A couple of doubts about research

  • Crusader Kings II Expansion Subscription

    Subscribe to the CK II Expansion and enjoy unlimited access to 13 major expansions and more!


  • Paradox Midsummer Sale has arrived! Up to 75% off!

    Enjoy some sun and song this Midsummer, but when the sun goes down, the fun doesn't have to stop! Paradox has a festive sale on plenty of games to keep your summer nights going!


    June 18th - June 30th
  • Crusader Kings III Available Now!

    The realm rejoices as Paradox Interactive announces the launch of Crusader Kings III, the latest entry in the publisher’s grand strategy role-playing game franchise. Advisors may now jockey for positions of influence and adversaries should save their schemes for another day, because on this day Crusader Kings III can be purchased on Steam, the Paradox Store, and other major online retailers.


    Real Strategy Requires Cunning

Pandrea

First Lieutenant
48 Badges
Jun 18, 2013
258
172
  • Crusader Kings II
  • Europa Universalis III Complete
  • Europa Universalis IV
  • Darkest Hour
  • Age of Wonders III
  • Magicka 2
  • Victoria 2
I have always had two doubts about how research works
1) What is better between level and skills for a tech team (TT) ? Should I use a lvl9 TT with 0 skills or a lvl6 TT with 4 skills on a specific research?
2) I know there's a minus for researches before their year. But what about I start a 1937 research on December 1936? It will go slower for a month o for the entire research?
 

Altruist

Captain
9 Badges
Sep 13, 2011
395
306
  • Darkest Hour
  • Stellaris
If the formula provided by Argandios works like intended, let's give it a try...

For simplicity we'll take 1936 Infantry, 5 fields of level 4 to be researched, no double/extra field, no pre-/post date research, it is year 1936 in which we research this tech:

inf1936.png


Code:
T = 100*D / [1.65*Nr + {(100+Er)*3*(6.1+Tl)} / {(100-Em)*(4+2*Dr)*MF*B}]

T = research Time of that part of the research.
D = Double time, 2 if it is one of those parts which require of double time (or double effort, or whatever you want to call it), 1 if it is not.
Nr = Number of rocket or nuclear research facilities you have, if that part of the research is related to either rocket or nuclear science projects, if that part is not related with any of those, this number is 0.
Er = Extra research efficiency due to industrial research for tabualte and computing machines, if any.
Tl = Research team level.
Em = Extra research efficiency due to any minister, if any.
Dr = Dificulty level of that part of the research.
MF = Main Field research, 1 if it is NOT a Main Field research for the team, 0.5 if it is.
B = Blueprints, 1 if if u do NOT have a blueprint, 0.5 if you do have a blueprint.

1) Level 9 tech team with 0 matched fields, no blueprint, same year as tech
Let's fill out the variables for your level 9 tech team with no matches:
D: 1 (no double time)
1.65*Nr: irrelevant
100+Er... replaced with 53: we need to adjust this, because in DH, misc.txt line 955, the research modifier was lowered from 1 to 0.4, so we'll change 100 to 40 and for Er (extra research ability) we'll assume all computer techs upto 1936 to be researched which adds up to +13, altogether 53... or the easiest way, just replace 100+Er with research speed as shown in the tech overview (right coloumn, bottom), in this case it would show 53%.
Tl: 9
Em: 0
Dr: 4 (same for all 5 tech fields)
Mf: 1 (no matched fields)
B: 1 (no blueprint)

This gives us the following:
Code:
T = 100*D / [1.65*Nr + {(100+Er)*3*(6.1+Tl)} / {(100-Em)*(4+2*Dr)*MF*B}]
T = 100*1 / [1.65*0  + {  (53)  *3*(6.1+9 )} / {(  100 )*(4+2*4 )*1 *1}]
T = 100   / [                  2400.9        /         1200            ]
T = 49.98 => rounded up: 50

One tech field should take 50 days, we have 5 fields, for all the same conditions, so we can simply say:
50 * 5 = 250 days it should take a L9 tech team to research Inf 1936 in 1936 with no matches and no blueprint.

2) Level 6 tech team with all tech fields matched, no blue print, same year as tech

Code:
T = 100*D / [1.65*Nr + {(100+Er)*3*(6.1+Tl)} / {(100-Em)*(4+2*Dr)*MF*B}]
T = 100*1 / [1.65*0  + {  (53)  *3*(6.1+6 )} / {(  100 )*(4+2*4 )*0.5 *1}]
T = 100   / [                  1923.9        /         600            ]
T = 31.19 => rounded up: 32

32 * 5 = 160 days it should take a level 6 team with all fields matched to research Inf 1936 in 1936, no blueprints.

Comments:
I haven't tested the formula very much, I hadn't the time since I saw it having been posted. So, I can't say wether it is a correct one.
Due to the much lowered research modifier in DH in comparison to HoI2 (from 1 to 0.4) research takes generally quite much longer... this was also the reason why the old approach was leading to wrong results. In other mods this has been taken even further: Edge of Darkness has lowered the research modifier to 0.3, same for World in Flames 2. This means that in relative comparison automatically various things have been strengthened or made much more important:
# computer techs, ministers raising the research speed
# blue prints (additionally EoD and WiF2 raise the bonus blue prints give from 2 to 2.5 which results in a changed multiplier from 0.5 to 0.4, lower=better)
# matching fields (so the old rule of thumbs don't work anylonger as you can quite clearly see from above calculations)

Even if the new formula doesn't hit it 100% correct, we have all noticed the longer research times in comparison to HoI2.
My experience: I can easily research 3 or 4 old techs with blueprints and well matched fields in the time it takes to research one juicy important tech of the present year without blueprints. Some techs are worth it, some not. Due to the lengthier research times, now more fall into the category: not worth research asap IF you have still interesting old techs available.
 
  • 4
  • 1Like
Reactions:

Altruist

Captain
9 Badges
Sep 13, 2011
395
306
  • Darkest Hour
  • Stellaris
# matching fields (so the old rule of thumbs don't work anylonger as you can quite clearly see from above calculations)
To stress this point... in above example, even a level 2 tech team with matched fields would be still (a little bit) faster than a level 9 tech team with no matched fields.
2) I know there's a minus for researches before their year. But what about I start a 1937 research on December 1936? It will go slower for a month o for the entire research?
Interesting question.
 

Pandrea

First Lieutenant
48 Badges
Jun 18, 2013
258
172
  • Crusader Kings II
  • Europa Universalis III Complete
  • Europa Universalis IV
  • Darkest Hour
  • Age of Wonders III
  • Magicka 2
  • Victoria 2
From wiki:
You get a cumulative 0.15% or 0.125% penalty (the latter is correct in Doomsday V. 1.3a) penalty per day for researching ahead of time. The relevant time for a tech is always January 1 of the corresponding year. So a 1938 technology started at 01.01.37 would get a 360*0.15% = 54% penalty for the first day, a 53.85% penalty for the second day and so on until you reach 01.01.38 when you don’t get any penalty for the remaining research of this tech anymore. The cap for the penalty is at 90%. This cap is reached by researching 600 days ahead of time if the penalty is 0.15%, 720 (or exactly two years, HoI years being exactly 360 days) if the penalty is 0.125%. Thus, starting to research a 1938 technology in April of 1936 would have the maximum 90% penalty. For these reasons, it is generally recommended not to "tech rush" more than 1 year early, and wherever possible to stay with current year technologies. There is no bonus for researching old technologies.

Is it still valid in DH?
 

Altruist

Captain
9 Badges
Sep 13, 2011
395
306
  • Darkest Hour
  • Stellaris
Is it still valid in DH?
Partly... if in doubt, usually a look into the misc.txt file helps. In this case line 1024 and following:
Code:
# Blueprint Bonus
    2.0
# Pre-Historical Date Mod (value * days prior to historical date. Lower limit for the total is -0.9. )
    -0.0022
# Post-Historical Date Mod ( 1 + value * days post historical year). The Highest possible value is equal to Blueprint Bonus. Doesn't apply to Nuclear and Rocketry researches
    0.000925 #0.0 #full bonus after 3 years post historical

The blueprint bonus is applied in the research formula as 1/blueprint_bonus. Which is important only for mods which change the blueprint bonus, otherwise you can just use 0.5 for blueprint and 1 for no blueprint as a mulitplier in above given formula. For EoD or WiF2 it would be 1/2.5 = 0.4 if blueprint available.

I guess the pre- and posthistorical modifier can be added just behind the blueprint multiplier as another multiplier of:
1 / (1 + historic_mod * offdays)
With a maximum of (1 + historic_date_mod * offdays) not higher than Blueprint bonus, which is why max is reached after 3 years in DH full. Might be that there is no max the other way around for researching ahead.

Altogether the formula is too big and long to use, especially since you need to do it for each tech field separately. But it is worth doing it for several cases to get a feeling for how much longer it takes to research ahead and how much faster old techs with blueprints can be researched. So you achieve the status of making an educated guess wether it is really more useful to research one juicy uptodate tech which takes a loooong time in comparison to 4 (or sometimes even more) old techs where you have blueprints and the boni those give you.

It certainly has changed my choice of techs to research. In my present game, year 1940, I deem exactly 3 techs of all the very many highly intersting 1940 techs as important enough to actually research them now: LCAG/CAG 1940, Improved Computing Machines (1940) and Assembly Lines and I am not sure wether the last one is really worth it. This is still for my China game with lots of old techs available for research.
 
Last edited: