Attack Rating is Affecting weapons past MBS atm

[Raggy]

The problem probably ties in with the current problem with AR and MBS on test.

What are the bets there is another secret nerf table?

[srompu]

Approaching 100%.
There's no relation between MBS and anything I can find.

Scum, I'll clean up a spreadsheet so it's readable and post. Give me a few.

[srompu]

http://www.megafileupload.com/6y8/Public_version.xlsx

Sheet 1 is where I verify that AAO and Rifle are combined. I cover similar ranges, and get identical results.

Sheet 2 is for the SRPS. Sheet 3 is the Gan'Kar Rifle, and Sheet 4 is Rock's data from earlier in this thread.

There are various data tables from various runs aggregated into a master table with columns "AAO+rifle" which is what it says, true_min, which is min damage dealt minus +damage, and true crit, which is similar but a crit hit. I get both damages from hitting the Punching Bug on testlive, which caps against ACs.

The overall behaviour may be seen in a master graph near the top of the page. There's a list of graphs, one min, one crit, paired together spanning the damage range. You can try changing each range a little to verify that points outside that range follow a different trend.
By comparing the slopes of min and crit to the weapon templates, respectively, you can see that they are proportional. The intercepts have similar proportionality, which suggests that the templates for lower AR are also all proportional.

By comparing the slopes of lines for different damage ranges, you'll find the n multipliers: 1, 2, 2/3, etc.

Since you know that the individual weapon data is proportional (multiplied by) the weapon damage template, you can divide that out. Similarly, the multiplier n can be divided out. This leaves you with one more thing, which I've called "s" so far, that's the missing piece of the conversion. This term is different for each weapon.

[Esssch]

What are the bets there is another secret nerf table?

https://www.youtube.com/watch?v=gENVB6tjq_M

[Raggy]

https://www.youtube.com/watch?v=gENVB6tjq_M

https://www.youtube.com/watch?v=VdddZIcyykM

http://www.megafileupload.com/6y8/Public_version.xlsx

Would it not have been easier to use a locked Google Docs?

[Scvm]

Raggy: The question was about profession-based hard-caps on attack rating contributing to the damage calculation, something entirely separate and unrelated to MBS, but due to the existence of MBS has gone almost completely unremarked upon until the recent (apparent) disabling of MBS on the test server. I can personally attest that last I knew MA's fist damage scaled predictably with AR from 3000 up to at least 4200 or 4300 while Adv AR stopped contributing to damage past 3501. I wonder if that is still the case on live or not.

Edit: silly me, there's a page 2.

thanks for sharing that, srompu

Another edit: has anybody compared any of this with current behavior on live, perhaps with weapons that don't have MBS?

[Scvm]

What immediately stuck out to me was that srompu's SPS numbers were right in line with what I had thought agents were supposed to do (referenced referenced here, would have been with XoA) and neither agree with McKnuckle's recent post (here). McKnuckle, if you're reading this, did you come up with that 5.82 on live?

Srompu's angst/gan'kar numbers really are weird and they baffle me.

[McKnuckleSamwich]

What immediately stuck out to me was that srompu's SPS numbers were right in line with what I had thought agents were supposed to do (referenced referenced here, would have been with XoA) and neither agree with McKnuckle's recent post (here). McKnuckle, if you're reading this, did you come up with that 5.82 on live?

Srompu's angst/gan'kar numbers really are weird and they baffle me.

Hi Scum

The numbers I came up with are based on "TESTLIVE (18.7.0.23) perks/procs." as mentioned in the first line of the post.

However, the basis of the prof multiplier was based on the crit damage gain (i.e. critical hit - normal hit = cirt damage gain) at or above 2750 AR which (at the time of writing) was assumed to provide maximum damage for that weapon.

That assumption has been disproved but broadly speaking the prof specific (or weapon specific on live) multiplier is limited by the logic statement: min(AR, MBS), meaning, if you can find a weapon with no MBS, the prof/weapon multiplier should (technically) keep going up until AR is capped (via OSB/MR/etc.).

The obvious drawback is that traditionally weapons with no known MBS do not have great damage ranges, and so generally speaking people will tend to use weapons with bigger damage ranges which usually also have MBS caps.

I didn't actually know that the post 1k AR multiplier was weapon limited, but in hindsight, it's always been the case. A 4.8 post 1k AR multiplier (5.8 total multiplier) is, broadly speaking, "average" for a 2750ish MBS weapon, but some profs cannot get near that multiplier, regardless of the weapon MBS (i.e. MP-Based on Justin's tests, and likely NT and doc as well-I'm speculating here).

Hope that helps.

[McKnuckleSamwich]

The 2037 crit damage came from observations of Rockdizzle's tests on Testlive.

Specifically the calculation here: http://forums.anarchy-online.com/sho...=1#post6206943

Which I did just before the new thread referenced in your post.

Which was corroborated in a couple posts of rockdizzle's in the same thread.

[lainbr]

This leaves you with one more thing, which I've called "s" so far, that's the missing piece of the conversion. This term is different for each weapon.

Just joining to ask: isnt the 's' the profession template especific multiplier? Or it is 'n' and I got everything wrong? :c

[Scvm]

Hi Scum

The numbers I came up with are based on "TESTLIVE (18.7.0.23) perks/procs." as mentioned in the first line of the post.

Well, I totally should have seen that.

I didn't actually know that the post 1k AR multiplier was weapon limited, but in hindsight, it's always been the case. A 4.8 post 1k AR multiplier (5.8 total multiplier) is, broadly speaking, "average" for a 2750ish MBS weapon, but some profs cannot get near that multiplier, regardless of the weapon MBS (i.e. MP-Based on Justin's tests, and likely NT and doc as well-I'm speculating here).

Hope that helps.

My posts in this thread have maintained a position of denial on any weapon-specific multiplier nonsense, and the only evidence for such a thing that I have seen is the XoA tab in srompu's spreadsheet—evidence which is contradicted by McKnuckle's and Rockdizzle's independent 18.7.0.23 test results with the same weapon, past results on live, and, presumably, current results on live. Everything I just listed shows the same multipliers at a given AR as srompu's SRPS numbers.

[srompu]

Ok, I found "the glitch." I forgot the damage template reversion for testlive .23 for the XoA.
It's 308(308) instead of 350(350).

Just joining to ask: isnt the 's' the profession template especific multiplier? Or it is 'n' and I got everything wrong? :c

Probably.
The n exists because there are ranges of AR skill that seem to act differently. In my tests, the damage scaling between 3000-3300 is twice that between 2550 and 3000, for instance.