Arsenal DPS spreadsheet

[Krix]

(02-24-2012 03:01 AM)bural Wrote:  IOn the other hand I have a hunch Terminal Velocity may be overvalued in the maple sheet. Please don't misunderstand me and take this as a fact though.

Actually, Terminal Velocity (or armor-piercing cell) is not actually used in the formula. We are always going to use it anyway so it is simply ignored.

[bural]

Hmm, obviously there's the very possible chance of me misreading something.
Also, there's the ever returning problem with regen and ability cost. While I'm unable to dig up the official post I thinking of, I believe it was stated BH and Trooper resource systems are identical meaning eg. Demo Rond should be 1.92 and Cell Charger 0.96. This doesn't mean they in fact are of course, but it's possible the UI suffers from rounding errors.

What I was referring to is the code just above equation tag (2.1) where critregen is defined utilizing critspersec. Definitely seems to me as though it's part of equation 2.1. As I understand it, you're calculating Cell Charger similarly to my Terminal Velocity in my heat building cycle. The problem with the approach of simply calculating time between crits and adding the internal cooldown is that crits aren't distributed uniformly. It's an average, I know, but the 3 successive hits from Full Auto will skew the image.

Ultimately, if your guildmate is interested it would be extremely interesting to compare Terminal Velocity proc frequency. In that regard I should note the frequency I display in the HUD is purely for the heat building cycle. I'll update it for the next release.

[Simo]

Hey, this spreadsheet looks really great. But I couldnt find any info about the meaning of the shortcuts. Like EP, "2 OP1 15% TM crit" etc.

So if you are able to make it more user-friendly Thanks

[Krix]

(02-25-2012 08:10 AM)bural Wrote:  Hmm, obviously there's the very possible chance of me misreading something.
Also, there's the ever returning problem with regen and ability cost. While I'm unable to dig up the official post I thinking of, I believe it was stated BH and Trooper resource systems are identical. This doesn't mean they in fact are of course, but it's possible the UI suffers from rounding errors.
What I was referring to is the code just above equation tag (2.1) where critregen is defined utilizing critspersec. Definitely seems to me as though it's part of equation 2.1.

It's a proven fact that it is not "identical". Troopers got a smaller amount of ammo at their max regen then bounty hunters got heat. So bounty hunters got a slight advantage there.

I will get back to you on critregen. But aslong as it's there I am quite sure it should be right.

[bural]

Can you point me in the direction of a link to that conclusion? I don't mean to sound as though I don't trust you, I'm just very interested if someone has made a definite conclusion. Generally, as a UI rounding error would affect both costs and regen it's pretty tricky to test, especially when latency adds additional uncertainty.
Besides, baseregen is set to 0.6ammo=5heat/100 (i.e. identical max regen) in the maple sheet so it's not a discrepancy we're currently accounting for.

[Qed]

Hey bural, I've been trying to poke at your sheet today to turn off the vent heat and terminal velocity, but setting the TV HPS and VH Vent Heat rows to 0 still seems to leave me with a rotation that uses >0.6 ammo per second. Have I missed something?

(also, it seems like terminal velocity and cell recharger list slightly different values:

Terminal Velocity restores 8 heat = 8/100 = .08 of the bar
Cell Charger restores 1 ammo = 1/12 = .083333 of the bar

I'd believe that the underlying numbers to work out, but I'm curious as to which number is actually correct.

[Krix]

(02-25-2012 09:16 AM)bural Wrote:  Can you point me in the direction of a link to that conclusion? I don't mean to sound as though I don't trust you, I'm just very interested if someone has made a definite conclusion. Generally, as a UI rounding error would affect both costs and regen it's pretty tricky to test, especially when latency adds additional uncertainty.
Besides, baseregen is set to 0.6ammo=5heat/100 (i.e. identical max regen) in the maple sheet so it's not a discrepancy we're currently accounting for.

Can't remember where I found that out...but it should be easily proved by testing. (or simple math).

1:
If I remember right bounty hunters stay at maximum heat loss until they get to 40%(40heat).
If it would work the same for troopers that means we should be at full regen until we are at 7,2 ammo. But we can never actually use those last 0,8 ammo so they are in fact lost. That means we only got full ammo regen until we go to 8ammo.

That means we can only use 33% of our ammo before we stop regaining ammo at our maximum level.

Also, (if I remember right) when an ability cost 1ammo for us it costs 8heat for you.
We got 12ammo. 12x8=96. So you got 4 more heat then you should if this were true mirror classes. Or, you got 12 & 1/2 ammo rounds.

So, in conclusion.
Bounty hunters got ~4% more resources then Troopers do.
Bounty hunters stay at max regen 7% longer then troopers do.

The issue that causes this is simply the big difference between a big number (100) and a markedly smaller number (12). You simply can't keep it the same. However you do it one class will get a slightly better/worse resource regen.

[Qed]

(02-25-2012 10:59 AM)Krix Wrote:  Can't remember where I found that out...but it should be easily proved by testing. (or simple math).

1:
If I remember right bounty hunters stay at maximum heat loss until they get to 40%(40heat).
If it would work the same for troopers that means we should be at full regen until we are at 7,2 ammo. But we can never actually use those last 0,8 ammo so they are in fact lost. That means we only got full ammo regen until we go to 8ammo.

That means we can only use 33% of our ammo before we stop regaining ammo at our maximum level.

Also, (if I remember right) when an ability cost 1ammo for us it costs 8heat for you.
We got 12ammo. 12x8=96. So you got 4 more heat then you should if this were true mirror classes. Or, you got 12 & 1/2 ammo rounds.

So, in conclusion.
Bounty hunters got ~4% more resources then Troopers do.
Bounty hunters stay at max regen 7% longer then troopers do.

The issue that causes this is simply the big difference between a big number (100) and a markedly smaller number (12). You simply can't keep it the same. However you do it one class will get a slightly better/worse resource regen.

I swear i've seen this argued both ways. A citation would be very useful to decide this one way or the other, since we clearly have _some_ interaction with fractional ammo (eg, the tooltip say .6 ammo/sec at our max regen), though this may be a difficult issue to actually find evidence for without some very OCD testing or a developer quote.

[Krix]

(02-25-2012 12:23 PM)Qed Wrote:  

(02-25-2012 10:59 AM)Krix Wrote:  Can't remember where I found that out...but it should be easily proved by testing. (or simple math).

1:
If I remember right bounty hunters stay at maximum heat loss until they get to 40%(40heat).
If it would work the same for troopers that means we should be at full regen until we are at 7,2 ammo. But we can never actually use those last 0,8 ammo so they are in fact lost. That means we only got full ammo regen until we go to 8ammo.

That means we can only use 33% of our ammo before we stop regaining ammo at our maximum level.

Also, (if I remember right) when an ability cost 1ammo for us it costs 8heat for you.
We got 12ammo. 12x8=96. So you got 4 more heat then you should if this were true mirror classes. Or, you got 12 & 1/2 ammo rounds.

So, in conclusion.
Bounty hunters got ~4% more resources then Troopers do.
Bounty hunters stay at max regen 7% longer then troopers do.

The issue that causes this is simply the big difference between a big number (100) and a markedly smaller number (12). You simply can't keep it the same. However you do it one class will get a slightly better/worse resource regen.

I swear i've seen this argued both ways. A citation would be very useful to decide this one way or the other, since we clearly have _some_ interaction with fractional ammo (eg, the tooltip say .6 ammo/sec at our max regen), though this may be a difficult issue to actually find evidence for without some very OCD testing or a developer quote.

Well, as I said this is quite easy to prove. It's not like I have to test it in the game or anything like that. I am not particulary bothered by this problem but some people obviously are, most likely because they pvp more then me .

The simplest solution for this would be for bioware to give troopers a 100ammo with bounty hunter prices. It's as simple as that. They should already have all the correct coding available. It's a simple copy and paste job.

[bural]

Indeed, I have no idea why they think having 12 max ammo is more realistic (or w/e) than 100 max ammo.

(02-25-2012 10:59 AM)Krix Wrote:  Can't remember where I found that out...but it should be easily proved by testing. (or simple math).

As always, I might be missing something but I just don't see how.
I see how it's easy to calculate, but that doesn't mean it's representative of actual ingame mechanics. What may seem as a sound logical conclusion may not reflect facts and consequently, empirical data are more or less required. Naturally, these are hard to come by without a combatlog.

Here's how I see it.

We know some discrepancies can be observed simply by looking at the UI. However, we don't know when and how the UI rounds off actual values.
Additionally, you have the problems of testing:
-In order to test regen, you'd need to spend a set amount of resources first so as to observe the regeneration time for X resources.
-In order to test ability cost (resource spenditure) you'd need to to know how many resources you have available within a certain timeframe.
The interaction between the two coupled with latency makes it problematic at best to test.

At one point I suggested completely emptying your resource bar so as to eliminate one uncertainty and know exactly how many resources you could regen from a specific timestamp. However, the assumption that you can infact empty your resource bar completely is equally problematic so it kinda got me nowhere.

Admittedly, the differences arising from this issue is in the 1-2/(0.16*12)~4% modified by the fraction of total damage done by resource spending abilities, so probably somewhere in the 3% area. This goes some way to explain our estimated dps differences, but hardly the majority of it.
Still, it's a fundamental issue that would be nice to eliminate.