Great test, simple and informative. However, just to cover all the bases, I would repeat the same test vs. 2 NPC's, then 3, then 4, etc. just to look for consistency (just in case BFAW behaves differently vs. 1 or multiple targets). Assuming that checks out, then I would investigate the following:

An error in the reporting of hits/misses

1. parse manually and see if the "accuracy bug" shows up.

- If it does, then the error is in the delivery of hit/miss information by the game.

- If it does not, then the error is in the delivery or interpretation of information by the parsing programs.


I would not proceed to speculate until I knew the above outcomes.

I would suggest, and that is a little cheeky, that Borticus try the parser used by players reporting the bugs, simultaneously to the dev debug parsing tools. That way it'll be clearer where/when the alleged bug creeps into the results, and whether the parser used by players is functioning correctly.

If one analysis says all is fine, and another says it's not, this will provide some insights into whether they really disagree or whether they agree but that the testing environment needs to be changed...

Edit: partially ninja'd, but statement stands.

It's probably been mentioned in one form or another, but it doesn't appear that either the Accurate trait or [Acc] weapon modifier were isolated as variables in your test Bort.

I have to be honest here, I'm highly suspicious that there's a reporting error at play here. Whether it's the Combat Log, or the Parser. Sure - my original tests dealt with some extreme figures, but if Accuracy "wasn't being used" then the results would have obviously swung a different direction.

Let's keep the discussion going, but I'd like to respond to a few points:

1) There is no Cryptic Dev combat parser tool.

The reason we don't have one is probably because our Programmers use a version of the client that can show them the step-by-step code used in every Combat event. So if a single event ever needs to be debugged in detail, they use that in combination with a surgical testing scenario, instead of relying upon large sample sizes.

2) "using hit calculations for cannons instead of beams"

There is no such variable, and this is another thing that I'd like to figure out the reason for. It may also be a reporting error of some sort.

3) Test vs multiple foes?

Yes, I did. The above tests were worded misleadingly. I was actually firing at a field of about a dozen NPCs, all with the same +1000 Defense buff, and another field of them without any buff as a Control group.

I'm also going to try throwing a few other variables into the fray, such as Defense added via Engine speed (instead of applying it directly), and Accuracy granted via various means.

4) The variance being posted is not at the "0" mark, so I'll do more testing.

That is to say, my test included even figures for Defense vs. Accuracy. Yours have not. I've heard numbers around the "75 Defense vs. 25 Accuracy" range. So I'll run more tests with variable offset figures, and see if I get the same results as previous.

5) I'm not sure when I'll get more time.

I'm fixing this in between other priorities, so please don't expect instant miracles. And it's just me doing this - I don't have any Programmers helping me debug.

They wouldn't really need to be, for Systems purposes. Internally, they are both powers, and I tested using variants on how those powers are applied, and saw no change in the results.

No, that's not the issue. FAW has hitrate problems even against a single target.

A useful test could be the following:

Give the target +1000 Defense, give yourself a weapon with +1000 Acc (and make sure that this really gets applied the same way a normal weapon would and not at some other stage of the calculations) and then see what the hitrate is.

if I understand it correctly, the tests were only +1000 Defense and then +1000 Accuracy, but not both at the same time.

This was already done. To clarify:

There are three ways in which an Accuracy buff can be applied to a player:
1) Innate power (Trait, Skill or Console bonus - all work the same)
2) Temporary power (Buff)
3) Item Enhancement power (Accuracy mod on weapons)

The original test was done using #2 and #3 methods. Though the difference between #1 and #2 is negligible in terms of the code/data, I'll be performing more tests today to see if there's anything amiss (hah) in that functionality.

As for the Test itself, I believe you misunderstood. I was applying both +Defense and +Accuracy bonuses individually, and then combined, to witness the effects.

I'll compile more data in the next couple of hours to make my case, but this is what I expect to be the outcome:

- There is a difference between beam and cannon accuracy.
- FAW has an accuracy problem when it comes to [Acc] mods.

Maybe you could give us the formula that is used for hitrate calculations? Just in case the old Geko one is wrong.

(As I understand it, only Defense and Accuracy are relevant. So things like range are not being taken into account, correct?)

And also: maybe your dev magic powers are the problem? Maybe you are affecting the calculations in a way that differ from actual weapon mods. Why are you testing with dev +1000 anyway instead of simply equipping Acc3 weapons and compare them to no Acc weapons?

Given the number of times that one power had an inadvertent (and oftentimes extreme) effect on another power (latest examples include the Romulan Boffs and Doffs throwing Cds out of whack for everything) that seems to lend credence to the idea that they need to be isolated during testing.

Maybe it is overkill, but isn't it better to rule something out experimentally than simply take for granted what you think the results are because something different, but related, did something similar?

Do small targets (eg mines/torps) fall under the category 1?

Also, is there a lantency issue potentionally at play? In otherwords FaW at target moving in and out of range and the log/parser report a miss vs something that wasn't targetable?