Japanese R&D

[darbycmcd]

Sorry Michael, with the to and fro of the patch and discussions happening in a couple different places, I am a bit confused about what the actual definative formula is for generating R&D points for Japanese ac. my assumption was that when a factory is completely repaired, it generates points much like it would generate planes, basically the number of devices is expected production per month. now there is some talk that there is actually a 1-3 point per day production, size notwithstanding.... could you outline it definately in one place. thanks!

[pompack]

Excellent idea to create a thread just for R&D

Here is what I consider to be the best treatment from Michael on the subject, as quoted from the patch thread:

ORIGINAL: michaelm

In simple terms
(a) A factory will produce a random R&D between 0 and the number of devices in the factory ie 1 device = 0, 2 devices = 0,1, 3 devices =0,1,2, 4 devices = 0,1,2,3, 10 devices = 0,1,2,3,4,5,6,7,8,9 etc.
(b) The random R&D is then divided by 10. Any random R&D of less than 10 will be '0', otherwise it will be the ten's component of the random R&D.
(c) If there are no damaged devices in factory (this is a given as there must be NO damaged ones present), add '1' to the number from (b).
(d) If the number from (c) exceeds 3, it is capped at '3'.

The number from (d) is added to the a/c development counter - this will be a number from 1 to 3 inclusive.
Once the counter exceeds 100, it moves the available date sooner by a month and resets the development counter.

[darbycmcd]

ah, but this is very fast! it means that every factory (fully repaired) will produce at least 1 point PER DAY. so 10 size 1 factories will give 300 points a month.... is that really the intent?

[darbycmcd]

For example Nymdalar reports this from his game (post 61 http://www.matrixgames.com/forums/tm.asp?m=2825013&mpage=3)

So if my understanding is correct this is giving me the following points toward acceleration
7(0) = 1
55(0) = 3
3(0) = 1
1(0) = 1

for a total of 6 points/day. Which means that I will get an acceleration every 16-17 days which matches up with what is accutually occuring in my game.

and he could actually be more efficient to use 3 size 1 factories rather than the 55. so imagine that just 6 size 1 factories leads to an acceleration every 2 weeks. is that the idea? given the number of R&D factories the japanese has, this seems to be a problem

[michaelm75au]

Here is the revised sequence that was valid before beta p8, and will be again with beta q4. The missing step from earlier postings in in BOLD.

In simple terms
(a.1) Daily number of initial devices from factory is as with a/c production [ (active devices + random(30))/30]. If this number is 0, or there are any damaged devices present, then no R&D will result this turn from this factory. [This agrees with the manual ]
(a.2) A factory will then produce a random R&D between 1 and the number of devices from step (a.1) ie 1 device = 1, 2 devices = 1,2, 3 devices =1,2,3, 4 devices = 1,2,3,4 10 devices = 1,2,3,4,5,6,7,8,9,10 etc.
(b) The random R&D is then divided by 10. Any random R&D of less than 10 will be '0', otherwise it will be the ten's component of the random R&D.
(c) If there are no damaged devices in factory (this is a given as there must be NO damaged ones present), add '1' to the number from (b).
(d) If the number from (c) exceeds 3, it is capped at '3'.

The number from (d) is added to the a/c development counter - this will be a number from 1 to 3 inclusive.
Once the counter exceeds 100, it moves the available date sooner by a month and resets the development counter.

It is possible to get '0' R&D now, as the step (a.1) was being bypassed post-p8 due to a bug introduced in p8.

Thank you to those who kept asking me to check the code.
The misplaced '{}' that caused this was very difficult to pick up. I was looking at it all week and just couldn't see it in the current code.

Based on the proper formula, to get a possible additional +1 from (b), it would require you to have a daily production of at least 10+.
Less than 10 would just give you '1' R&D daily from step (c).

[michaelm75au]

ORIGINAL: darbymcd

For example Nymdalar reports this from his game (post 61 http://www.matrixgames.com/forums/tm.asp?m=2825013&mpage=3)

So if my understanding is correct this is giving me the following points toward acceleration
7(0) = 1
55(0) = 3
3(0) = 1
1(0) = 1

for a total of 6 points/day. Which means that I will get an acceleration every 16-17 days which matches up with what is accutually occuring in my game.

and he could actually be more efficient to use 3 size 1 factories rather than the 55. so imagine that just 6 size 1 factories leads to an acceleration every 2 weeks. is that the idea? given the number of R&D factories the japanese has, this seems to be a problem

See earlier post.

The R&D numbers are not absolutes, there is randomness to consider. Even under the buggy version, the 55(0) could have produced 1,2 or 3 points a day.

[Numdydar]

So rather than 6/day as I projected, it would be a range of 4-6 daily. The 6 points would be the most likely outcome as the factory is 30+ correct? Or is a 1,2, or 3 just as likely?

Also Micheal you stated earlier that 30(0) was the optimal R&D factory size as that would almost always get you three points. Is that still the case os is a 1(0) size better? Inquiring minds want to know [:)]

[michaelm75au]

ORIGINAL: Numdydar

So rather than 6/day as I projected, it would be a range of 4-6 daily. The 6 points would be the most likely outcome as the factory is 30+ correct? Or is a 1,2, or 3 just as likely?

Also Micheal you stated earlier that 30(0) was the optimal R&D factory size as that would almost always get you three points. Is that still the case os is a 1(0) size better? Inquiring minds want to know [:)]

Becareful, I think you are still thinking in terms of the buggy code which was missing a vital step.

With the correct code, a 30(0) will give you a minimum of 1 R&D per day and a potential 2 every so often.
A 1(0) might give you 1 R&D per MONTH.

See post #5 above for the correct formula

[darbycmcd]

But is it also that step c ensures 1 point per day no matter the number of devices, as long as they are all repaired? that is the problem. i think it is as conceptualized that a 55 factory can produce 1-3 per day (ie around 55 / month) but a 1 factory producing 1 per day is a problem i think..... Michael do you have some insght into the design intent? step c ensures that no matter the size every factory functions at a minimum as a 30(0) factory. that is too much i think.

so i will take for example a game i am playing as the japanese in may 42. i have 15 factories researching the franka (due 1/44). now 6 of them have had some repair, if i had been smart i would have ensured that all the factories were just 1(0) so they would be giving me points now. so 6 factories {1(0)} would give me 180 points per month!!! i switch them up to frank r and i am getting that plane at the end of 43 using just 6 factories. given that japan has something in excess of 30 r&d factories to start, this is a lot of plane acceleration. is that the intent? i can accerlerate probably 5 plane lines that way to get planes from mid 45 by the end of 43 if this is the way it is.....

[kmitahj]

Factory of size 1(0) won't produce 1 r/d pt a day because it will reach step (c) only twice per month (on average). Most of the time it will fail in step (a1) because ((1+random(30)) div 30) is zero unless random(30) gives 29 or 30.

[michaelm75au]

ORIGINAL: darbymcd

But is it also that step c ensures 1 point per day no matter the number of devices, as long as they are all repaired? that is the problem. i think it is as conceptualized that a 55 factory can produce 1-3 per day (ie around 55 / month) but a 1 factory producing 1 per day is a problem i think..... Michael do you have some insght into the design intent? step c ensures that no matter the size every factory functions at a minimum as a 30(0) factory. that is too much i think.

so i will take for example a game i am playing as the japanese in may 42. i have 15 factories researching the franka (due 1/44). now 6 of them have had some repair, if i had been smart i would have ensured that all the factories were just 1(0) so they would be giving me points now. so 6 factories {1(0)} would give me 180 points per month!!! i switch them up to frank r and i am getting that plane at the end of 43 using just 6 factories. given that japan has something in excess of 30 r&d factories to start, this is a lot of plane acceleration. is that the intent? i can accerlerate probably 5 plane lines that way to get planes from mid 45 by the end of 43 if this is the way it is.....

Refer to post #5 above for the steps to get potential daily R&D points.
The very first step was NOT being done in the betas p8 to q3.

Will be fixed in q4 to come out later today.

[darbycmcd]

ok, i guess i am just unclear. to me it seems that step b could result in 0 points... fine. but then step c would add 1 no matter the size, kmataji suggests it is a test but it seems to me it is just a test for all repaired or not. so just to be clear, does step c mean that every completely repaired factory produces at least 1 point per day?

[kmitahj]

Hi,

If this number is 0, or there are any damaged devices present, then no R&D will result this turn from this factory

It is quote from Michael post #5. Take a look again on step (a1) written in bold there. This is a test I was referring to. If it fails (the number is zero) there no step (b) nor (c) so no r/d points produced. As far as I understood Michael this step (a1) is missing in few last betas but it will be reimplemented soon. Missing of this step is likely whats explain odd (too fast) research acceleration reported by Numdydar. Or at least so I hope...

[darbycmcd]

so i am super confused, because yes that is a quote from post 5, but in the same post he says
Less than 10 would just give you '1' R&D daily from step (c).
which suggests that a zero value in step b would still allow a +1 value from step c
which is what i am talking about, should a 1 factory give you 1 point per day as opposed to an ac production factory which would give you 1 aircraft per month.

[darbycmcd]

so it seems that basically research mirrors production, ok that is as i thought, but...

Michael, can you confirm the 1 point daily production for any size, repaired factory? is that what you meant about

Less than 10 would just give you '1' R&D daily from step (c).

this i think is important.

[michaelm75au]

ORIGINAL: darbymcd

so it seems that basically research mirrors production, ok that is as i thought, but...

Michael, can you confirm the 1 point daily production for any size, repaired factory? is that what you meant about

Less than 10 would just give you '1' R&D daily from step (c).

this i think is important.

ONLY if the production check (a1) produced at least '1' a/c.

If the production check returns '0' or any damaged devices, then there is no R&D that turn regardless of the factory size.

If the production check returns '1' or more, then the other checks follow on, and you end up with a minimum of '1' R&D regardless of factory size.

Note that this is how R&D use to work before p8

[FeurerKrieg]

Okay, I think it is clearer now.
A size 30(0) factory ensures a minimum of 1 R&D point per day, with some chance of 2 R&D points. (And maybe 3 as well? I'm not so sure on that part based on the comment where michaelm said you would need 570(0) to get 3.)

Thanks guys!

[bk19@mweb.co.za]

Now that we all have got to understand the R&D mechanisms I note that there is still a subtle (or apparently so) variance between factories assigned to normal production and factories assigned to R&D.

In particular, when an existing manufacturing factory space is expanded say from 30 working devices to 60 devices (intending to double capacity) these factories will continue generating output whilst the factory floor moves from 30(30) to 60(0) during the interval it takes to bring the new factory devices online.

In contrast however, according to step a.1, attempting to double the capacity of a research facility will result in a total disablement of what would normally be functional devices for the entire time it takes to actually bring on line the last of the new devices.

What is the motivation which supports what appears to be a rather draconian limitation?

[pompack]

ORIGINAL: bk19@mweb.co.za

Now that we all have got to understand the R&D mechanisms I note that there is still a subtle (or apparently so) variance between factories assigned to normal production and factories assigned to R&D.

In particular, when an existing manufacturing factory space is expanded say from 30 working devices to 60 devices (intending to double capacity) these factories will continue generating output whilst the factory floor moves from 30(30) to 60(0) during the interval it takes to bring the new factory devices online.

In contrast however, according to step a.1, attempting to double the capacity of a research facility will result in a total disablement of what would normally be functional devices for the entire time it takes to actually bring on line the last of the new devices.

What is the motivation which supports what appears to be a rather draconian limitation?

Just think of it as a real-world effect. Long, long ago in a facility far, far away (well actually only about 20 miles but it sounds good) as a team leader on research projects I learned to lock the door and refuse to accept any additional "help" when I was on a tight schedule. Each additional person added to a team makes at least one veteran unproductive until he/she gets the newbe up to speed and further disrupts the cohesion of the team for a time. If I were unfortunate enought to have my team "doubled" I would have lost even more time since I and all of the veterans would be busy updating resumes as insurance. OTOH, if I needed to double factory floor output (for hardware, not software) I could easily (given money and people) set up a line parallel to the first without disrupting on-going production at all.

Now was that the intent in this code? Probably not, but I will say the result is certainly not unreasonable as a real-world effect.

[kmitahj]

ORIGINAL: Feurer Krieg
A size 30(0) factory ensures a minimum of 1 R&D point per day, with some chance of 2 R&D points. (And maybe 3 as well? I'm not so sure on that part based on the comment where michaelm said you would need 570(0) to get 3.)

Unless Micheal decides to change original implementation no chance for 2 r&d pts.
You get guaranteed 1 r&d pt from factory of size 30(0)+ (actually 29 will do but 30 is so much easier to remember and preach)
You need at least 270(0) factory to get a chance for 2 r&d pts a day.
You need at least 570(0) factory to get a chance for 3 r&d pts a day.

In particular, when an existing manufacturing factory space is expanded say from 30 working devices to 60 devices (intending to double capacity) these factories will continue generating output whilst the factory floor moves from 30(30) to 60(0) during the interval it takes to bring the new factory devices online.

Think about factory as an entity with large number of hands working at assembly line (workers) and a few brains taking care to make the assemby process smooth (engineers). In research number of hands is not so important as number of brains working on plane improvements - engineers are perfectly happy to get one new plane prototype a day for their tests (hence factory of size 30 is enough to give steady stream of r&d pts day by day). But because the number of engineers in factory is limited you may use thier skills either on research of new aircraft model or on supervision of assembly lines expansion but not on both at the same time.
Well, it is just a story line - as good or as bad as any other I guess. [:)]