Research Upkeep
Moderator: Oberlus
Research Upkeep
The issue
To control growth of empires, FO uses (among others) the strategy of making stronger techs cost exponentially more than previous (weaker) techs.
This works well with (very) few, long research paths, but gets in trouble when there are many. On one hand, because that encourage players to go wide to get many small perks faster, rather than few stronger ones later. On the other hand, because once empires get big after going deep in few paths (and so they have RP to get humongous end-game techs that cost as much as 10 early techs together) they can get all the remaining small techs (and full tech paths) rather quickly. This means research strategies are constrained early game (you pick up the most useful tech among the cheap ones) unless a delicate balance is achieved, and are inevitably very limited end game (you pick up half the tech tree at once, little to decide).
This is particularly annoying in my design process with the themed tiered tech tree, because I want wide and deep research strategies (and mixes) to be competitive.
The general idea (from several suggestions buried in several threads in the forum):
Make techs harder to research the more techs you have. That should/could solve or at least alleviate the issue commented above.
Implementation?
We could do something like what it is done with the hull/parts upkeep: add a factor to the tech costs that increases the more techs are researched.
I don't like changing tech costs (the same I don't like changing hulls/parts costs).
Another option, that I like more, is to add a malus to research meters (similar to what is done in some cases with influence meters) that increases with the number of researched techs. The fluff/accounting label could be something like "complex knowledge body". A game rule could modify how fast increases the malus or completely disable it.
Thoughts?
/First paragraph edited to add some extra info/
To control growth of empires, FO uses (among others) the strategy of making stronger techs cost exponentially more than previous (weaker) techs.
This works well with (very) few, long research paths, but gets in trouble when there are many. On one hand, because that encourage players to go wide to get many small perks faster, rather than few stronger ones later. On the other hand, because once empires get big after going deep in few paths (and so they have RP to get humongous end-game techs that cost as much as 10 early techs together) they can get all the remaining small techs (and full tech paths) rather quickly. This means research strategies are constrained early game (you pick up the most useful tech among the cheap ones) unless a delicate balance is achieved, and are inevitably very limited end game (you pick up half the tech tree at once, little to decide).
This is particularly annoying in my design process with the themed tiered tech tree, because I want wide and deep research strategies (and mixes) to be competitive.
The general idea (from several suggestions buried in several threads in the forum):
Make techs harder to research the more techs you have. That should/could solve or at least alleviate the issue commented above.
Implementation?
We could do something like what it is done with the hull/parts upkeep: add a factor to the tech costs that increases the more techs are researched.
I don't like changing tech costs (the same I don't like changing hulls/parts costs).
Another option, that I like more, is to add a malus to research meters (similar to what is done in some cases with influence meters) that increases with the number of researched techs. The fluff/accounting label could be something like "complex knowledge body". A game rule could modify how fast increases the malus or completely disable it.
Thoughts?
/First paragraph edited to add some extra info/
- Krikkitone
- Creative Contributor
- Posts: 1559
- Joined: Sat Sep 13, 2003 6:52 pm
Re: Research Upkeep
I like the basic idea.
How exactly we implement it would depend on what we want in terms of different tech paths.
A few alternatives for the “maintenance factor”
1. Total number of techs (encourages highest value techs only...makes split paths a very bad idea)
2. Total cost of techs (split paths less of a problem)
3. Most expensive tech cost (split paths preferred)
Now that all also depends on what the benefits of higher base cost techs are... the faster the benefits ramp up, the more single path > split paths
For the mechanics, I like the idea of an “effective output” reduction....so your empire generates 100 research points, but because you have 1.5 “maintenance factor”, 60 research points 100*1.5/(1+1.5) is needed in maintenance so only 40 is available for research
How exactly we implement it would depend on what we want in terms of different tech paths.
A few alternatives for the “maintenance factor”
1. Total number of techs (encourages highest value techs only...makes split paths a very bad idea)
2. Total cost of techs (split paths less of a problem)
3. Most expensive tech cost (split paths preferred)
Now that all also depends on what the benefits of higher base cost techs are... the faster the benefits ramp up, the more single path > split paths
For the mechanics, I like the idea of an “effective output” reduction....so your empire generates 100 research points, but because you have 1.5 “maintenance factor”, 60 research points 100*1.5/(1+1.5) is needed in maintenance so only 40 is available for research
Re: Research Upkeep
Me tooI like the basic idea.
Maybe it's a crazy idea but here it is:
The science tree could be broken down into flexible tiers. That means you have 10 techs in tier1, but if you invented 5 of them all the others are moved into tier2 making them more expensive. Now you have the 5 old + 10 new in tier2. As in tier1 you have 5 techs to research till all of them went to tier3. That would mean you have to choose which techs you get for cheap research cost, gives you the discribed effect. I think as favor you could tweak the effect real good, because you could define how many techs could be developed at base cost per tier and how many a specific tech costs.
I don't know if such an idea could be implemented into the engine easily
Want some fresh experience? Try Kosymod
Re: Research Upkeep
I would prefer the themes with tiers approach (in that context we were already talking about making tech more expensive depending on the researched tech.)
As that seems far away it may make sense to introduce something like research upkeep for the current layout. I am still waiting for your first design of the theme tiers, oberlus *wink*
With current tree layout, from a UI point of view I find both approaches to upping the research cost horrible. Reducing RP production by a factor is hard to estimate for a player (e.g. you know you get +2 RP for getting a tech per planet (e.g. 20 planets -> 40 RP), but how much will it be after reduction ??32,2 RP ??) and it is a mismatch that it is a global final multiplier, but it you see it as a local effect. On the other hand upping the RP cost by a factor will make for strange numbers in the tech overview.
One detail on implementation: research prediction also needs to take a research cost increase into account.
I think we do not want to encourage split paths, but rather specialisation, so 3) is out.
2) could be ok. One detail here - I suggest to count only RP spent. That means if you get a tech via an effect it does not increase research cost.
Also what is missing - an estimation of the RP growth we currently have. I think influence mechanic as is does not really stop exponential growth because basically in the end you switch one planet to influence focus in order to support three other planets. Also nerfing the flat boni moved us to a flatter (part of the) exponential curve. So the current cost of later tech does not fit the research curve that well any more (means later tech may be almost cost prohibitive).
As that seems far away it may make sense to introduce something like research upkeep for the current layout. I am still waiting for your first design of the theme tiers, oberlus *wink*
With current tree layout, from a UI point of view I find both approaches to upping the research cost horrible. Reducing RP production by a factor is hard to estimate for a player (e.g. you know you get +2 RP for getting a tech per planet (e.g. 20 planets -> 40 RP), but how much will it be after reduction ??32,2 RP ??) and it is a mismatch that it is a global final multiplier, but it you see it as a local effect. On the other hand upping the RP cost by a factor will make for strange numbers in the tech overview.
One detail on implementation: research prediction also needs to take a research cost increase into account.
I think 1) is out, because it makes balancing very hard/restricts the tree design too much - one would have to add filler techs just to balance tech cost growth.Krikkitone wrote: ↑Fri Oct 09, 2020 1:35 am A few alternatives for the “maintenance factor”
1. Total number of techs (encourages highest value techs only...makes split paths a very bad idea)
2. Total cost of techs (split paths less of a problem)
3. Most expensive tech cost (split paths preferred)
I think we do not want to encourage split paths, but rather specialisation, so 3) is out.
2) could be ok. One detail here - I suggest to count only RP spent. That means if you get a tech via an effect it does not increase research cost.
Check out the tons of threads on themed tech tree.
Also what is missing - an estimation of the RP growth we currently have. I think influence mechanic as is does not really stop exponential growth because basically in the end you switch one planet to influence focus in order to support three other planets. Also nerfing the flat boni moved us to a flatter (part of the) exponential curve. So the current cost of later tech does not fit the research curve that well any more (means later tech may be almost cost prohibitive).
Any code or patches in anything posted here is released under the CC and GPL licences in use for the FO project.
Look, ma... four combat bouts!
Look, ma... four combat bouts!
Re: Research Upkeep
I need to run spreadsheets to get a better insight but I also get (2) is the way to go for starters.
Re: Research Upkeep
I don't really like it since Researching a Tech having negative consequences (making other techs more expensive) is a bit opposed to what Research is supposed to achieve in a 4X (always bringing positive results).
But the problem you describe is real and endgame research indeed becomes quite boring, you are very right to want to address it.
And what you propose could be a temporary solution, that actually tackles the problem, until we find a better way.
The qualm I would have with malus to research meter is that it doesn't allow to differentiate between themes (or does it ? I don't really know). It would be indeed much more coherent to have researched Techs raise the cost of Techs of different Themes (and raise them more the more their Theme is far away) but not the cost of Techs of the same Theme (so yes, at endgame one would have easy access to all the cheap Techs of his main Theme, provided he didn't disperse his efforts - I see that as a perk, not a problem).
Especially if we finally get a TAR model (I don't like the feel of it, but reading Zach's interview made me understand why it would have more interesting strategic consequences that what we have now), it doesn't make much sense to have Applications and Refinements make other Applications and Refinements of the same Theory cost more...
But the problem you describe is real and endgame research indeed becomes quite boring, you are very right to want to address it.
And what you propose could be a temporary solution, that actually tackles the problem, until we find a better way.
The qualm I would have with malus to research meter is that it doesn't allow to differentiate between themes (or does it ? I don't really know). It would be indeed much more coherent to have researched Techs raise the cost of Techs of different Themes (and raise them more the more their Theme is far away) but not the cost of Techs of the same Theme (so yes, at endgame one would have easy access to all the cheap Techs of his main Theme, provided he didn't disperse his efforts - I see that as a perk, not a problem).
Especially if we finally get a TAR model (I don't like the feel of it, but reading Zach's interview made me understand why it would have more interesting strategic consequences that what we have now), it doesn't make much sense to have Applications and Refinements make other Applications and Refinements of the same Theory cost more...
Re: Research Upkeep
What if instead of research malus give bonus for other empires which still didn't researched tech?
Gentoo Linux x64, gcc-11.2, boost-1.78.0
Ubuntu Server 22.04 x64, gcc-12, boost-1.74.0
Welcome to the slow multiplayer game at freeorion-lt.dedyn.io.Version 2024-04-14.ad50e93.
Donations're welcome:BTC:bc1q007qldm6eppqcukewtfkfcj0naut9njj7audnm
Ubuntu Server 22.04 x64, gcc-12, boost-1.74.0
Welcome to the slow multiplayer game at freeorion-lt.dedyn.io.Version 2024-04-14.ad50e93.
Donations're welcome:BTC:bc1q007qldm6eppqcukewtfkfcj0naut9njj7audnm
Re: Research Upkeep
I like that... need to think about it the get how wel it would work.
But that would not solve the problem of once getting to late game being able to get tons of smaller techs from other branches and depleting whole tiers at a fast pace.
- Krikkitone
- Creative Contributor
- Posts: 1559
- Joined: Sat Sep 13, 2003 6:52 pm
Re: Research Upkeep
The problem with research is that the cost is fixed but the benefit depends on how big your empire is.
benefits from research are per planet/ per population/ per ship, etc.
So if a tech is worthwhile for some empires at time X it will be worthwhile for all empires at time X+N.....unless the actual cost scales.
Rather than directly scaling the cost (which has mechanical micromanagment issues, as well as fluff-why does it work across tiers) I like the idea of “maintenance costs” where there is a % of your research that is spent ‘maintaining current knowledge’.
Asymptotes at 100%
Amount depends on research you produce (a % of it)
% depends on research cost of all researched/stolen/traded techs (not counting ones gifted by game engine/starting events)
benefits from research are per planet/ per population/ per ship, etc.
So if a tech is worthwhile for some empires at time X it will be worthwhile for all empires at time X+N.....unless the actual cost scales.
Rather than directly scaling the cost (which has mechanical micromanagment issues, as well as fluff-why does it work across tiers) I like the idea of “maintenance costs” where there is a % of your research that is spent ‘maintaining current knowledge’.
Asymptotes at 100%
Amount depends on research you produce (a % of it)
% depends on research cost of all researched/stolen/traded techs (not counting ones gifted by game engine/starting events)
Re: Research Upkeep
I fully agree to thatReducing RP production by a factor is hard to estimate for a player (e.g. you know you get +2 RP for getting a tech per planet (e.g. 20 planets -> 40 RP), but how much will it be after reduction ??32,2 RP ??)
As I love to plan everything possible in the game it indeed would be horrible not to know what effect it has to research something. At the moment I always watch when I get free production capacity. I than try to get new weapons / shields etc ready to produce the next batch of ships. That's why I thought that tier-system could allow to plan the research path in a better way than just an expotential growth of cost.
For outpost ships / colonies for me it is already difficult because I can't predict what the next outposter costs. I know it is expotential growth, but did't want to remember my studies in math each time I open my savegame...
I like that idea as well. If that explanation fits to the final model...I like the idea of “maintenance costs” where there is a % of your research that is spent ‘maintaining current knowledge’.
I think it's better first to discuss what to be implemented and than motivate that in the game.
Want some fresh experience? Try Kosymod
Re: Research Upkeep
I think these ideas try to address different problems. o01eg's idea addresses "catch up" if lagging too far behind. oberlus idea adresses "its stupid to get tech for (almost) free".
"its stupid to get tech for (almost) free" - could also be based on average research cost in your empire. E.g. if tech is cheaper than average research cost, it gets more expensive (fluff: need to "reinvent" prerequisites for such primitive tech).
In that case
Code: Select all
Researched tech is one where the empire actually spent RP for (so not initial tech and not gifted tech).
For a tech: RP_base(tech) is the base tech cost
For an empire:
for each researched tech of that empire: RP_sum += RP_base(tech)
for each researched tech of that empire: Tech_count += 1
RP_avg = RP_sum / Tech_count
for an unresearched tech for that empire: RP_to_research = RP_base + max(0,sum(RP_avg-RP_base))
You would need to pay 333RP (for a 100RP tech), 233RP (for a 200RP tech) and 400 RP (for a 400RP tech). Note that researching the 100 RP tech in this case would lower RP_avg to 200.
Going deep in a single spike while having cheaper base tech would not increase the RP cost for legacy much.
Going deep in a some branches is not more expensive than it is now, but increases RP cost for legacy tech.
Getting all the early tech later on would be very expensive (you could, but it would probably not be efficient).
If you need some earlier tech it is still available at an affordable price (the highest RP cost of legacy tech would be less than double the RP_avg).
Researching a single new tech (after adding some offset for early game) would not change RP_avg much, so a player can easily estimate short-term costs (for what happens after researching multiple techs one would rather rely on the research queue). AI could use heuristics for planning or exact planning (NB our AI does not plan) - it currently mostly uses playbook heuristics.
So i think this would work mechanically/game-theory wise, but it is certainly hard to understand for the player/would need some good UI to be understandable/optimizable. So while working and not hard to implement it would not be KISS - it could be an optional feature for advanced players in multiplayer.
Actually I got the feeling for single-player, we do not need research upkeep for "its stupid to get tech for (almost) free". That is because in multi-player it makes more sense to have specialised/distinct empires. Our AI is not able to create distinct empires, you basically have the player being distinct from the AI so there is no need to emphasize that. Yes, feeling more distinct from one game to the other also is good, but not as pressing.
"catch up" is actually much more relevant for single-player, on one hand for beginners and for increasing the challenge if AI can actually profit from it.
Any code or patches in anything posted here is released under the CC and GPL licences in use for the FO project.
Look, ma... four combat bouts!
Look, ma... four combat bouts!
- Geoff the Medio
- Programming, Design, Admin
- Posts: 13603
- Joined: Wed Oct 08, 2003 1:33 am
- Location: Munich
Re: Research Upkeep
Bonuses are generally more fun than penalties. This sort of mechanism can also be tied to policies for spying or keeping an empire's research protected against spying, which would respectively increase the rate of acquiring tech progress from other empires and reduce the rate at which other empires can acquire progress in techs you've researched.
Re: Research Upkeep
For that specific point I'd be more in favor of making spying easier on the more advanced Empires (not acquiring their most advanced Techs, but let's say they have Sentient Hulls and you only have Organic Hulls : getting MultiCellular Casting from them should be quite easy) as their technology permeates their whole Empire...Geoff the Medio wrote: ↑Wed Oct 14, 2020 5:50 pm This sort of mechanism can also be tied to policies for spying or keeping an empire's research protected against spying, which would respectively increase the rate of acquiring tech progress from other empires and reduce the rate at which other empires can acquire progress in techs you've researched.
Re: Research Upkeep
Just check out the themed tech catories threads (e.g. viewtopic.php?f=6&t=11207 or themes fundamentals - summary of discussion by oberlus). It is one thing we want to try. I think that was about five themes with multiple tiers (less than seven?). Unlocking tiers and using multiple tiers would make further research more expensive. There would be a visible step when this happens - researching a tech without unlocking tiers/themes would not change research cost.drkosy wrote: ↑Tue Oct 13, 2020 8:11 pm As I love to plan everything possible in the game it indeed would be horrible not to know what effect it has to research something. At the moment I always watch when I get free production capacity. I than try to get new weapons / shields etc ready to produce the next batch of ships. That's why I thought that tier-system could allow to plan the research path in a better way than just an expotential growth of cost.
UI could be very space efficient, alignment in blocks. if we make the techs on the same tier and theme the same cost, we would also save a lot of screen space and reduce unnecessary complexity.
Building a colony ship or a colony increases cost:
The next colony building becomes 3PP more expensive: additional 6% of the base colony cost (50PP)
The next outpost ship becomes moŕe expensive: additional 6% of the base outpost part cost (50PP)
The next colony ship becomes moŕe expensive: additional 6% of the base colony part cost (120PP)
Also the ships become more expensive if you build a ship: on top of the 6% the ship becomes 1% more expensive (for ship without part based upkeep)
Building twelve colonies doubles colony cost. Building 70 ships doubles ship cost.
Any code or patches in anything posted here is released under the CC and GPL licences in use for the FO project.
Look, ma... four combat bouts!
Look, ma... four combat bouts!
Re: Research Upkeep
The upkeep thingy (more expensive techs the more techs you get, or RPs invested), fluffed as "complex body knowledge", could be formulated as more cheap techs the less techs (or RPs) you have, fluffed as "focused body knowledge", to be seen as a bonus.