Stellaris

Stellaris

Not enough ratings
Research Points Study
By AngleWyrm
Study of research point accumulation
   
Award
Favorite
Favorited
Unfavorite
Introduction
How many Research Points (RP) does the player have at any point in the game's timeline?

There are many mods that add researchable items, and some cost a lot of RP. Those prices are usually chosen from a list of options in 00_scripted_variables.txt

When some of those modded research options appeared in my game, their RP costs put them over five years to complete, some nearly ten years. Their cost was too high to be chosen, compared to other items I could research.

I suspected that RP cost is for modders to decide when along the game's timeline their tech should become a competitive option. Other contributing factors being all the RP spent on prerequisite techs, and the chances of the tech card and it's predecessors being drawn.

So I tracked RP across Year, Population count, Planets owned, and Sprawl as a general indicator of empire size.

Research Points Study[docs.google.com]
Science!
Are any of the tracked variables (Year, Pop, Worlds, Sprawl) related to total Research Points?

I ran the game at top speed with AI taking the player's role. Vanilla + all DLC rated 'Mostly Positive' or better, 600 stars, Commonwealth Of Man, 1x research speed, 1x habitable worlds, max Hyperlane density, Ensign difficulty, Mid:2400, End:2600, Victory:2800

Correlation with Research Points
Run
Ended
Year
Pops
Worlds
Sprawl
Estimators of relationship
1
2365
0.1640
0.8166
0.6313
0.7749
Pop= 0.352×RP +2.53[docs.google.com]
RP= 1.38×Pop +82.8[docs.google.com]
2
2720
0.2363
0.8548
0.8425
0.8628
Pop= -0.0000271×RP^2 +0.208×RP +63.4[docs.google.com]
RP= 0.00806×Pop^2 +2.1×Pop -57.4[docs.google.com]
3
2505
0.4509
0.2480
0.4809
0.5054
Correlation too low across all tracked measurements

Observations on first run
  1. Population stood out as having the tightest match to RP, and Sprawl came in second
  2. With only 165 years to the run, a simple linear equation was sufficient to approximate the relationship
Observations on second run
  1. Sprawl rated just barely better than Population count, but population is a more atomic and therefore easier to handle concept, so I continued to use that measurement
  2. The larger set of data fit better to a quadratic equation, suggesting a second influence. I'm guessing the Pop^2 factor approximates acceleration due to research bonuses
  3. The difference between estimated and actual RP ranged across about +/- 240 RP
  4. Several wars caused sudden and dramatic shifts in population over time[docs.google.com]
Observations on third run
  1. Population stagnated around 60~110 for most of game[docs.google.com]
  2. Empire spread across five sectors early on
Summary
Research is proportional to Population, thus population can serve as a tool for modders to decide how much RP to charge for their tech.

The spread of the data (+/- 240 RP) suggests significant noise from random variance.

Lookup table for Years until a given pop size can complete a project[docs.google.com]

< >
3 Comments
AngleWyrm  [author] May 27 @ 6:09pm 
Click the links. Found at the end of the Introduction section, and again at the end of the Summary section.
post data all data
i need more data can you use spread sheet