TAR model: Theories only require other theories?

For what's not in 'Top Priority Game Design'. Post your ideas, visions, suggestions for the game, rules, modifications, etc.

Moderators: Oberlus, Oberlus

Message
Author
User avatar
Oberlus
Cosmic Dragon
Posts: 1692
Joined: Mon Apr 10, 2017 4:25 pm

TAR model: Theories only require other theories?

#1 Post by Oberlus » Mon Jan 28, 2019 10:30 am

Main thread about Major Tech Tree Overhaul.
The main question I need to answer before anything else is that one in OP's title: will it be accepted the restriction of the Theory/Application/Refinement (TAR) model that theoretical techs only require other theories?
For now, I'd say it should be, but let's compare both options to find out the advantages of each one:

With current tech tree, theories do require applications, and there is some intense crossover between different categories and between different branches of the same category. This forces you to get many applications that you won't (or wouldn't) use.
The Silent One wrote:
Thu Jan 26, 2017 4:39 pm
here's my tech screen mockup. It's not supposed to show an actual arrangement of technologies or their connections, it's supposed to demonstrate a possible layout and some mechanisms for the TAR-system
I like very much that mockup. Imagine something alike for the whole tree. The very definite theory branches and how they gather the applications beneath them greatly improves legibility and comprehension. Fewer crossed dependencies makes easier the balance of each branch. And the fact that applications don't require many previous applications implies that you no longer must get most of the lower applications to get the higher ones, and thus you have more interesting choices to make when researching (e.g. you won't need NAI and AA to get Exobots, you will be able to go for only AA or only Exobots).

So I understand that adopting a strict TAR model, and specifically the theories-only-require-theories restriction, would bring in
  • better legibility and comprehension (KISS),
  • easier balance, and
  • more variety of strategies.
I fail to identify any advantage of the current system over this one. Please, let's discuss.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 12487
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

Re: TAR model: Theories only require other theories?

#2 Post by Geoff the Medio » Mon Jan 28, 2019 1:47 pm

My main issue with a Theory-Application-Refinement tech system is that it means there are a lot of techs that do nothing. This means the player has to wait longer to get to the next fun / interesting content. The system looks nice in a diagram and is interesting to think about as a design, but I don't find it particularly interesting to play compared with just having techs that do things.

If it's especially important / desirable to make choosing a path through the tech tree as flexible as possible, ie. not having to get any particular tech before any other, then a tiered system, similar to the original MOO, would be possible. There are 1-3 techs on each tier (per category), and any of them being researched would mean the techs of the next tier become researchable. These techs would all be "applications", that actually do something for the player, however. This is more-or-less equivalent to making all multi-prereqs optional, so players need just one, not all, of a tech's prereqs to unlock the tech. Having more than one prereq could do nothing, or could make a tech cheaper / faster to research. (I'd also throw in some non-tech prereqs, like having a building or species or special or similar, which do the same as having a tech prereq, including unlocking or making a tech cheaper / faster to research.)

User avatar
Oberlus
Cosmic Dragon
Posts: 1692
Joined: Mon Apr 10, 2017 4:25 pm

Re: TAR model: Theories only require other theories?

#3 Post by Oberlus » Mon Jan 28, 2019 3:36 pm

Geoff the Medio wrote:
Mon Jan 28, 2019 1:47 pm
My main issue with a Theory-Application-Refinement tech system is that it means there are a lot of techs that do nothing.
Fair enough.

So right now I'm between two alternatives:

1. Pseudo-TAR system that removes the doing-nothing theoretical tech problem and keeps the rest of TAR system and can use (a variation of) The Silent One's mockup:
- Grant every theory an immediate application. The empty space for theories in the mockup would be filled with these applications.
- Above and under the theory (in the mockup) can be up to two sub-branches of applications for each theory (let's call it A the one above and B the one below).
- Every application under the same theory can be researched directly, but they will become cheaper if the immediatly previous app on its sub-branch has been researched first (see example)
- Any app can have at most one theory or application prerequisite (being it a tech or a building) coming from a different sub-branch of this or another category.
- Refinements (always of apps that are not the one given by a theory) only have the corresponding app or the previous refinement as requirement.

Example:
Theory_T1 unlocks Apps branches A and B, with three and one apps respectively, for App_T1A1, App_T1A2, App_T1A3 and App_T1B1.
Researching App_T1A1 will reduce cost of App_T1A2, that will reduce cost of App_T1A3, but non of these will affect cost of App_T1B1, that will not affect the cost of any App_T1A*.

2. Tiered system. I have to think of this more carefully, but the following seem right:
- The linked mockup could be used only if enough prerequisites are removed from current tech tree. Otherwise, a different one should be disigned, or stick to the current one and forget about better readability/understandability.
- Theories should be removed all together, either by merging them with one application (as in my other choice) or by spreading somehow its cost among the following techs.

User avatar
The Silent One
Graphics
Posts: 1019
Joined: Tue Jul 01, 2003 8:27 pm

Re: TAR model: Theories only require other theories?

#4 Post by The Silent One » Mon Jan 28, 2019 6:56 pm

Geoff the Medio wrote:
Mon Jan 28, 2019 1:47 pm
My main issue with a Theory-Application-Refinement tech system is that it means there are a lot of techs that do nothing.
But isn't exactly that one of most compelling points of the TAR system - to make the player choose between researching relatively cheap applications that will have an immediate (if minor) effect, and researching an expensive theory that will not immediately yield anything, but unlock more powerful applications and may let you win the game /tech race in the long run?

E.g.:
Do I have a use for the Cyborg Implant app, which will increase my max population somewhat, or a more expensive Ground Troop Pod refinement to buff my troop transports for an invasion; or do I rather go for the Transorganic Sentience theory which will take while, but unlock the powerful Pure Energy Metabolism, which will greatly increase production and science?

I absolutely agree that there shouldn't be many techs that do nothing. For that reason, the relationship between theories and applications should be a small number of theories compared to many applications, with maybe 3-4 refinements per app.
If I provided any images, code, scripts or other content here, it's released under GPL 2.0 and CC-BY-SA 3.0.

Morlic
AI Contributor
Posts: 285
Joined: Tue Feb 17, 2015 11:54 am

Re: TAR model: Theories only require other theories?

#5 Post by Morlic » Mon Jan 28, 2019 7:37 pm

I agree with The Silent One that the TAR model can lead to more interesting decision making if implemented well. In a good implementation, there would be relatively few but expensive theory techs which unlock a variety of new techs. It is a proven design in particular in the RTS genre where you have some sort of unit/tech tier system and may offer many strategic choices.

However, I am worried that the design is hard to implement if you have parallel branches with (roughly) symmetric tech costs. If the theory tech cost scales something like 10/100/1000/10000 RP in a single branch, then the cost of parallel theory branches becomes trivial compared to going deep into a particular branch. If the costs do not scale exponentially, then you will run into problems with empire's exponential RP scaling and again trivialized cost of theory techs.

I think a single string of theory techs (basically a tier system) could work better.
If I provided any code, scripts or other content here, it's released under GPL 2.0 and CC-BY-SA 3.0

User avatar
Krikkitone
Creative Contributor
Posts: 1508
Joined: Sat Sep 13, 2003 6:52 pm

Re: TAR model: Theories only require other theories?

#6 Post by Krikkitone » Mon Jan 28, 2019 8:31 pm

The "Theories do nothing" issue of
strategy (long term investment with no immediate payout)
v.
game reward (spending time without a reward)

can probably be solved by
1. having the "Theories" have a fairly limited investment (ie a large number of minimum turns)

So you will be working on a Theory (no tangible reward) but you will probably also be working on a few Applications

2. UI wise, whenever the Theory is finished it can give you the reward of showing all the new Applications you have available.

As for making the cost work... an example (sorry didn't read the whole point)

Theory
10/100/1k/10k
Application (2-5 per Theory tech)
30/300/3k/30k

So getting all the applications granted by a theory requires just as much research as getting the next level (of course the actual reward for those applications should be proportional to their exponentially increasing cost)

As for the deep v. wide, An option would be for Tech cost to be dependent on total Theory Levels

User avatar
EricF
Space Dragon
Posts: 354
Joined: Fri Jul 29, 2016 10:12 am

Re: TAR model: Theories only require other theories?

#7 Post by EricF » Tue Jan 29, 2019 12:32 am

Geoff the Medio wrote:
Mon Jan 28, 2019 1:47 pm
My main issue with a Theory-Application-Refinement tech system is that it means there are a lot of techs that do nothing.
This has been my major complaint about the current tech tree since about the first time I played FO.
I really really dislike researching techs that give me nothing. It's not FUN.
Morlic wrote:
Mon Jan 28, 2019 7:37 pm
I think a single string of theory techs (basically a tier system) could work better.
This reminds me of a saying "Everything is Physics".
Quantum Theory leads us to our modern understanding of Chemistry and hence also Biology.

One more thing. Sorry to be a Party pooper, I know Theory/Application/Refinement is how it works
in Real Life, but when has it become a goal of the Project to model RL?
I actually think pruning the Tech Tree of all the Do Nothing techs would make it easier for people to
navigate and make decisions and make more room for expansion without making it too cluttered.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 12487
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

Re: TAR model: Theories only require other theories?

#8 Post by Geoff the Medio » Tue Jan 29, 2019 6:36 am

The Silent One wrote:
Mon Jan 28, 2019 6:56 pm
But isn't exactly that one of most compelling points of the TAR system - to make the player choose between researching relatively cheap applications that will have an immediate (if minor) effect, and researching an expensive theory that will not immediately yield anything, but unlock more powerful applications and may let you win the game /tech race in the long run?
I don't consider that a very "compelling" point. The do-nothing techs just seem like a waste of time and player attention, so that the tech tree designer can feel good about their elegant "TAR" tree structure. There is also about as much choice with tiers of mutually-optional techs, as instead of choosing between immediate or delayed benefit, one can choose between several things per tier. Both presumably can or do also allow choice between additional thing on a tier vs. moving to the next tier, or skipping version 3 of something since you already have version 2 which is good enough, or waiting to get version 5 on a future tier.

User avatar
Oberlus
Cosmic Dragon
Posts: 1692
Joined: Mon Apr 10, 2017 4:25 pm

Re: TAR model: Theories only require other theories?

#9 Post by Oberlus » Tue Jan 29, 2019 9:37 am

So some of us find purely theoretical techs not fun, and others actually like the idea of a do-nothing tech whose reward is unlocking more techs.
We may not have a consensus that way, looking for what is funnier.

I think we should go the way that does not cause unfun to anyone.

Would you all accept a system (tiered, branched or a mix) that does not have purely theoretical techs?
That is, all techs would have at least one functionality other than unlocking other techs.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 12487
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

Re: TAR model: Theories only require other theories?

#10 Post by Geoff the Medio » Tue Jan 29, 2019 10:27 am

I suppose I should refine my previous post a bit... I don't like required do-nothing techs. If there was a "theory" / do nothing tech in some cases, which is optional but still gives the effect of unlocking the next tier, that would be fine. I'd actually like having several ways to unlock tech tiers or individual techs, such as researching a very RP-expensive but fast (minimum) research time do-nothing tech, researching a medium RP-cost and moderate time tech that unlocks stuff, owning a building or ship part (perhaps that was unlocked by a tech on the previous tier), or having a particular species in the empire.

Ophiuchus
Programmer
Posts: 1045
Joined: Tue Sep 30, 2014 10:01 am
Location: Wall IV

Re: TAR model: Theories only require other theories?

#11 Post by Ophiuchus » Tue Jan 29, 2019 5:15 pm

Geoff the Medio wrote:
Tue Jan 29, 2019 6:36 am
The do-nothing techs just seem like a waste of time and player attention,
player attention 1) I agree that the MoO tiers were more tightly packed/making a lot better use of screen space.

player attention 2) I do not think that do-nothing techs are confusing for players if they give a uniform structure
Geoff the Medio wrote:
Tue Jan 29, 2019 6:36 am
so that the tech tree designer can feel good about their elegant "TAR" tree structure.
TAR tree structure 1) Again I think the structure helps players and designers to understand, so there is value in there. But another structure (like the MoO tiers) might deliver more value.
Any code or patches in anything posted here is released under the CC and GPL licences in use for the FO project.

Furthermore, I propse... we should default to four combat rounds instead of three ...for the good of playerkind.

User avatar
Krikkitone
Creative Contributor
Posts: 1508
Joined: Sat Sep 13, 2003 6:52 pm

Re: TAR model: Theories only require other theories?

#12 Post by Krikkitone » Tue Jan 29, 2019 6:18 pm

Geoff the Medio wrote:
Tue Jan 29, 2019 10:27 am
I suppose I should refine my previous post a bit... I don't like required do-nothing techs. If there was a "theory" / do nothing tech in some cases, which is optional but still gives the effect of unlocking the next tier, that would be fine. I'd actually like having several ways to unlock tech tiers or individual techs, such as researching a very RP-expensive but fast (minimum) research time do-nothing tech, researching a medium RP-cost and moderate time tech that unlocks stuff, owning a building or ship part (perhaps that was unlocked by a tech on the previous tier), or having a particular species in the empire.

That is one possibility.
Tier unlock by Theory (cheap option no reward beyond techs) or by completing X techs on that level (more expensive, but actually give rewards)

User avatar
The Silent One
Graphics
Posts: 1019
Joined: Tue Jul 01, 2003 8:27 pm

Re: TAR model: Theories only require other theories?

#13 Post by The Silent One » Wed Jan 30, 2019 11:35 am

Morlic wrote:
Mon Jan 28, 2019 7:37 pm
I think a single string of theory techs (basically a tier system) could work better.
Geoff the Medio wrote:
Tue Jan 29, 2019 10:27 am
I'd actually like having several ways to unlock tech tiers or individual techs, such as researching a very RP-expensive but fast (minimum) research time do-nothing tech, researching a medium RP-cost and moderate time tech that unlocks stuff, owning a building or ship part (perhaps that was unlocked by a tech on the previous tier), or having a particular species in the empire.
Krikkitone wrote:
Tue Jan 29, 2019 6:18 pm
Tier unlock by Theory (cheap option no reward beyond techs) or by completing X techs on that level (more expensive, but actually give rewards)
An attempt to consolidate these suggestions below: top line is theories/tiers. They become unlocked if you research a certain number of applications from the previous tier, or alternatively, if your empire meets certain conditions as having a certain species/buildings etc. - in the example below, terraforming only requires 4 applications to unlock if you should (arbitrarily) have the secret gardener in your empire. Applications can have refinements which increase the app's effect, but do not help unlocking the next tier.
Attachments
tech_mockup_tiers.jpg
tech_mockup_tiers.jpg (199.59 KiB) Viewed 1374 times
If I provided any images, code, scripts or other content here, it's released under GPL 2.0 and CC-BY-SA 3.0.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 12487
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

Re: TAR model: Theories only require other theories?

#14 Post by Geoff the Medio » Wed Jan 30, 2019 12:45 pm

The Silent One wrote:
Wed Jan 30, 2019 11:35 am
[Tech tiers] become unlocked if you research a certain number of applications from the previous tier, or alternatively, if your empire meets certain conditions as having a certain species/buildings etc.
I'd still keep the option to just spend research points to unlock a tier, without having to research the previous tier's applications. I'd probably balance this by having it cost about as much as researching all the required applications to unlock the tier, but having a low minimum research time, so the tier can be unlocked faster if an empire has very high RP output.

Also, consider having a tier be highlighted to mean that that tier has the conditions met to unlock the next tier, ie. that tier is finished and you can move on, and the checkboxes for a tier refer to the items in that tier, not the tier below.

User avatar
Oberlus
Cosmic Dragon
Posts: 1692
Joined: Mon Apr 10, 2017 4:25 pm

Re: TAR model: Theories only require other theories?

#15 Post by Oberlus » Sat Feb 02, 2019 1:12 pm

Summing up desirable characteristics of the tech tree:
  1. Visibility: Techs are shown more "tighly"/"packed up", making a better use of the screen and allowing with few or no extra clicks to know how to get to a given tech. A pseudo-tiered system seems the best option.
  2. Replayability: several "tech pathes" to get to advanced researchs (different list of techs = different set of effects at different times in the sequence = different strategies). Example: getting these applications or these other applications or that theory.
  3. Do-nothing theoretical techs (for one way to unlock the next tier) are an option, but keep them in the low numbers.
  4. Anything else?
Geoff the Medio wrote:
Wed Jan 30, 2019 12:45 pm
I'd still keep the option to just spend research points to unlock a tier, without having to research the previous tier's applications.
So we have these ways to unlock apps of tier X+1:
  1. Research theoretical tech (tier unlocker) with no effect.
  2. Research a subset of apps of tier X (if the subset is the whole set of apps of tier X, then this is just one way, if not all apps are required then there are much more alternative pathes).
I'd probably balance this by having it cost about as much as researching all the required applications to unlock the tier, but having a low minimum research time, so the tier can be unlocked faster if an empire has very high RP output.
Assuming there are M+N applications, {a, b, ... n}, with an average cost of P0 RPs and T0 turns each, and you require to unlock next tier (its apps) the next theoretical tech, u, or M applications of this tier (a subset of M elements of {a..n}, unless N=0). Then, as of your suggestion, cost of u should be M*P0.
I think it can work, but will need balance. The point is it must not be a no-brainer, i.e. there shoul be effective choices at different situations.
If I have low RPs I assume the commonly good choice would be to go for as much apps of this tier as possible, so that you get effects much sooner. Once enough apps have been researched, you can start researching next tier apps saving yourself the M*P0 RPs of the theory. But even if a next tier app is really necessary for further expansion and none of the current tier apps will give you a useful benefit at this point of game you will go for the apps instead of the theory because you will require the same or time either way but going the apps path you save yourself the RPs from researching the theory and you get the current tier apps (that will have some use later on). Therefore, maybe theory should cost less than the sum of the apps that will also do the unlocking.
If I have enough RPs to go for the theory in its minimum time T1, I also have enough RPs to go for all the M apps simultaneously in some more time T2 but getting all the effects. This seems more of a choice. If going the theory, I get to unlock the next tier apps and then I have to research one of them. Assuming the next tier apps are comparable in cost but greater time (again T2) than the current tier theory (T1), then I'll be getting an expensive app in T1+T2. Going the apps way I'd be getting all effects in T2 and then one expensive effect at T2+T2, that maybe is a 20%-50% slower than going the theory. So if I need some of current apps effects (or are going to boost your research), it'd be better to go the apps path, and the theory path if I only need that special effect of next tier. If theory costs less (as comented in previous paragraph for low RP empires) then high RP empires could be going the theory and some of the apps of a different tech branch (category) at the same time, or the theories of two different branches.
Summing up, I guess the theory way to unlock next tier apps would be a choice whenever there is urge to get to some next tier app for both low and high RP empires. I guess that is good. Carefully allocation of the effects among the tiers should do the trick to allow for interesting choices in different situations.
Geoff the Medio wrote:
Wed Jan 30, 2019 12:45 pm
consider having a tier be highlighted to mean that that tier has the conditions met to unlock the next tier, ie. that tier is finished and you can move on, and the checkboxes for a tier refer to the items in that tier, not the tier below.
I'm not sure I understand this.
If no explanation was added, from The Silent One's mockup, I think I would assume that a theory, once highlighted, unlocks (allows to invest RPs into) the apps below/under it (in its "column") as well as the theory tech of the next tier. And the checkboxes within a theory, I expect them to be the indication of what is missing to bypass the theory (hovering the mouse over it would shop an informative popup).
Note that with "bypass" I refer to the unlocking of the tier via the previous tier apps, which effectively renders useless to research that tier theory and so it should appear as highligthed (researched) although no RPs were invested on it.

Post Reply