Explicit test on research system

Discuss Research and Social Aspects of the game here.

Moderators: Balthagor, Legend, BattleGoat, Moderators

User avatar
tkobo
Supreme Ruler
Posts: 12397
Joined: Jun 04 2002
Location: In a vast zionist plot ...RIGHT BEHIND YOU ! Oh Noes !

Explicit test on research system

Post by tkobo »

Okay, i took washington in the US because its the region i know the most about (having the most experience in playing it out of all the regions ive played).

I went in and switching back and forth between the regions on the map created a stable diplomatic evironment by getting alliances.

I went in and created a stable resource environment by arranging land trades and resource treaties (so that other regions on the map would send me large amount of resources i could encounter shortage in.Except oil,which is tough for most regions in this map to meet their own usage)

I set funding to about the reccomended level = to 16.71 mil

I than set the sliders to 0% 60% 40% and let the game run til research efficiency adjusted as fully as reasonable(took about 1 year),which stablized at 51% efficiency.

I than picked a short time research tech- Advanceded building materials.
Which than showed a research time of 2.9 years.

Now immediately following the first day,putting the tech into the research column cuased research effciency to fall.

This is becuase as weve been told,the money allocated to the specific tech research assigned has a cap. Because of this cap any money over said cap is directed to the other 2 possible funding allocations (tech level,research efficeincy)

Now before the tech of advanced building materials was put into the research column, the cap for the amount of the money that could be allocated to research specifics techs was zero.

SO even though the slider was set to put 60% of the 16.71 mil into specific tech research,there was no specific tech picked,so the cap was equal to zero.
This meant that the 60% of 16.71 mil was NOT being spent on researching a technology but was instead being redirected into the 2 other sliders.
What the formula is for this,we dont know.So we cant say how much of the 60% of 16.71 mil is redirected to tech level research and how much is redirected to research efficiency.




So lets look at the math at this point.

We start out with 16.71 mil allocated to everything.
The sliders further break that down to 0% allocated to tech level research,60% allocated to specific tech research, and 40% allocated to research efficiency.

So 10.026 mil allocated to research specific techs
So 6.684 mil allocated to research efficiency
So 0mil allocated towards tech level.

We than add in a specific research tech.That being ABM.

ABM has a est total cost of 1,800mil

Now at the current rate of research,2.9 years (1058.5 days), the cap for ABM on a daily basis is circa 1,700,520 per day(circa 1.7mil).

So any money allocated beyond the circa 1.7mil cap,will be redirected to the other 2 possible fund allocations.

Now we now we were originally allocating 10.026 mil/day to a specific research.

And we now know our current cap for specific research spending is circa
1.7mil.

So we can now figure out our actual spending on specific research tech and a general spending on the other 2 possible funding allocations.

Money spent on the direct research of the specific tech of ABM is now 1.7 mil/day
Money spent on the possible funding allocation of tech level research is now ZERO + some unknown percentage of circa 8.326mil.
Money spent on the possible fudning allocation of research efficiency is now equal to 6.684mil + some unknown percentage of 8.326mil.

So we see that despite using the slider to allocate funding for a specific research tech to circa 10.026mil,that due to other adjustments we are actually only spending part of that allocated money on that single funding allocation.
The new total being spent on the specific research funding is only circa 1.7mil (1,700,520)out of that 10,026mil.


Now as the research efficiency falls, due to a change in how much(if any now) of the 60% is being redirected to the research efficieny funding the time reguired to actually complete the research of the tech : Advanced building materials(from this piont on referred to as ABM) is increased and the spending cap per day lowered.


Now at the current adjusted rate of research(due to the change in amount of funds redirected),3 years (1095 days), the cap for ABM on a daily basis is circa 1,643,835 per day

NowMoney spent on the direct research of the specific tech of ABM is now 1.643 mil/day

NowMoney spent on the possible funding allocation of tech level research is now ZERO + some unknown percentage of circa 8,382mil

Now Money spent on the possible funding allocation of research efficiency is now equal to 6.684mil + some unknown percentage of circa 8,382mil


Now there are things that can/will cause ongoing adjustments to these figures.
If for instance your population and/or GDP/c is changing,it can/will change the break down of the final money spent per possible funding allocation AND can/will change the total days required to research a specific tech as well.
As will any further fall to research efficiency due to the new total of funding created not having bottomed out yet/leveled out.



I am going to try and continue this test each day. Adding a new wrinkle,like a new research center ,a new additional specific tech,increased funding,etc...
This post approved by Tkobo:Official Rabble Rouser of the United Yahoos
Chuckle TM
szabfer
Captain
Posts: 126
Joined: Aug 24 2005

Post by szabfer »

Well.... long post, long testing... however I've got a benefit: now I understand, how this technology investment cap can be calculated.

I'm really curious about the parallel projects results.... I still dont understand, why increases the time of a project if I start another project as well.... when I have 15-20 research facility and tons of funds invested... and 100+ efficiency...
User avatar
tkobo
Supreme Ruler
Posts: 12397
Joined: Jun 04 2002
Location: In a vast zionist plot ...RIGHT BEHIND YOU ! Oh Noes !

Post by tkobo »

Im not going to claim it all makes sense.The devs themselves have said they still argue over how the research system should work.

Originally,you could complete all research extremely fast(just a few game months).
So it made sense to increase the time reguired to complete a tech/s each time another tech was added to the current research que.
i dont believe the caps existed at that time(or if they did, were much larger).

Than after seeing how many people could blow thru the tech, and after a few people (me one of them) asked for them to look at this and slow things down ,they changed the system.

The changed the way efficiency reacted to changes of funding and i believe but cant say for sure they added the cap in.
The cap unfortunately makes the sliders inaccurate. This confused me greatly at first and i had to have it explained to me about the caps existance also.Something another forum member had figured out.
Sorry i forget who, or id give them their due credit.

We still dont know many things about how the research system works.
How the money redirected is split between the 2 other funding allocations is a big mystery.
And although we know theres a cap , we really havent had how it works confirmed either.Like what decides what exactly the cap is.

Some earlier testing and theory craft pointed to a total cost / total days= daily cap formula.
Which has been put forth as the way we think it works a few times now and hasnt been shot down by the devs .Yet.

My next test, which should be done later today (its like5 am here :oops: ) (unless i get side tracked )will likely be the same test but with mulitple techs.
To make sure that the dev tests from earlier are still correct and that multiple techs at once ,are the fastest way to research multiple techs.
AND the slowest way to research individual techs.
This post approved by Tkobo:Official Rabble Rouser of the United Yahoos
Chuckle TM
szabfer
Captain
Posts: 126
Joined: Aug 24 2005

Post by szabfer »

Well.... it is a big surprise for me that the research was too fast before the patches... I played a lot before update 3, and the research was painfully slow for me, every time. Maybe I was too lame...
User avatar
tkobo
Supreme Ruler
Posts: 12397
Joined: Jun 04 2002
Location: In a vast zionist plot ...RIGHT BEHIND YOU ! Oh Noes !

Post by tkobo »

Okay, i couldnt sleep so i tried a rather fast test on the effect of multiple techs.

i would post the data but it varied too much and often made no sense.

im going to run a more complete test again later.

BUT as it stands now, the research of multiple techs does indeed seem to be messed up.
I would even say there are now multiple bugs taking place depending on what you do.

One which really seemed strange was that when i picked a tech for research (added it to the que), let it run for a day and than added another research to the que,than removed both and added a third research to the que on its own, than removed that research and re added the first tech i picked back to the que, and than re added the second tech back to the que, both the first and second tech reported the same amount of time required for the research of each of them,even though up to this point they had always had a large difference between the time required for each of them.

if you think that was hard to follow reading it, your right, it was even more confusing doing and watching it happen.

It would appear that Eric Larsen and szabfer are correct that the research times are now messed up in cases of multiple researchs.

If two researched simultaneously is is not faster than one after the other, than your research spending is being limited by one of the other factors George mentioned. I did some testing before update #2 because this was a big issue for me. Don't remember what techs I used but found two that had the same time to research. One at a time, each was about 300 days but simultaneous was about 520 days each or something close to that. I believe I used Ontario in the Canada map to test it...


This is the way it used to work.But now something appears to be wrong.
And that one bug I mentioned above is just bizarre.

The only thing i can think of that can explain some of the results is the other factors mentioned by the devs and a few forum members
- Number of Research Centers
- Amount of dollars spent on Research
- Limits to how fast an item can be researched based on the number of centers you have (ie, like the example above, sometimes a research project just can't take more money spent on it, it can only progress so fast; this 'top speed' is usually based on a combination of research efficiency and the number of slots dedicated to the project).
- Region's total GDP and population size - there's only so much of your population and economy that can be dedicated to the field of research. A high GDP and population region with three research centers will be able to spend more on research than a lower GDP and/or population region.
If this all is the case, than depending on your gdp/c,amount of research centers,amount of population,amount of slots dedicated ,cap per tech per day, etc...
It might actually be the case that 3 techs researched seperately are now faster than 3 techs researched individually, under some combinations of conditions.

Which seems very odd to me.

i am going to use one of my world map save games for the next test.
As this should eliminate any issues caused by low population,low gdp/c ,low number of research centers, low max spending,etc....
This post approved by Tkobo:Official Rabble Rouser of the United Yahoos
Chuckle TM
User avatar
tkobo
Supreme Ruler
Posts: 12397
Joined: Jun 04 2002
Location: In a vast zionist plot ...RIGHT BEHIND YOU ! Oh Noes !

Post by tkobo »

yup, still not asleep :cry:

okay i played around with that bizzare case some more.

Two techs involved robotic surgery and collaborative medicare.

Robotic surgery alone takes circa 444 days.
Collaborative medicare alone takes circa 2.5 years.

Yet when i add them both to the que at the same time,they both list their individual completion times as circa 3.7 years(does not count in the extra time that will be added due to the efficiency drop as it fully takes effect).

So the time reguired to complete one tech increases circa 300% BUT
the time to research the second tech only increases circa 50% ????

This seems to show 2 possibilities.
1) the other factors mentioned have a HUGE effect on time required
2) that theres a bug

If its one,than figuring out when researching 3 techs seperately is worse than 3 techs simultaniously may be more effort than its worth to do mathimatically.
You might simply be better off using a "do and see what happens" method
on a case by case situation.
This post approved by Tkobo:Official Rabble Rouser of the United Yahoos
Chuckle TM
User avatar
Balthagor
Supreme Ruler
Posts: 22082
Joined: Jun 04 2002
Human: Yes
Location: BattleGoat Studios

Post by Balthagor »

tkobo wrote:...So the time reguired to complete one tech increases circa 300% BUT
the time to research the second tech only increases circa 50% ????...
I just skimmed the post but I would suspect this is because of the spending cap on the techs. You're able to spend enough to get Robotic Micro Surgery at a reasonable rate, but not really for collaborative medicare. Try both at once and your funding is really being overextended so both take extremely long. Again, only George has the formulas but that seems to be the way it's set up.
Chris Latour
BattleGoat Studios
chris@battlegoat.com
User avatar
tkobo
Supreme Ruler
Posts: 12397
Joined: Jun 04 2002
Location: In a vast zionist plot ...RIGHT BEHIND YOU ! Oh Noes !

Post by tkobo »

okay, taking Bals post into consideration i ran another test.

I first set spending to reccomended.
Than added one tech.
This resulted in a reasonable and expected research completion time estimate.
Than set spending to double the reccommended
Than added another tech
This also resulted in a reasonable and expected research completion time estimate.
Than set spending to max (as setting spending to 3x recommended was above max).
than added another tech
This also resulted in a reasonable and expected research completion time estimate.
than added another tech
This messed up all the tech times making them all very unreasonable.

SO, the cap does indeed seem to be the factor.
The one case i qoute above with the two research techs ending up with the exact same estimated research times due to uneven increase seems very likely to be just a very bizzare coincidence.

So possibly a rule of thumb might be to increase the funding allocated by a factor of 1 for each tech you add to the que beyond the first tech.

IE 2 techs= double reccommended funding
3 techs= triple reccommended funding

I ran this 4 times with different combinations of techs and it held true each time.

And everytime i added a tech but did NOT add a factor of funding, the times required for the techs went horribly bad.


szabfer- yes, back during update 1 or 2 i was almost always able to complete the full tech tree in a short period of time (like under a year of play) once i figured out that research done right at that time could actaully
generate positive income, which meant that there was no real end to the amount of money you could pump into research.As the money you spent on research not only paid for itself, but generated profit as well.
Others were able to max out the research tech level (circa 125) and all the techs they wanted in 6months to a year.
This post approved by Tkobo:Official Rabble Rouser of the United Yahoos
Chuckle TM
Eric Larsen
Colonel
Posts: 350
Joined: Oct 25 2005
Location: Salinas, CA

The Cap

Post by Eric Larsen »

tkobo wrote:It would appear that Eric Larsen and szabfer are correct that the research times are now messed up in cases of multiple researchs.
tkobo,
Quite an interesting and extensive test you did there. I'm envious of your diplomatic skills at getting all the AI's lined up to play nice. It's nice to see that your test has lead you to see what I've been trying to point out as far as research "cost" efficiency goes. A lot of money goes into researching tech yet not much real progress comes out. The cap you kept talking about doesn't relate to just one tech, it relates to researching tech as a whole. Really sucks that our tech research is being dumbed down. I could care less about having the capped excess shunted over to tech level and efficiency as I want techs.

I really hope that BG will get this multiple project research thing fixed so that the money we pour into researching techs is being spent on techs. I can understand and support a cap per item just to keep the game honest. But that cap should allow multiple projects to be researched so that we can expand the number of techs we get researched over time. What good does getting to the next tech level do if we can't even keep close to doing up all the old techs we want or need?

I'd like to see a "sweet spot" where the best research cost efficiency happens around the middle. Players should not be allowed to research just one tech really fast at a time unless they are willing to pay a penalty for duplication of effort. Researching too many projects would just spread spending too thin as not enough resources get devoted to each project. Somewhere in the middle, like 1.5 projects per research center plus the one freebie should be the "sweet spot" for researching tech.

I personally do not like this artificial research spending cap, especially when I'm playing the world scenario. The cap is even worse in that scenario ruining the joy of having such an awesome research budget get wasted. You should try that one just to see how bad it is compared to the smaller scenarios.
Thanks,

Eric Larsen
BigStone
General
Posts: 1390
Joined: Dec 22 2004
Location: Holland

Re: Explicit test on research system

Post by BigStone »

tkobo wrote:What the formula is for this,we dont know.
Oh yes.. we know... George is using some quantummechanical formulas
which normaly used for calculating the strength of gravityfields around black holes....

:D
NO MORE NOISY FISH [unless they are green & furiously]
I HAVE STILL A FISH IN MY EAR
User avatar
tkobo
Supreme Ruler
Posts: 12397
Joined: Jun 04 2002
Location: In a vast zionist plot ...RIGHT BEHIND YOU ! Oh Noes !

Re: Explicit test on research system

Post by tkobo »

BigStone wrote:
Oh yes.. we know... George is using some quantummechanical formulas
which normaly used for calculating the strength of gravityfields around black holes....

:D
okay, i found that VERYYYY funny.

:lol:
This post approved by Tkobo:Official Rabble Rouser of the United Yahoos
Chuckle TM
szabfer
Captain
Posts: 126
Joined: Aug 24 2005

Post by szabfer »

Well.... I have stronger and stronger feeling that the research calculations should be really simplified. We don't need "quantummechanical formulas". It is overcomplicated .... it is a game, and we are the Rulers! How can I play with a game when there is no way to understand the rules?

So, can't we have some kind of simple rules in the game? Like these...

- more research facility = more research slots and more possible money investment (as the game goes currently). Lets say 1 facility = 1 slot and X million dollar possible MAXIMUM investment on research.

- eliminate the efficiency stuff completely. I suppose that is the quantummechanical part... :) Or make it simpler: every new research facility results +X% inefficiency (can be exponential). Funding efficiency decreases this value...

- techs have base "money" cost and "time" duration. More money invested = faster research. Of course, there is a cap (minimum time duration for every tech) and the speed improvement cost can be exponential.

- the slots are independent: eg. if the player invest enough money then the research goes on full speed for every running projects. That is what I really miss currently. I have tons of research facilities (so tons of scientists) and tons of money invested. I cant believe that the research of a railway technology should slow down the research of an antitank weapon (for example)....

- If there is not enough money invested, then it is distributed evenly between the currently researched technologies. Or more user friendly - "waterfall" method: the first tech in the list gets all the money. If it reaches the cap, the remaining goes to the second. If it reaches the cap, the remaining goes to the third and so on.
Or the best user friendly way: a checkbox on the research panel with the name "distribute funding evenly" and leave the freedom of choice to the player.


Can you implement it today? 8)
Last edited by szabfer on Feb 08 2006, edited 2 times in total.
szabfer
Captain
Posts: 126
Joined: Aug 24 2005

Post by szabfer »

Or another idea: take out the research from the economy. It just drains the money from the economy and that's all. No effect on unemployment, no effect on GDP. It is just a money drainer :) If the economy is strong, then a country can invest in research. Otherway cant.

Or the last idea: a dev PLEASE take an hour or two and explain in xx rules, how this research stuff supposed to work.
User avatar
tkobo
Supreme Ruler
Posts: 12397
Joined: Jun 04 2002
Location: In a vast zionist plot ...RIGHT BEHIND YOU ! Oh Noes !

Post by tkobo »

It does still effect gdp/c.
Its nowhere near the huge effect it used to have,but it still does cause a small percentage rise.

I think it also still affects unemployment,but becuase of the new "return to center" design,the effect is no longer longterm or an accelerating effect like it used to be.

The odd thing is , it used to work exaclty as explained in the dev qoutes.

But than they changed how somethings worked in the economy , fixed some bugs affecting how things worked in the economy,changed how things worked in the research efficeincy,fixed some bugs that were affecting how things worked in the research efficiency,etc....

I begin to wonder if all the changes and fixes added together managed to over fix the existing problems.
Kinda like adding some weight to a side of a scale to achieve balance, only to find out that the other side wasnt being weighed correctly to begin with.
Than fixing the way that sides weighs its loads,only to find that they had created a new imbalance, by adding the weight they added to the other side.
This post approved by Tkobo:Official Rabble Rouser of the United Yahoos
Chuckle TM
szabfer
Captain
Posts: 126
Joined: Aug 24 2005

Post by szabfer »

Tkobo, I know that research affects to gdp, unemployment etc... In my previous post, I suggested, what would be good to change in the system - eg research should be independent from the economy. Ok, only one dependency - the money. Nothing else.

The system is too complex, as we can see one fix causes other bug now... lets make it simpler.
Post Reply

Return to “Department of Interior - Research and Social Services”