AC cap and shields

[Phantron]

No, you're the one who used DPS/AC (I was lazy and copy & pasted a few things) which is wrong because even if AC benefit extends indefinitely (which it does not) you'd still get the conclusion that it stops being not as effective.

Suppose benefit of AC goes on indefinitely so that it can actually reduce DPS to 0 (which is obviously wrong). When DPS becomes 0, then DPS/AC = 0. By your logic, it's a waste to get AC to this level because DPS/AC = 0 which is as small as this value can possibly be. This can't be further from the truth because at this level of AC, if it's possible to reach it, you'd be indestructible and not being able to die in any possible way to melee sure is worth it. DPS/AC simply has no meaning as a useful metric to measure AC's effectiveness.

On a fundamental level, the only benefit of AC is reducing the damage taken, namely DPS. Parses show quite conclusively that this effect is linear. If 100 AC reduced DPS by 1% at 1600 then 100 AC still reduces DPS by 1% at 3200 AC. There is a point where you can't reduce DPS anymore (namely DPS = mob hitting all mins), but this is not remotely attainable even on old world trash (there is no indication that a giant rat wouldn't do 1% less damage to you if you go from 3300 to 3400 AC, except you probably don't care)

It should be noted that I'm against the AC camp. However the notion that AC's usefulness changes depending on your AC is simply wrong. You never have too much or too little AC to favor a HP vs AC trade one way or another. The value of less DPS, compared to HP, depends on your total HP. Therefore the value of AC only changes when your total HP changes significantly. Imagine tomorrow everyone has 100000 HP. Is it better to get 100 more HP or take 1% damage? It's obvious the latter because taking 1% less damage with 100K HP is like gaining 1% more HP (1000). Conversely if tomorrow they took away a 0 from everyone's HPs so that the highest tanks are now at 2000 HP instead of 20K, then getting 100 HP (5% more HP) suddenly looks a lot more attractive than taking 1% less damage.

[Naubi]

I still think you theorising only amounts to anything in the case of someone with a ch chain (or other regular stacked healing) on them. Which is fine for warriors as that's you primary role (on raids at least).

For where you have unreliable healing (or healing from HoTs or HoT like effects), I prefer AC. This is my usual situation.

"The value of less DPS, compared to HP, depends on your total HP." this is just plain wrong except in a particular kind of situation.

[Gunnar]

Originally Posted by Xanathol

I am not saying there is a second softcap but rather an 'effective' cap where more AC means nothing. I don't know how esle to state this than how I already have. Just as you mentioned, when all hits become mins and the distribution is a flat line, that would be the 'effective' cap where 1 more point of AC does nothing for you. Refer to the 'gnoll pup' example.

Not sure what this has to do with this thread then. It's er obvious. We all know you can get mobs that are so trivial so as to be relagated to min hits where more ac may not make a difference. This has nothing to do with actual content we play in.

Originally Posted by xanathol

What I said was untrue about Phantron's post was where he said that gains in AC ( over the softcap ) were linear and the benefit of 1 point of AC is only a function of your HP. What I am saying, and have shown in the numbers I posted from the parse data, is that as your AC gets higher with a fixed amount of HP, the 'value' of an additional point gets lower; that is, its not linear, and the average hit / AC total shows just that. Phantron used DPS / AC, but when you do that, the numbers that show the difference just become smaller, so much so that it looks like "noise". Looking at the average single hit per point of AC shows the decline in AC value as AC rises.

I'm not seeing what you're seeng from the available data. I'm also not seeing it in practice, not in a way that really bears any significant meaning.

Originally Posted by xanathol

I am not talking about 'the' softcap - I am talking about this effective cap where an additional point of AC is meaningless, and no one I know anywhere has AC so high that this is observed in OoW or GoD.

Of course not. We won't come anywhere near that point (if ever erm doubtful) while this content is reasonably current. If we did It wouldn't matter, well we would have some serous development issues. I don't see how you would make this cap meaningful to us or this discussion except as a tool to better understand the mechanics used by soe. It has little do do with quantifying the value of ac vs hp for tanking on raids.

Originally Posted by xanathol

If you want proof of the diminishing returns Kavhok and I have stated exist, run parses against a mediocre mob - say velious or luclin era, perhaps PoP - and calculate the average hit per point of AC, as I did for these numbers here.

Kavhok, made perfect sense to me, I have no problems with what he said and am glad for his contrbution. On the other hand i don't feel you are saying the same thing. I'm not really an AC or a HP guy. I always liked AC from a kind of silly want to be the big mitigator kind of standpoint and still do, but, for real advantage raid tanking it's not so cut and dried hehe. Not that it matters much as tanks gearing up on coa and tacvi style gear prolly won't go too far wrong regardless if they throw out ac augs for hp or not.

PS. Of course the type of healing available is relevant. My wife is a shammy and heals mostly with HOT's and a few fast heals. Mitigation is very desirable when I play with her healing. Being CH'd and spammed on a raid has different emphasis. Tho, I think the argument is mostly pointless beyond benchracing given a little common sense. The clerics being attentive is far more important than some hp/ac juggling.

[Xanathol]

The point, as it pertained the post it was originally to address, has long been lost due to having to restate the obvious over & over again to address subsequent post that are totally out in left field. If you want proof that AC effectiveness is not linear, see the numbers from the parse that I in posted in #153(?):

383.98 / 1608 = 0.23879 average damage per AC
346.05 / 2400 = 0.14419 average damage per AC
308.81 / 3297 = 0.09366 average damage per AC

Done. This is giving me a headache.

[Yoda]

of course the AC effectiveness to the DPS isn't linear !
since AC doesn't affect avoidence, at infinite AC you would get hit for min hit only and does your DPS received wouldn't be at zero.
You should separate the DPS into the part from DB and the part from DI and then do the calcul dps_from_DI / AC to look at its progression (linear ?, etc...)

[Raaj]

Last post was on this subject was in August, moving this to the library.