I discovered the following with in game testing. Can someone else confirm this? It seems too crazy to be true. I'm probably wrong, I hope I am. T2 of any factory builds T1 units just as fast as 3 T1 factories. T2 demands 4x the metal as a single T1 factory. Also discovered through in game testing, the build power is also stated in the UI, and can be seen with the Unit Database mod via PAMM. Sooo.... T2 produces units 3x faster, but costs 4x as much metal... in other words, 15 metal per second disappear into thin air when you use T2 factories. It would be more metal efficient and just as energy efficient to just use T1 factories - and that's not even going into the greater metal and real estate costs of T2.
Odd. I did some retests in sandbox, with even weirder results. T1 Factory test. Step 1: Drain all metal storage by building something. Have no mexes. Step 2: Build doxes with 3 T1 factories for 53 seconds. You will get 24 of them. Metal will remain drained to 0 (except the brief pause when the factory stops building). T2 Factory test. Step 1: Drain all metal storage by building something. Have no mexes. Step 2: Build doxes with 1 T2 factory for 53 seconds. you will get 21 of them. Metal will remain drained to 0. TL;DR 20 metal and 2k energy per second produce units more effeciently from T1 factories - T2 is inefficient. I keep my original conclusion: T2 factories waste metal, and are useless unless building T2 units.
1.5k metal builds into 32 doxes in the T2 factory. 1.5k metal builds into 33 doxes in the T1 factory. They should build into 33 doxes. However for some strange reason the dox in the T2 factory costs 48 metal. How did I test? Metall storage of 1.5k, 2 fabbers build something to even out the 20 metal gain of the com. When the metal storage is below 1000 you can exactly see how much metal a dox costs. This difference really is strange. This explains your second observation. Your first observation is based on roll-off time. A dox in the T2 factory costs 3 metal more, however as you pointed out, the T2 factory is slow. My test results: 30 dox in 3 T1 factories: 1min 30 dox in 1 T2 factory: 1min58s So for the dox it is only 33% faster (1 T1 factory would take 3min)! But the metal isn't wasted (except for the 3 metal difference). A dox takes 3s to build in a T1 factory and has a roll-off time of 3s (roughly). So it takes 6s for one dox, meaning that 3 factories build a dox every 2s, which leads to 30 dox in 1min (this matches my test result). In the T2 factory a dox takes something under a second to build. I don't have an exact clock but PA database says 0.8s (48metal/60metal/s; It should of course be 0.75 but maybe the game rounds the time to have 0.1 intervals at most, this would explain why the metal cost is higher for the dox in this case: Time has to adjust to the 0.1 scale, so the metal cost is adjusted too). The roll-off time is 3s (roughly) like for the T1 factory. So it builds 1 dox every 3.8 seconds, which is 30 dox in 114s (1min54s); which is almost my test result . As you see the building time is purely because of the roll-off time. You save some of the real building time in the T2 factory (0.8s compared to 3s for the dox (which is almost 4 times faster)) but you don't save any roll-off time. This explains why it isn't just 4 times faster. This of course is valid for every unit in the game. I expect all roll-off times for T1 and T2 are the same, so you have the same effect (with different numbers for how much slower T2 is than you expect, because the ratio of building time to roll-off time is different for every unit with different metal cost) for the other factories too. Hope this helps! Good catch with the different metal consumption! Here is more background-data on roll-off times and real costs of units: https://forums.uberent.com/threads/basic-data-for-build-order-optimization.67451/
Yes, the game operates in ticks, which are normally around 0.1s. As far as I can tell nanolathes are either on or off, spending their full metal. Dox cost 45 and adv. factory runs 60, so it wastes. You my notice that unit costs are even multiples of the typical builder/factory - e.g. dox is 45 instead of 50 because of t1 factory runs at 15. There is a similar issue with reclaim. Trees are actually worth so little it basically rounds down to zero; reclaim value is the tick the fabber starts running plus the tick it stops running (I'm guessing) or 20% of the fabber build power.
Thanks for the sanity check and link, Knub23. I hadn't thought about the roll off times. I guess this confirms that T1 factories are better than T2 when producing T1 units. That said (though I haven't tested it yet) T2 units are invariably more powerful and cost effective than T1 units - with the possible exception of T2 AA being nearly useless outside interplanetary travel.
Unnecessary. Just delete the com. No but seriously, the commander has a power button, and I'm fairly sure the power button turns the yields off.
