TML/SML building bug

Moderator: keyser

TML/SML building bug

Postby Eukanuba » 23 Apr 2013, 16:49

Production of tactical or strategic missiles effectively stops if your mass income is less than 100%. Is it accepted that this is how it should behave, or is there some other reason nobody has tackled this bug?
User avatar
Eukanuba
Priest
 
Posts: 301
Joined: 20 Apr 2012, 19:59
Location: Wales, UK
Has liked: 1 time
Been liked: 0 time
FAF User Name: Eukanuba

Re: TML/SML building bug

Postby BRNKoINSANITY » 24 Apr 2013, 01:40

This behavior is not a bug, it works the same as an build in progress does when you don't have sufficient mass to complete it without stalling. However, the nuke is set up so that you can't spam assistance to get a ton of nukes out of one building.

1. A nuke costs 12,000 mass. No small investment, so stalling mass hurts your build time significantly.

2. The nuke silo has over 1k buildpower. That is the equivalent of 200 t1 engies, so you would have to invest more in unit cap to doubling the build speed than it is worth.

3. HUGE time investment. 5 minute build, which works out to 4500 en and 40 mass a second.

These three things work together to balance the nuke launcher so that nuke spam does not win every game. However, what it means to answer your question is that being just a few mass per second short of full eco has a huge impact on build time. Since it is only drawing 40 mass per second on a 5 minute build time, if we imagine the nuke in a perfectly balanced isolated eco, being 8 mass a second short adds a solid minute to the build time. If you are 35 mass short, well, you are in for a hell of a long wait.

The best solution to this is to build multiple silos and when you are ready to start nuking balance your eco. A lot of time I do this from a navy slot. Build strat subs or the Sera battleship, and if you can win navy consider nukes advanced shore bombardment. Since you won navy, it is easy to balance eco and only produce enough to maintain your navy against air, and just let your nukes build.
BRNKoINSANITY
Evaluator
 
Posts: 951
Joined: 09 Oct 2012, 01:14
Has liked: 43 times
Been liked: 207 times
FAF User Name: BRNKoINSANITY

Re: TML/SML building bug

Postby Eukanuba » 24 Apr 2013, 10:26

I don't agree with that.

Any other building project will be prorated. Say my mass is +80 but -100, I will be at 75% mass income and therefore all building projects will take a third as long again compared to if I was at 100% mass (in this case +100 / -100). This is true for any buildings, experimentals, units or upgrades.

But it is not true for missile launchers. Ignore the SML for now and look at the TML: the database doesn't have the info but from memory I think it uses either -6 or -10 mass continually while it is making missiles, but the exact figures don't matter. If you are running at 90% mass efficiency the TML will as good as stop production, whilst still draining its full mass amount.

Look at the replay I posted on this page, specifically the TMLs I build near the middle of the map. Notice how they aren't making missiles at all until I pause all my factories. This can't be intentional behaviour.
User avatar
Eukanuba
Priest
 
Posts: 301
Joined: 20 Apr 2012, 19:59
Location: Wales, UK
Has liked: 1 time
Been liked: 0 time
FAF User Name: Eukanuba

Re: TML/SML building bug

Postby BRNKoINSANITY » 25 Apr 2013, 01:39

It is prorated the same way, but as I was showing you with the mass costs and build time, if you prorate by percentage it affects the TML and SML far more to be short mass than it does building another unit. The vast majority of the cost of either one of these missiles is energy, so you have a huge energy draw and a small mass draw (like 6m, ~125 en for tac if I remember right). Taking a percentage, if you are 5 energy short, no biggie, but 5 mass short and you multiply your build time by 6 to get one missile. You also have to consider your distribution of buildpower. If you have engy 1 drawing 6 mass building something, and engy 2 drawing 2 mass, and you are -2 eco, engy one will have higher priority on available mass because it is building something that costs more mass for the same build power.

^^^ this is not 100% a correct explanation if you get uber technical, but I am trying to make it simple.

Also, just because something says -10 mass does not mean it is using 10 mass. If you are -10% on eco (-5 and you are bringing in 50), the description will still display the maximum draw as it always does, but the build will only be drawing the available mass, so it would be drawing 9 mass (-10%). This is taking for granted that you are not assisting with any addition build power. Everything you are building across the map will be drawing 90% of it's max potential.
BRNKoINSANITY
Evaluator
 
Posts: 951
Joined: 09 Oct 2012, 01:14
Has liked: 43 times
Been liked: 207 times
FAF User Name: BRNKoINSANITY

Re: TML/SML building bug

Postby Eukanuba » 25 Apr 2013, 14:33

It looks like there's a fair chance I'm being stupid here, but I still don't get what you're saying.

If my TML costs -6 mass to build a missile, then when I am running at 80% mass income it should still be getting ~5 mass and therefore building at 80% speed. The only way that the TML *should* be getting only 1 mass as in your example (where you say that the TML is “5 short”) is if my overall mass income is running at ~17%.
User avatar
Eukanuba
Priest
 
Posts: 301
Joined: 20 Apr 2012, 19:59
Location: Wales, UK
Has liked: 1 time
Been liked: 0 time
FAF User Name: Eukanuba

Re: TML/SML building bug

Postby BRNKoINSANITY » 26 Apr 2013, 00:50

Ok one more shot at explaining this and then someone will have to jump in with a graph :lol:

Referring to above example with 50 mass eco. ^^^ We will disregard energy as 99% of the time you are +power.

You only have one t3 factory building percies with 24 t1 engies assisting. That totals to 49 mass consumed on 180 build power. This would be the equivalent of a t2 fac running along with a few t1 facs and something else building structures, pretty much a normal game situation. With the one mass expense though, we have consolidated it down to one easy to track number.

Then, you build a TML and let it load. Remember that before, you were +1 (49/50). Now you are -5, or 10% overdrawn. However, everything does not run at 90% efficiency. You have to remember that the TML only has 80 build power, whereas the unit building has 180 build power working on it, over double what the TML has. So in reality, the TML is only able to draw a tiny tiny amount of mass away from the mass intensive factory build. So, it takes a long time to load.

I don't think I can explain any clearer than this.... if you still don't get it go to the unit database with a calculator and start adding stuff together till you figure out how things are prioritized.

As to whether or not this is buggy behavior, the TML has to be limited in it's build power.... if it had so much build power that it could still build at a good rate in a -mass situation, when your eco was good it would be BLAZING, like a -25 draw and building a missile every few seconds. That would totally destroy your economy if you built more than one TML.
BRNKoINSANITY
Evaluator
 
Posts: 951
Joined: 09 Oct 2012, 01:14
Has liked: 43 times
Been liked: 207 times
FAF User Name: BRNKoINSANITY

Re: TML/SML building bug

Postby Eukanuba » 26 Apr 2013, 08:55

Well, it looks like up until this point I have never understood the more subtle ways build power affects the game. The economy is significantly more complex than I had realised up until this point. This will probably be the reason that I sometimes randomly crash my eco and can't figure out where the drain is, and also most likely why I'm shit at team games that involve heavy eco.

Well silly me.
User avatar
Eukanuba
Priest
 
Posts: 301
Joined: 20 Apr 2012, 19:59
Location: Wales, UK
Has liked: 1 time
Been liked: 0 time
FAF User Name: Eukanuba

Re: TML/SML building bug

Postby CrazedChariot » 26 Apr 2013, 13:32

To be honest, I dont follow this logic either. Maybe I am missing something.

BP is already reflected in the actual drainage on an economy. This is what actually calculates the drainage, as opposed to raw total mass cost. I don't believe you need to account for it twice.

If a tml uses -6 mass, and your eco is slightly negative. In theory, it should still build at relatively the same speed as before.

I agree with the original poster, something funny seems to happen where the tml seems to take forever to load.
User avatar
CrazedChariot
Contributor
 
Posts: 141
Joined: 02 Sep 2011, 13:13
Has liked: 0 time
Been liked: 32 times
FAF User Name: CrazedChariot

Re: TML/SML building bug

Postby Blackheart » 26 Apr 2013, 14:41

After doing some investigations, it seems bugged for me. The production stops sometimes completely, so no progress at the missile itself is made. Also the mass drain is sometimes fluctuating. Something is definetly wrong here. Attached replay for demonstration. At min 7 or so it gets interesting,since cheat window isnt working anymore for me :x Note: Only mass stall is tested, i dont know if it gets the same results with power stall.
858278-BC_Blackheart.fafreplay
(21.72 KiB) Downloaded 103 times
User avatar
Blackheart
Priest
 
Posts: 376
Joined: 04 Jul 2012, 13:26
Has liked: 294 times
Been liked: 299 times
FAF User Name: various

Re: TML/SML building bug

Postby BRNKoINSANITY » 27 Apr 2013, 01:41

Sigh.... I just spent an hour doing the math and adding everything up with a sandbox game to back up the data and the freaking forum lost it.... so here is the condensed version.

Sandbox -> 50 mass income, 50 outgoing without fluctuations through a series of various tier engies building queues without moving. so net 0 with depleted storage. Then I added the TML so should have been net -6, but it stayed at 0 and built the tac. Here is the actual game data.

Perfect eco had 40 buildpower on unit build for a total of 50 mass. Unassisted tac launcher in perfect eco builds in 20 seconds with 80 build power. With the - eco in our sandbox the build times were as follows. each increment is a t1 engy assist.
80 bp -> 3 min
85 bp -> 2:10
90 bp -> 1:40
95 bp -> 1:30
100 bp -> 1:22 etc.

This is slightly flawed as with each addition of an engy you increase the amount you are -mass and therefore benefit less, but you can see the curve. As for the math, if your buildpower is on a 1/1 ratio with the tac trying to pull 6 mass away from the 50 already in use, you end up with 49.89 to unit and .11 to tac (1/1 loss as a percentage). Since the tac is double buildpower, or 1/2 ratio, then you have something like 49.34 to .66, and if you multiply .66 by 180 seconds for the top example, you almost bullseye ~120, the cost of a missile. If I had the time to add up all the situations by ratio of the build power I would bet money that they add up correctly the same way.

TL;DR -> this is not a bug, the math works.
BRNKoINSANITY
Evaluator
 
Posts: 951
Joined: 09 Oct 2012, 01:14
Has liked: 43 times
Been liked: 207 times
FAF User Name: BRNKoINSANITY

Next

Return to FAF Suggestions

Who is online

Users browsing this forum: No registered users and 1 guest