AC softcap and low level hardcap discussion

[bcbrown]

Originally Posted by DeathsSilkyMist [You must be logged in to view images. Log in or Register.]

One extreme way would involve removing the worn AC * 4 / 3 AC inflation step. While this step was confirmed by Haynar, it doesn't necessarily mean it still occurs on P99. You would also add a step to use spell AC for softcap.
...
In this scenario you wouln't hit softcap, or you would just barely go over. I am not saying I am leaning towards this solution. But it was one way I was thinking about it at the time.

Thanks for explaining your thinking. Certainly possible, but I agree it's a little implausible.

I wanna take a step back for a second. The results of my experiment were surprising, and we know it's surprising because it's contrary to what you predicted the outcome would be. Any time that the (made in jest) theory "bcbrown never sees an effect from shield AC but DSM always sees an effect" is consistent with the results, we know there's still a lot we don't know.

What we do know is this: sometimes shields can have an effect, and sometimes they don't. They can have an effect at level 5, and they can have an effect at level 60. I think at this point almost anything else is still speculation. We don't know what impact class has (priest vs int-caster vs melee/hybrid), nor what impact armor-class has (plate vs chain vs leather vs cloth). We don't know how spell AC interacts with caps, and we don't know whether 1 ac from spells has an equivalent impact to 1 ac from armor.

Since my gear isn't great, I won't be able to help much with running experiments to help answer those questions. I did run a small experiment. First side had no Bladecoat. Second side had Bladecoat (37 ac on wiki), and I took off my pants and gloves (17 and 20 ac, respectively). Both sides had potg, and I ended up with 988 and 826 hits. I haven't looked at the results yet. I don't think we know enough to make a prediction worthwhile, but I think we can say that if there is a difference, that's evidence that spell ac is treated differently from worn ac, while if there is no difference, that's not really evidence for anything. And even if there is a difference, the number of hits for each side aren't enough to have really solid evidence. That all sound right to you?

[DeathsSilkyMist]

Originally Posted by bcbrown [You must be logged in to view images. Log in or Register.]

Thanks for explaining your thinking. Certainly possible, but I agree it's a little implausible.

I wanna take a step back for a second. The results of my experiment were surprising, and we know it's surprising because it's contrary to what you predicted the outcome would be. Any time that the (made in jest) theory "bcbrown never sees an effect from shield AC but DSM always sees an effect" is consistent with the results, we know there's still a lot we don't know.

What we do know is this: sometimes shields can have an effect, and sometimes they don't. They can have an effect at level 5, and they can have an effect at level 60. I think at this point almost anything else is still speculation. We don't know what impact class has (priest vs int-caster vs melee/hybrid), nor what impact armor-class has (plate vs chain vs leather vs cloth). We don't know how spell AC interacts with caps, and we don't know whether 1 ac from spells has an equivalent impact to 1 ac from armor.

Since my gear isn't great, I won't be able to help much with running experiments to help answer those questions. I did run a small experiment. First side had no Bladecoat. Second side had Bladecoat (37 ac on wiki), and I took off my pants and gloves (17 and 20 ac, respectively). Both sides had potg, and I ended up with 988 and 826 hits. I haven't looked at the results yet. I don't think we know enough to make a prediction worthwhile, but I think we can say that if there is a difference, that's evidence that spell ac is treated differently from worn ac, while if there is no difference, that's not really evidence for anything. And even if there is a difference, the number of hits for each side aren't enough to have really solid evidence. That all sound right to you?

No problem!

Shield AC will only have a special effect if you are above the softcap. Your data is consistent with how shield AC would work if you don't hit the softcap.

What your data shows is that 198 worn AC + 61 Spell AC does not hit the softcap for Druids. At this point we need to find the softcaps. I can work on that since I have the AC. Priest softcaps should be the same.

It would be cool to see some more Spell AC tests on your end. I am curious if spell AC acts different even under the softcap. If the EQEMU code is correct, then spell AC is a bit different from worn AC, as it gets divided by 4 and doesn't get inflated by the inflation step. I don't think it works like that on P99, but not sure.

Just to be clear, AC doesn't affect how often the mob misses you. It only affects mitigation after you get hit.

[bcbrown]

Originally Posted by DeathsSilkyMist [You must be logged in to view images. Log in or Register.]

What your data shows is that 198 worn AC + 61 Spell AC does not hit the softcap for Druids. At this point we need to find the softcaps. I can work on that since I have the AC. Priest softcaps should be the same.

My data does not show that. It raises that question. I realize how pedantic this sounds, but an experiment cannot result in a conclusion that it was not attempting to answer. I agree that at this point we need to find the softcaps, but my experiment does not answer that question, it raises that question. Not exactly a rigorous source, but here's a comic addressing a similar topic: https://xkcd.com/882/

Any speculation on results before I analyze the parse I did with and without Bladecoat?

Edit: to be clear, that experiment I did was not with 198 worn AC, because I was removing various slots for each parse. I usually have prexus totem equipped, but took that off to free up a slot, then removed 23 ac either via lodi shield or whatever else. I also added 1 ac by upgrading a bracer, so I think that puts it at 171 worn ac + 61 spell ac.

[DeathsSilkyMist]

Originally Posted by bcbrown [You must be logged in to view images. Log in or Register.]

My data does not show that. It raises that question. I realize how pedantic this sounds, but an experiment cannot result in a conclusion that it was not attempting to answer. I agree that at this point we need to find the softcaps, but my experiment does not answer that question, it raises that question. Not exactly a rigorous source, but here's a comic addressing a similar topic: https://xkcd.com/882/

Any speculation on results before I analyze the parse I did with and without Bladecoat?

Data is data. Your data supports the softcap theory.

No speculations right now. I haven't done much spell AC testing.

[DeathsSilkyMist]

Originally Posted by bcbrown [You must be logged in to view images. Log in or Register.]

Edit: to be clear, that experiment I did was not with 198 worn AC, because I was removing various slots for each parse. I usually have prexus totem equipped, but took that off to free up a slot, then removed 23 ac either via lodi shield or whatever else. I also added 1 ac by upgrading a bracer, so I think that puts it at 171 worn ac + 61 spell ac.

Thanks for clarifying. So you were even further from any possible softcap. That makes sense.

[bcbrown]

Originally Posted by DeathsSilkyMist [You must be logged in to view images. Log in or Register.]

Data is data. Your data supports the softcap theory.

No speculations right now. I haven't done much spell AC testing.

My data supports the softcap hypothesis. A theory has been confirmed by experiments done to test the hypothesis. But we're not going to resolve that philosophical difference in this thread so I'll try not to keep harping on it too much [You must be logged in to view images. Log in or Register.]

With bladecoat (37 ac on wiki):
988 hits
41.155 dmg/hit

Without bladecoat (wearing crystal chitin gauntlets and gladiator's chain leggings):
826 hits
43.393 dmg/hit

Looks like it's worth further investigation!

[DeathsSilkyMist]

Originally Posted by bcbrown [You must be logged in to view images. Log in or Register.]

My data supports the softcap hypothesis. A theory has been confirmed by experiments done to test the hypothesis. But we're not going to resolve that philosophical difference in this thread so I'll try not to keep harping on it too much [You must be logged in to view images. Log in or Register.]

With bladecoat (37 ac on wiki):
988 hits
41.155 dmg/hit

Without bladecoat (wearing crystal chitin gauntlets and gladiator's chain leggings):
826 hits
43.393 dmg/hit

Looks like it's worth further investigation!

Haynar, the programmer who worked on this, said there was a softcap. The softcap is indeed a theory confirmed by Haynars own experiments[You must be logged in to view images. Log in or Register.] We are simply reconfirming the theory.

The squelch point hypothesis has yet to be confirmed.

I agree, more inestigation is needed on spell AC.

[DeathsSilkyMist]

Posting my 3 initial tests to find the softcap. They are all against Shiel Glimmerspindle, 1000 hits each:

=========================
100 AC Test, No Item in Secondary Slot
=========================

DV, Count
19, 121
22, 33
25, 33
28, 39
32, 36
35, 46
38, 31
42, 35
45, 38
48, 38
52, 44
55, 41
58, 35
62, 38
65, 36
68, 30
72, 36
75, 42
78, 31
82, 217

Total Damage = 53149

=========================
150 AC Test, No Item in Secondary Slot
=========================

DV, Count
19, 216
22, 35
25, 39
28, 38
32, 30
35, 40
38, 36
42, 34
45, 30
48, 44
52, 32
55, 49
58, 40
62, 37
65, 43
68, 39
72, 37
75, 43
78, 32
82, 106

Total Damage = 47028

=========================
200 AC Test, No Item in Secondary Slot
=========================

DV, Count
19, 289
22, 41
25, 41
28, 33
32, 32
35, 43
38, 43
42, 38
45, 40
48, 33
52, 38
55, 38
58, 37
62, 39
65, 33
68, 39
72, 41
75, 41
78, 34
82, 27

Total Damage = 41805

So far this suggests the softcap is not between 100 AC and 200 AC. The damage difference between 100 AC and 150 AC is similar enough to 150 AC and 200 AC.

My 363 AC tests Show that the damage I took at 363 AC without a shield was 39236.

Going from 100 AC to 200 AC reduced my damage by 11346.

Going from 200 AC to 363 AC reduced my damage by 2569.

There is almost certainly a softcap somewhere between 200 AC and 363 AC. I'll try to pinpoint it when I run the next few tests.

[DeathsSilkyMist]

Did a few more tests to hone in where the softcap is:

========================
223 AC Test, No Shield Secondary
========================

DV, Count
19, 340
22, 33
25, 36
28, 40
32, 40
35, 27
38, 38
42, 42
45, 41
48, 28
52, 31
55, 47
58, 37
62, 29
65, 28
68, 26
72, 40
75, 37
78, 46
82, 14

Total Damage = 39948

=======================
223 AC Test, Lodi Shield Secondary
=======================

DV, Count
19, 341
22, 32
25, 35
28, 36
32, 40
35, 26
38, 33
42, 33
45, 31
48, 37
52, 31
55, 39
58, 42
62, 44
65, 43
68, 38
72, 42
75, 35
78, 37
82, 5

Total Damage = 40312

=======================
235 AC Test, Lodi Shield Secondary
=======================
DV, Count
19, 366
22, 33
25, 36
28, 35
32, 40
35, 36
38, 34
42, 39
45, 27
48, 31
52, 35
55, 36
58, 23
62, 42
65, 47
68, 35
72, 35
75, 31
78, 30
82, 9

Total Damage = 38829

=======================
250 AC Test, Lodi Shield Secondary
=======================

DV, Count
19, 370
22, 53
25, 36
28, 43
32, 32
35, 25
38, 29
42, 24
45, 25
48, 39
52, 21
55, 31
58, 35
62, 24
65, 33
68, 44
72, 41
75, 40
78, 41
82, 14

Total Damage = 39056

======================
250 AC Test, No Shield Secondary
======================

DV, Count
19, 333
22, 25
25, 23
28, 39
32, 35
35, 36
38, 46
42, 54
45, 30
48, 22
52, 44
55, 44
58, 32
62, 40
65, 35
68, 39
72, 40
75, 38
78, 39
82, 6

Total Damage = 40581

We can see that the 223 AC tests show the same pattern that Bcbrown saw, where the shield parse matches the non-shield parse. This makes sense, as Bcbrown said he had 171 worn AC + 61 Spell AC = 232 AC. While we are still not sure exactly how spell AC works, this total value is in the ball park.

Once we get to 250 AC, the shield parse starts to become better than the non-shield parse. The 235 AC shield parse shows basically no difference compared to the 250 AC shield parse.

This means the softcap for Shamans (and probably Druids/Clerics as well) is between 223 AC and 235 AC.

I believe the softcap returns may be 0.17 for Shaman, which is the EQEMU value for low chain classes. This would be like Shaman and Ranger.

This is because in my 363 AC Tests, my no shield parse is 39K damage with 380 minimum hits.

The 250 AC test with a shield in this post is 39K and close to 380 minimum hits. 23 AC is Shield AC, so taking that out would give us 227 AC. 363 AC - 227 AC = ~136 softcapped AC * 0.17 = ~23 AC, which is roughly the same AC as my shield. This suggests that 136 softcapped AC gave me the equivalent of ~23 worn AC.

[kjs86z2]

you can do it DSM

we believe in you