants to provide those min/max/interval numbers on existing parses I'd be happy to plug the data in.

[Nirgon]

I just wanna stroke your beard

[koros]

bump?

[Schortt]

I'm still working on collecting data. Will hopefully have enough to throw up another post after the weekend. I've got comparisons on monk and war tanking ST trash. I still need to do some more straightforward comparisons on lower level stuff as mob level is definitely making a difference w/r/t AC effectiveness.

If anyone wants to provide those min/max/interval numbers on existing parses I'd be happy to plug the data in.
I'll try to make this available too.

[koros]

Nice. Do you have the min/max hit of that mob you were tanking? If I can figure out the interval we can probably figure out really quickly how much of an effect AC is having.

[Throndor]

The Siren's den in Siren's Grotto may also prove to be a somewhat effective place to gather parse data, as the mob types are very limited and lvl ranging is minimal while still providing consistently above-lvl-55 mobs.

Sleeper's Tomb's trash mob-level varies considerably, and may prove problematic in obtaining consistency in your independent variables (the mob's lvl, class-type etc).

[Throndor]

Send me a tell if u catch me on this weekend. I can hop over to test with a beta buffed cleric, and my chanter to slow and extend the window which u can keep your tank's propped up if u want a prolonged test.
For the sake of consistency, you probably shouldn't even kill the mob your using to parse, and just camp or blur the mob between parses, although maintaining agro without dmg may be a bit problematic with a warrior/monk.

[Schortt]

The offer is appreciated, but I would encourage you to just collect data on your own as I usually do this whenever I happen to have some spare time and it's often at very random hours. I'll send you a tell this weekend if I remember in any case.

I do always use the exact same mob to run a series of parses, never killing it, and I don't use slows because it just makes parsing sessions take that much longer. Mobs that flurry are actually a plus because you can get a statistically useful number of hits that much faster! If your tank is dying, roll more cleric alts and betacopy/buff them. Boxing on beta is legit, use it to your advantage.

[Tapatio]

Hmm. Mitigation seems to be scaling pretty well within my means of testing it.

All data is a 60 Iks Monk getting hit by the same level 59 Crystal Destroyer in Velks (max hit 350) for roughly 10 minutes.

All AC values were with identical buffs (aego) and achieved by removing gear down to 877. To get below that I had to load up on plat to get some AC penalty going on.

1171 AC
--- Total damage: 32455 --- Avg hit: 186 --- Swings: 399 --- Defended: 100 (25.1%) --- Hit: 174 (43.6%) --- Missed: 125 (31.3%) --- Accuracy: 58.2% --- Dodged: 18 (5.7%) --- Parried: 0 (0%) --- Blocked: 67 (16.8%) --- Riposted: 15 (4.5%) --- Absorbed: 0 (0%)

1070 AC
--- Total damage: 44798 --- Avg hit: 199 --- Swings: 481 --- Defended: 112 (23.3%) --- Hit: 225 (46.8%) --- Missed: 144 (29.9%) --- Accuracy: 61% --- Dodged: 17 (4.4%) --- Parried: 0 (0%) --- Blocked: 75 (15.6%) --- Riposted: 20 (4.9%) --- Absorbed: 0 (0%)

969 AC
--- Total damage: 40925 --- Avg hit: 210 --- Swings: 447 --- Defended: 115 (25.7%) --- Hit: 194 (43.4%) --- Missed: 138 (30.9%) --- Accuracy: 58.4% --- Dodged: 28 (7.8%) --- Parried: 0 (0%) --- Blocked: 68 (15.2%) --- Riposted: 19 (5%) --- Absorbed: 0 (0%)

877 AC (naked + aego)
--- Total damage: 45533 --- Avg hit: 229 --- Swings: 498 --- Defended: 142 (28.5%) --- Hit: 198 (39.8%) --- Missed: 158 (31.7%) --- Accuracy: 55.6% --- Dodged: 28 (7.3%) --- Parried: 0 (0%) --- Blocked: 92 (18.5%) --- Riposted: 22 (5.4%) --- Absorbed: 0 (0%)

770 AC (naked + aego + enough weight to drop my AC that far)
--- Total damage: 45136 --- Avg hit: 246 --- Swings: 436 --- Defended: 103 (23.6%) --- Hit: 183 (42%) --- Missed: 150 (34.4%) --- Accuracy: 55% --- Dodged: 19 (5.4%) --- Parried: 0 (0%) --- Blocked: 66 (15.1%) --- Riposted: 18 (4.9%) --- Absorbed: 0 (0%)

I will try out MotG or some other much higher level mob when I have the time.

Nice.

I wonder if a plate class would experience different findings than a monk.

[Nilasbik]

Monks underweight have the same ac return as knights. So you cant compare any plate class. You have to compare them to sk/pal. Wars in lead then sk/pal/mnk (underweight), then the rest of the plate classes, then chain, leather, cloth. Monks overweight are on the leather scale. Over weight on a monk doesnt mean that current weight is higher than current STR either. So his agi doesnt change when he goes over the mnk weight. Which at 60 is 20.9. Higher than that drops your ac on a curve to a certain point. So he could incrementally increase his weight to further drop the AC without dropping to the agi penalty.