Research Factories. Repairs and Optimising

[PresterJohn001]

Can anyone shed some light on what determines which and how many research factories are repaired.... and how to optimise research so i can get my uber planes to the front quicker

All i really "know" about this area is that only fully repaired factories contribute to research and that each batch of ?100 points give a "chance" to accelerate by one month. Points appear to accumulate from month to month.

No idea on what the "chance" % actually is.
No idea what determines which factories get repaired.

For optimising i'm thinking expansion should be done slowly (1 or 2 factories at a time)... but is factory repair at a set rate (eg 1 per turn, randomly chosen) or a % rate (eg 1% of factories that need repairing). This could make a difference.

Research factories can be converted like production factories, so if you choose the next model in a series they take no damage - i wish i'd known a year ago

hints tips and suggestions needed! How do i get Frank in May '43? [;)]

[viberpol]

ORIGINAL: PresterJohn
All i really "know" about this area is that only fully repaired factories contribute to research and that each batch of ?100 points give a "chance" to accelerate by one month. Points appear to accumulate from month to month.

Hard to say really... [&:] after about 440 turns in my PBEM I managed to advance miraculously the J2M R&D by 3 months with all 150 ready (fully repaired) R&D factories.. so it's definitely not that a "100 batch" gives an only chance of 1 month... Very random I'd say.

[CapAndGown]

It has been definitively shown that the "chance" that research will advance when 100 plane-equivalents are produced is 100%.

[viberpol]

One important thing though. I feel (don't really know, the opinion is based on my gut feelings) that the closer the "due date" of an airframe being R&D-ed, the higher chance of repairing a single factory point every turn.
This is clearly visible in my game, where factories with planes due to arrive in '45 or '46 repaired only 1 or 2 points (after those 400 turns) while the points of factories with next generation, available in 2-3 months, somehow accelerated very substantially.

But again, you should think if accelerating the airframes available soon is worth the prize in terms of supply costs and factory load.

[pompack]

A few points on research

1) a lot of trials seem to indicate that NO research points are produced unless a factory is 100% repaired.
2) the probability of repair seems to increase substantially as you get closer to the bingo date, conversely it can take a long time to repair one factory to research an a/c two years away
3) some trials seem to indicate that the probablilty of repair increases with the size of the factory, i.e. for a given bingo date a 10(1) factory will repair sooner than a 1(1)

Putting this all together just makes my head hurt [&:]

[tigercub]

OK from what i know a size 30 RD-fatory is the most efficent so more of that size factory(30 the sweet spot) is the way to go but you only get so many RD factorys. bigger is not better...only when there is a lady in the room. [;)]

Tigercub

[Grfin Zeppelin]

ORIGINAL: tigercub


OK from what i know a size 30 RD-fatory is the most efficent so more of that size factory(30 the sweet spot) is the way to go but you only get so many RD factorys. bigger is not better...only when there is a lady in the room. [;)]

Tigercub

To big is much worser than slightly to small ya know [;)]

[RUDOLF]

I pushed Frank forward 14 months!!
 
 

[castor troy]

ORIGINAL: RUDOLF

I pushed Frank forward 14 months!!

that´s what I mean by wayyyy early... wouldn´t be acceptable for a PBEM I guess. And it also shows that the R&D got a serious problem I think. I don´t even want to imagine what it would mean to produce 300 Franks per month starting in 10/42.

[viberpol]

ORIGINAL: RUDOLF
I pushed Frank forward 14 months!!

Unbelievable, me thinks... is this vs. AI or PBEM or a "dry test"? [&:]
How many AARs we have seen up to today with Franks moved for so early date?

ORIGINAL: castor troy
that´s what I mean by wayyyy early... wouldn´t be acceptable for a PBEM I guess. And it also shows that the R&D got a serious problem I think. I don´t even want to imagine what it would mean to produce 300 Franks per month starting in 10/42.

Again. Frank cannot be accelerated using the "R&D switch according to the autoupgrade patch" technique recently discovered, as Ki-84 is a completely new design and have to be researched from scratch.
Take the possibility to expand factories out and none would like to play Japan side.

It also shows that "sir Robin" style of play and letting the Japan side freely use the resources gathered benefits the JFBs economy. Again, this is not about replaying the history, but trying to change it... [;)]

[hunchback77]

Viberpol, he was probably researching the Frank with at least 5 to 7 research factories of around 30. That's how he pushed the Frank forward 14 months.

[tigercub]

i would like to know how many factorys you did use Viberpol and what size? 14months is gamey i would not go there but i would like ta know.

Tigercub

[FatR]

My Principles of Optimizing Research (Not Yet Really Validated By PBEM Practice):

1)Expand research facilities by a lot. It is cheap. 16 is my typical size for pure-research facilities that will not be ever switched to production. The speed of repair seems to be at least somewhat proportional to side. It is possible, that 16 is an overly cautious standard, and factories of size 20-30 will be more effective. I believe it was shown, that there is a limit on research points that can be provided by a single research facility, and it is around 30. Try to switch as few facilities to production as possible and set the rest to researching new models when their old ones are about to enter service.
2)Therefore, devote multiple research facilities to important planes.
3)Do not research planes you do not actually hope to accelerate at all, unless there is no production facilities that can be switched to a new plane. Also, pick planes you intend to use at the beginning of the game, and do not research the rest, of course.
4)Facilities researching an earlier modification of a plane usually (but not always - the big Ki-84a plant you get in the beginning is an exception, and maybe there are others) can be switched to a modification right next to it in the upgrade tree without getting damaged. As facilities repair faster when the availability date is closer, always start research with the earlies available modification in the tree, even if you aim for one of the latter modifications, and switch facilities when they are fully repaired.

EDIT: I can't believe Frank can be pushed forward 14 months. Facilities repair too slowly. I'm in the middle of April of 1942, and no facilites that research planes with availability in 1944 are even close to being fully repaired and actually producing research point.

[viberpol]

ORIGINAL: tigercub

i would like to know how many factorys you did use Viberpol and what size? 14months is gamey i would not go there but i would like ta know.

Tigercub

Hi Tigercub

Me? In a PBEM scen 2 at 2.43 I've got 3 factories doing Frank's R&D with only 50(50) points ready(damaged) and no single month ahead.
It's RUDOLF who said he accelerated it 14 months and I expressed my opinion that this is unbelieveble.

[Grfin Zeppelin]

The Frank research facories wont be repaired until mid/maybe early 43 so it sounds a bit fantastic to accelerate this plane by over a year.I think if one goes bonkers it can be accelerated by about 6-7 months but this would need extreme actions and some repair luck.

[tigercub]

nar viberpol but thanks anyway.

Tigercub

[PaxMondo]

ORIGINAL: Gräfin Zeppelin

The Frank research facories wont be repaired until mid/maybe early 43 so it sounds a bit fantastic to accelerate this plane by over a year.I think if one goes bonkers it can be accelerated by about 6-7 months but this would need extreme actions and some repair luck.

Echos my experiences. I've never had better luck than 6 months accel no matter what I've tried. And I've done some really gamey test cases (12*30 R&D on Frank). So 14 months ... he should hve bought Lotto tickets instead from my perspective!!! [:D][:D]