Forged Alliance Forever

Re: 3629 broadsword damage nerf?

2014-03-07T14:24:19+02:00

Oh I see. Well, yeah, I do! :p

Statistics: Posted by IceDreamer — 07 Mar 2014, 14:24

Re: 3629 broadsword damage nerf?

2014-03-07T09:11:21+02:00

Ice, that whole post of yours is the exact reason why I suggest using the charging period, or whatever you want to call it, rather than the rate of fire. You've even put the 1/RateOfFire in, which is exactly the charging period. In mathematics, that is.

Statistics: Posted by Plasma_Wolf — 07 Mar 2014, 09:11

Re: 3629 broadsword damage nerf?

2014-03-07T05:32:50+02:00

@pip and ice

Thanks for replying to my inquiry.

-Cheers

Statistics: Posted by errorblankfield — 07 Mar 2014, 05:32

Re: 3629 broadsword damage nerf?

2014-03-07T01:53:45+02:00

It's not necessary Plasma. The best way to get rid of the 'Problem' is to just be a little more careful. I highly doubt pip will make this mistake again. Firing cycles CAN be used on SupCom's units, but they make the code, and DPS calculation, exponentially harder, to the point where I try to get rid of as many of them as I can whenever I am modifying units for my own little projects. Having the whole thing done RateOfFire OR RackReloadTime is best, the cases where both are active... Things get silly.

It's for this reason that I never use RateOfFire*Damage to find DPS. Instead. I always calculate 1/RateOfFire, round the result to the nearest 0.1, then do Damage/Result to find DPS. It's an extra step, but it eliminates problems. For numbers between .333 and .499 I can never remember if the engine round down or up, so I often find myself checking the DPS manually to be 110% sure :p

Statistics: Posted by IceDreamer — 07 Mar 2014, 01:53

Re: 3629 broadsword damage nerf?

2014-03-07T00:09:06+02:00

There's only one way to get rid of the DPS rounding problem and that is by ensuring that reload is determined by a fire cycle, rather than a rate of fire. So you're using the period of a periodic function, rather than its frequency. The two are perfectly interchangeable in mathematics, but not in numerical computing, due to the limitations of computers.

Statistics: Posted by Plasma_Wolf — 07 Mar 2014, 00:09

Re: 3629 broadsword damage nerf?

2014-03-06T21:35:30+02:00

IceDreamer wrote:
GPG are definitely aware of it. It was one of them who pointed the rounding out to me in the first place all those years ago!

Ah, ok, good to know, this is important.

Statistics: Posted by pip — 06 Mar 2014, 21:35

Re: 3629 broadsword damage nerf?

2014-03-06T20:54:32+02:00

GPG are definitely aware of it. It was one of them who pointed the rounding out to me in the first place all those years ago!

Statistics: Posted by IceDreamer — 06 Mar 2014, 20:54

Re: 3629 broadsword damage nerf?

2014-03-06T20:11:10+02:00

errorblankfield wrote:
pip wrote:It's not a typo. It's related to the way FA engine computes DPS. In most cases, you get the DPS by multiplying the damage by the rate of fire. For instance : 100 damages * 0.5 rate of fire = 50 DPS. But in some circumstances, due to the fact that the engine computes things every other 0.1 second tick, it happens that DPS is different from that simple multiplication. What I calculated is : Broadsword deals 100 damages * 3 rate of fire = 300. But with the way the engine works, 3 shots are fired every one second, but the engine launches these shots every 0.3 seconds and not 0.333, which ends being faster (0.3+0.3+0.3 = 0.9 compared to 0.333+0.333+0.333 = 0.999). So the real math i should have used is more complex than usual : 100 * 10/3 = 333.

Glad to see that's a simple oversight.

Possibly a stupid question, but why did this just come up now? By that I mean, wouldn't it be insanely simple to change the calculation to == [Damage] * round([ROF], 1).

I suppose this could be the first time it's ever come up, but that makes me wounder about other rounding errors in the data base.

Or was this just done by hand for the patch log. (I was under the impression the unit description was off as well, hence my initial question.

Now that I think of it, it's not the first time it happens. The exact same problem occured for the Cerberus nerf. Basically, all weapons with rate of fire = 3 are super annoying for this reason. I'll be more careful next time I encounter one.

Moreover, I might be wrong but I'm not even sure GPG was aware of it. For instance, if you look at the Mantis, it would make sense for it to have 24 DPS like the others tanks (8*3), but due to its rate of fire of 3, it has a bit more (26.6). I don't mean it's not balanced, just that I suspect several units were maybe designed to have a DPS value that ended up to be slightly higher than initially intended (I would bet it's the case for the Jester for instance, which ends up having 53 DPS, but I believe it was designed to have "only" 3*16 = 48).

Statistics: Posted by pip — 06 Mar 2014, 20:11

Re: 3629 broadsword damage nerf?

2014-03-06T19:14:15+02:00

All the calculations are done by hand, we don't have a formula (Beam weapons mess it up), and it's easier to just do it on the fly anyway. The database does use a formula, which is why it's usually wrong for beam weapons. This happened because of the way the engine rounds things, and pip's relative inexperience with that, leading to a 100% honest mistake. I usually notice when mistakes with DPS calculations have been made, but this time I didn't, and nor did pip, because the main testing was on the new Speed and Cost, not the DPS.

Statistics: Posted by IceDreamer — 06 Mar 2014, 19:14

Re: 3629 broadsword damage nerf?

2014-03-06T17:59:33+02:00

pip wrote:
It's not a typo. It's related to the way FA engine computes DPS. In most cases, you get the DPS by multiplying the damage by the rate of fire. For instance : 100 damages * 0.5 rate of fire = 50 DPS. But in some circumstances, due to the fact that the engine computes things every other 0.1 second tick, it happens that DPS is different from that simple multiplication. What I calculated is : Broadsword deals 100 damages * 3 rate of fire = 300. But with the way the engine works, 3 shots are fired every one second, but the engine launches these shots every 0.3 seconds and not 0.333, which ends being faster (0.3+0.3+0.3 = 0.9 compared to 0.333+0.333+0.333 = 0.999). So the real math i should have used is more complex than usual : 100 * 10/3 = 333.

Glad to see that's a simple oversight.

Possibly a stupid question, but why did this just come up now? By that I mean, wouldn't it be insanely simple to change the calculation to == [Damage] * round([ROF], 1).

I suppose this could be the first time it's ever come up, but that makes me wounder about other rounding errors in the data base.

Or was this just done by hand for the patch log. (I was under the impression the unit description was off as well, hence my initial question.

Statistics: Posted by errorblankfield — 06 Mar 2014, 17:59

Re: 3629 broadsword damage nerf?

2014-03-06T16:50:48+02:00

Yes, it's that. That's been discussed before, I used the word typo more as an example though. This discussion reminded me of the energy typo for building strat subs, it had a zero too many at the end.

Anyway, the balance changes should always be error free, i.e. If an error has been found, it should be corrected and en we discuss whether here is a balance or not.

Statistics: Posted by Plasma_Wolf — 06 Mar 2014, 16:50

Re: 3629 broadsword damage nerf?

2014-03-06T19:38:42+02:00

It's not a typo. It's related to the way FA engine computes DPS. In most cases, you get the DPS by multiplying the damage by the rate of fire. For instance : 100 damages * 0.5 rate of fire = 50 DPS. But in some circumstances, due to the fact that the engine computes things every other 0.1 second tick, it happens that DPS is different from that simple multiplication. What I calculated is : Broadsword deals 100 damages * 3 rate of fire = 300. But with the way the engine works, 3 shots are fired every one second, but the engine launches these shots every 0.3 seconds and not 0.333, which ends being faster (0.3+0.3+0.3 = 0.9 compared to 0.333+0.333+0.333 = 0.999). So the real math i should have used is more complex than usual : 100 * 1/0.3 = 333.

Statistics: Posted by pip — 06 Mar 2014, 16:13

Re: 3629 broadsword damage nerf?

2014-03-05T21:39:23+02:00

A balance change due to a typo should never be accepted. Typos can go unnoticed and unnoticed is something else than "tested and acceptable", or "tested and balanced". Almost everything goes unnoticed to someone who isn't familiar with the way the patching works in FAF, but that does not mean that typos should go through.

Things that have actually been tested and then generate complaints amongst those who haven't tested it, that's another matter entirely (wait for the next patch period before telling all about it).

As for a typo, just put it up here and see what happens. If the people who put the change (Accidentally) in haven't noticed it, they'll change it again. If not right away, remind them during the next patching period and then it's all done. Again: don't complain over and over again (you could've done so during the patching period).

That being said: 25 DPS is not much (though 12.5% points additional cut is), but it's no reason to say: "forget about it, move on". If all gunships had this additional 12.5%. cut by accident, then it's less of a matter, but accidents have to be corrected, plain and simple. During the next patch period, that is.

Statistics: Posted by Plasma_Wolf — 05 Mar 2014, 21:39

Re: 3629 broadsword damage nerf?

2014-03-05T20:04:10+02:00

Gunship DPS changes weren't tested even close to as thoroughly as most of the other changes because it's the speed that's more important. Broadswords should still get the extra DPS because A: It's losing out badly to the Wailer otherwise and B: There's no way in hell an extra 25 DPS will break anything.

Statistics: Posted by IceDreamer — 05 Mar 2014, 20:04

Re: 3629 broadsword damage nerf?

2014-03-05T19:34:49+02:00

^^ Right, if it was tested and found to be a good change then there is no need to change it further. But I think the creator of the post just wants an acknowledgement that there was a mistake (wherever that might be) and move on.

Statistics: Posted by rockoe10 — 05 Mar 2014, 19:34