Repairing R&D Factories

[Chris21wen]

Most know that repairing R&D factories takes longer the further away an aircraft's due date is (See the attached graph.), it is random and only fully repaired factory generate research points.

I wanted to know by how much so over the past year game time, 391 days (31 Dec 42) I've been recording the turns when R&D factories repair using Tracker. I stress this is not the research time but the repair time of the factories.

Aircraft types are irrelevant, only due date matters and I was only interested in factories where I intended to build the optimum factory size of 30. During this testing all bases had sufficient supply (>10k) at all times.

Now I've always believed that it did not matter when you increased the R&D factories to 30, total repair time would be exactly the same whether you increased it in Dec 41 or Dec 43, given the same random (not possible I know) die rolls. This does not appear to be the case. The earlier you increase the earlier your fully repaired R&D factory will be.

I also believed it was pointless to expand R&D factories to 30 until it was within 12-18 months of it's due date. Again not correct as the larger the R&D factory is the faster it appears to repair.

Even if this is incorrect it is still wise to do it as earlier as possible due to the point above. Plus, if a 5 level R&D factories full repairs to 5 over a month you'll only get a fraction of the research points a 30 level factory would generate. If you subsequently increase the factory size it will stop generating research until full repaired again.

The conclusion I've come when it comes to repairing and expanding R&D factories is:

The earlier you expand the quicker the factory reaches fully repaired.

[GetAssista]

The formula I came up with after testing is that daily probability to get +1 repair on a 30 size factory is approximately 1/(T-4/15) where T is number of calendar months left until the arrival month of the airframe. When T=1 probability ==1, factories repair by +1 every day when the frame is due next month. There is absolutely no point having other sizes on R&D factories so I only tested size 30

[Bella]

The formula I came up with after testing is that daily probability to get +1 repair on a 30 size factory is approximately 1/(T-4/15) where T is number of calendar months left until the arrival month of the airframe. When T=1 probability ==1, factories repair by +1 every day when the frame is due next month. There is absolutely no point having other sizes on R&D factories so I only tested size 30

I concur.

[PaxMondo]

The formula I came up with after testing is that daily probability to get +1 repair on a 30 size factory is approximately 1/(T-4/15) where T is number of calendar months left until the arrival month of the airframe. When T=1 probability ==1, factories repair by +1 every day when the frame is due next month. There is absolutely no point having other sizes on R&D factories so I only tested size 30

I concur.

This agrees generally with previous work done on this. You can search the forum back 7 or so years ago. It was delved into in great detail.

The one piece you are missing is the random nature of the repair ... and there is a great deal of variability. We also know that the randomness is associated with each factory. Further, the random factor changes monthly. Meaning, a factory might repair more readily one month, but then the next month it will slow down considerably.

Hence, if you put 12 factories on Ki-84a from the start, you can expect to have it in early '43. The Ki-84a in early '43 is HUGE for the IJ. Completely negates the arrival of the P-47. Put +16 RnD factories on the A7M and you can get it in very late '43 ... again, a huge thing for the IJ as it robs the F6F of its glory. But to do these things means foregoing a lot of other aircraft ... strategic decisions.