AR and NR

[Ebag333]

So some folks know I've been doing some research into the way that AR and NR works. I've been quietly watching threads such as these, and see a lot of incorrect/incomplete information.

Anyway, I'm made a lot of steps towards figuring this out (in fact, I suspect I have the basic idea down), but need more data and test cases to confirm and flesh out what I have.

There's been a few folks helping me, but as they say, many hands make light work.

If you're interested in helping, shoot me a PM. If you have theories on how it works, post them up, and we'll see how right you are.

Here is my initial Nano Skill AR calculator. Feel free to use it, but I make no claims as to it's accuracy beyond that the land rate should be accurate +-5%.

http://pvp.aodb.us/ARDEF/nano_calc.html

So here's my first pass at weapon AR calculations.

http://pvp.aodb.us/ARDEF/weapon_calc.html

I'm actually pleasantly surprised at how well it came out. I added up the difference between the test scenarios and the calculated (for -100 agg/def), and ended up with a total of -1.54%.

So if one test was 1% higher than the calculated, and another was 2% lower, the average of the two would be 1% lower. Having an average of 1.54% lower is phenomenal.

[mr_road]

please don't figure this one out, because to misquote douglas adams...

there is a theory that if anyone finds out the why and how of the universe, then it will collapse in on itself and be replaced by something more bizarre and weird.




there is another theory that this has already happened.

[Ebag333]

Aww, and I was just about to post the answer to Nano AR. I had graphs and everything.

Guess I'll just share 'em with Metaing then.

[Corily]

please don't figure this one out, because to misquote douglas adams...

there is a theory that if anyone finds out the why and how of the universe, then it will collapse in on itself and be replaced by something more bizarre and weird.




there is another theory that this has already happened.

Finding out how NR works does not come close to finding out how the universe works lol. Some people get too deep into AO.

Anyway, finding out how NR works is imo very beneficial. It'll help alot on finding a decent setup for anyone who wants to have the best for their character.

[BugsCoT]

Or perhaps it will be a mystery to everyone....

[Corily]

It's still made by humans. We aren't "that" good, yet.

[Ebag333]

Or perhaps it will be a mystery to everyone....

Or not.


(Teaser pics FTW. )

[Obtena]

I understand the graph but can you explain what parameter is dictating the need for the three lines? Is it agg/def setting?

[Technogen]

yes, I had to ask him too when he showed me.

[Ebag333]

I understand the graph but can you explain what parameter is dictating the need for the three lines? Is it agg/def setting?

The three different lines give you an idea of the effect of the agg/def bar on the land rate. It's actually quite interesting how flat that line gets when you're full aggro.

[Ebag333]

Slightly updated graph, has an extra point for the lower end on the -100 Agg/Def.

Do you guys have any idea how hard it is to reach ~20% AR Diff %?


http://img134.imageshack.us/img134/269/nanoar.tif

[srompu]

How did you calculate the Diff AR%?

[Ebag333]

How did you calculate the Diff AR%?

It's on the page that I sent you the link to.

AR	/	(	AR	+	Def	)	 = 	Diff AR %

AR is whatever the total attack rating is. So if a nano is 50% PM and 50% SI, and the user has 1k of each, the total attack rating would be 1k (.5*1k + .5*1k = AR). The def is the same, but for whatever defensive skills the user has.


It's simply a way for me to find a value where 100 vs 200 and 500 vs 1000 are the same, because the chance to land nanos (or hits, if we're talking about weapon skill) is the same.

[Corily]

Do you guys have any idea how hard it is to reach ~20% AR Diff %?

What about perking NR6-7-8? That should give you the last bits of the curve needed to analyse it.

[McKnuckleSamwich]

so what you're saying is:

evenif a doc's inits gimp you severely in attacking, stay full defence and keep your UWOS spammed as much as possible.

The problem for me is the procs.

Seems like a good way to check defences against an MP docs and traders and even fixers... hum, and enforcers, but NT's have constant barrage hello annoying.

Ebag: are you saying this is the same curve for AR vs. evades+AAD?

which is why if you have 3300 evades and youre dealing with somone with 3200 AR, they will hit you *most* of the time?

And it also provides the reason why DOF/ limber makes nearly no difference at higher levels, which is what you'd expect, since it barely effects the ratio, ...

like 3200 AR vs 3300 is 0.97
and 3200 vs 3500 is 0.91

but, at level 60,

with 400 AR vs 500 evades is 0.8
but 400 AR vs 700 evades is 0.57

[Ebag333]

What about perking NR6-7-8? That should give you the last bits of the curve needed to analyse it.

Shush you. I was trying to avoid reperking. Stripping my baby MP was bad enough.

so what you're saying is:

evenif a doc's inits gimp you severely in attacking, stay full defence and keep your UWOS spammed as much as possible.

The problem for me is the procs.

I'm not saying anything. Though if I were to say something, and if I were a doc....I'd probably play full def.

There has been nowhere near enough testing to quantify this, but it seems that between full def and full aggro, there's a progressively worse penalty. In other words going from -100 to -75 is much less painful than -25 to 0, which is much less painful than 25 to 50, which is much less painful than.....

So being a little aggro might not make a big difference. But being a lot aggro would.

Again, that's just my initial impression based on a handful of tests and could be way off base. But you can see how much closer 0 Agg/Def is to -100, than it is to 100.

Seems like a good way to check defences against an MP docs and traders and even fixers... hum, and enforcers, but NT's have constant barrage hello annoying.

Most profs have fairly static offense and defense, so knowing where you stand against the various profs has its advantages.

Ebag: are you saying this is the same curve for AR vs. evades+AAD?

No.

It's very similar, but not the same curve. I suspect that the formula's used are very close to each other.

which is why if you have 3300 evades and youre dealing with somone with 3200 AR, they will hit you *most* of the time?

And it also provides the reason why DOF/ limber makes nearly no difference at higher levels, which is what you'd expect, since it barely effects the ratio, ...

like 3200 AR vs 3300 is 0.97
and 3200 vs 3500 is 0.91

but, at level 60,

with 400 AR vs 500 evades is 0.8
but 400 AR vs 700 evades is 0.57

First off, your AR Diff %'s are off. 3200 AR vs 3300 Def is not .97 (which I'm assuming you mean to be 97%).

3200 / (3200+3300) = 49.23%

A ~49% AR Diff % could give you anywhere from a ~55% to a ~90% land rate, depending on the attacker and defenders Agg/Def setting. That's a difference between weapons and nanos, the attackers Agg/Def setting seems to have zero effect on the land rate for nanos. For weapons, it's a fairly small effect (compared to the defenders Agg/Def setting) but it's there.

But your basic point is correct. At higher settings it is much more difficult to have an effect on the land rate based on a single item. On the flip side, higher levels have many more options open to them, allowing them to change their setups drastically. An example is Wizzytrooper, who was setup very differently than the typical soldier. Most soldies are somewhere around 1800-2000 defense and 3500-3700 AR. Wizzy was around 3000 defense and 3200 AR.

To use those numbers as an example, for attacking (another soldier) Wizzy would have 64% Diff AR % and for defending, they would have 54% Diff AR % against him.

A stereotypical soldier would have (against another average soldier) 66% for attacking (for both sides, since the values are the same).

Is giving up 2% Diff AR % (for attacking) worth gaining 12% Diff AR % when defending?

That's a question I leave up to you.


(I'm sure someone's going to disagree with my example, and there are probably flaws in it. But it's just an example, folks, so don't take it too literally. Take the meaning and method behind it, and figure out your own examples.)

[Ebag333]

srompu encouraged me to test an old theory....that self casting nanos (casting nanos on yourself) was broken. For what it's worth, this is something that I had believed was probably true, though I'd never tested it.

Well, I have 7 test cases so far. While they aren't all direct 1:1 comparison with non-self testing, the ones that are within a few % of AR Diff % are within a few % for land rate, well within the margin for error.

For example, one test had 47.61% for the Diff AR %, with a land rate (non-self) of 47.38%. The self cast was a Diff AR % of 47.00%, with a land rate of (self) 49.85%.

A .61% difference in the Diff AR % (which is actually quite a bit bigger than you'd think) with a difference of 2.47% in the land rate (which is actually quite a bit smaller than you'd think).


So self casting nanos on self = borked? Busted. You get the exact same land rate as landing on anyone else.


And if anyone cares, I'm up to 36,678 lands/resists. Even figuring only 2 seconds to recast the nano (and many tests were done with larger NR check nanos which had long recharges, to adjust the AR Diff % more) you're looking at 73,356 seconds. Or 20 hours, 22 minutes, and 36 seconds.

Ouch.

[Corily]

Well for starters self-casting can't be borked since my testing involved 2 characters and gave the same results you obtained.

If anything, I'd say this is the beginning of the end of the NR-nanoskill struggle. This clearly shows how it works and the theory itself doesn't have any flaws, holes or errors. (Well it curve needs more testing below 20% ofcourse)

[mr_road]

so if i read this right:

blue is full def
65% AR difference or higher gives 100% land rate

so if the attacker has 2k nano skills on a 100% check NR nano and the defender is full def then:

Diff AR % = 100 * AR / (AR + Def)
65 = 100 * 2k / (2k + NR)
2k + NR = 200k/65
NR = 200k/65 - 2k
NR = 1076

if you can't get more NR than 1076, you won't resist anything vs 2k nano skills

nano skills = 1500, NR uselessness limit = 307

[Corily]

nano skills = 1500, NR uselessness limit = 307

You made a mistake in that part of your post. When I calculte it, it comes down to 807NR, wich seems a more accurate number for the NR uselessness limit. (blue line, target full defence) Edit: found your mistake, you didn't change the 2k nanoskills in your formula, whilst they should.

The rest of the calcs seem to be correct indeed.

Also, from tests I ran I saw the true NR uselessness limit was at 62.5%, not 65%.