Tech Tree Discussion: FO vs. MoO2

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

Moderator: Oberlus

Message
Author
User avatar
Bigjoe5
Designer and Programmer
Posts: 2058
Joined: Tue Aug 14, 2007 6:33 pm
Location: Orion

Re: Tech Tree Discussion: FO vs. MoO2

#16 Post by Bigjoe5 »

eleazar wrote:You still have do-nothing theories in there. Is that because you want them, or they were necessary with the current engine to organize your applications?
Those are there for organization purposes. The idea is that the player researches one of the applications directly rather than researching a theory (which is generally quite unfulfilling), then the applications. Making the tech tree actually work the way I want it to with the current mechanics would be very difficult or impossible. Also in general, nothing actually does what it claims to - everything is just hypothetical.
eleazar wrote:Sometimes you have things like Biology V, and Biology V-B. Are those intended to be mutually exclusive?
Nope. They are separate little branches that the player can optionally research. The player needs to research one application from Biology Level V (for example) to move up the Biology category, but he can also choose to research an application from Level V-B at any time (after that tech level is unlocked), or he can ignore it.
Warning: Antarans in dimensional portal are closer than they appear.

User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: Tech Tree Discussion: FO vs. MoO2

#17 Post by eleazar »

Of course i've never played a game with tech set up like that, but my suspicion is that combining a MoO2-style multiple-choice columns (MCC) with the branches and more complex dependancies of a Tech Tree, that you end up with the worst of both worlds, rather than the best. What you get is more complicated and confusing than a plain tree, which in our case is unlikely to be an exemplar of simplicity even if we throw out theories.

I can't really say that either MCC or Trees are inherently better. They have different strengths. But i think we should pick one or the other, not try to have both.

User avatar
Bigjoe5
Designer and Programmer
Posts: 2058
Joined: Tue Aug 14, 2007 6:33 pm
Location: Orion

Re: Tech Tree Discussion: FO vs. MoO2

#18 Post by Bigjoe5 »

I disagree. Can you explain what the worst of both worlds is, and what's confusing about the tree I've provided?
Warning: Antarans in dimensional portal are closer than they appear.

User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: Tech Tree Discussion: FO vs. MoO2

#19 Post by eleazar »

Bigjoe5 wrote:I disagree. Can you explain what the worst of both worlds is, and what's confusing about the tree I've provided?
eleazar wrote:What you get is more complicated and confusing than a plain tree, which in our case is unlikely to be an exemplar of simplicity even if we throw out theories.
* If you have on average 3 choices, all other things being equal, there will be 3 times as many items in the tech tree
* I realize your "Biology X" type theories are just placeholders to organize things, but if displayed in a tree, you'll need something to bind things together and show the relationships between the multiple choice techs, and for attaching your branches and crosslinks, so that brings us up to 4 times as many items.
* You'll need some way to identify techs that aren't researchable because you chose another option, doable but it's something else that moves further into "more complicated".
* You loose one of the nice things about a multiple-choice-column (MCC)-- it can be displayed more simply and clearly than a tree , in a list with columns.

User avatar
Bigjoe5
Designer and Programmer
Posts: 2058
Joined: Tue Aug 14, 2007 6:33 pm
Location: Orion

Re: Tech Tree Discussion: FO vs. MoO2

#20 Post by Bigjoe5 »

eleazar wrote:* If you have on average 3 choices, all other things being equal, there will be 3 times as many items in the tech tree
3 times as many as what? 3 times as many as if the player is expected to research every tech? I don't think that was going to be the case to begin with. Overall, we want there to be a way to make players research some techs and leave others anyway, so there shouldn't really be any more techs.
eleazar wrote:* I realize your "Biology X" type theories are just placeholders to organize things, but if displayed in a tree, you'll need something to bind things together and show the relationships between the multiple choice techs, and for attaching your branches and crosslinks, so that brings us up to 4 times as many items.
Again, 4 times as many as what? What I have in my tree - a bunch of theories bound to applications and linked together - is exactly what the other tree had. The only difference is that mine is bigger because it's meant to emulate the tech tree of the final, playable game. The size of the tech tree isn't the point of what we're discussing. If anything, I have fewer "objects" in my tech tree compared to a hypothetical version of the current tree with the same number of applications because of the much simplified relationships between techs reducing the number crosslinks between techs.
eleazar wrote:* You'll need some way to identify techs that aren't researchable because you chose another option, doable but it's something else that moves further into "more complicated".
They can be made a darker shade....? That doesn't sound like something that would be a problem to understand or implement, especially since the player is expected to know that he can only choose one application from each theory when he starts the game (and if he doesn't, it's not the type of thing that's hard to discover).
eleazar wrote:* You loose one of the nice things about a multiple-choice-column (MCC)-- it can be displayed more simply and clearly than a tree , in a list with columns.
True. You can accomplish something similar however, by selecting only "researchable" techs in the tech tree screen, which would show you only the tech levels that you can research the applications of next - the versatility of our UI makes it fairly easy to see simply and clearly what your next research options are.

You still haven't really described what's confusing about the tree I provided, which should be pretty simple to navigate once the player know how it works. It's basically 7 columns, with 2 of the columns having occasional branches and interdependencies with the other 5 columns, which is obviously a little more complicated than the basic MoO2 columns, but not confusing, IMO, by any means.
Warning: Antarans in dimensional portal are closer than they appear.

User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: Tech Tree Discussion: FO vs. MoO2

#21 Post by eleazar »

Bigjoe5 wrote:
eleazar wrote:* If you have on average 3 choices, all other things being equal, there will be 3 times as many items in the tech tree
3 times as many as what? 3 times as many as if the player is expected to research every tech? I don't think that was going to be the case to begin with. Overall, we want there to be a way to make players research some techs and leave others anyway, so there shouldn't really be any more techs.
Well that's sorta true, but in practice it doesn't work.

First of all, the amount of our plain tree that would be unused by a tech-focused player is undefined, but i really doubt it is 2/3rds.

Lets assume the average tech-focused player will be able to research 50 techs before somebody wins. No matter if it is a tree or a column system (or both) we need more than those 50 (or for a 3-choice column 150) because we don't want the above average tech-focused players to research everything possible, and then research will be useless, which would be a let-down. It also helps keep the game more continually engaging if the "ultimate" or "ultimates" are usually out of reach.

So while MCCs could replace the need for alternate choice branches in a tree, it doesn't replace the need for an extra buffer of techs beyond what players will normally reach. To keep exceptional players equally busy, and MCC needs 3x as many techs in that buffer.

Bigjoe5 wrote:
eleazar wrote:* You loose one of the nice things about a multiple-choice-column (MCC)-- it can be displayed more simply and clearly than a tree , in a list with columns.
True. You can accomplish something similar however, by selecting only "researchable" techs in the tech tree screen, which would show you only the tech levels that you can research the applications of next - the versatility of our UI makes it fairly easy to see simply and clearly what your next research options are.
Not at all similar.
True you can narrow the displayed techs to only those that are immediately researchable, but it's a lousy UI, since it just shows a single column of techs in a large field of black, unorganized in any discern-able way.

As much as i find it disconcerting to praise any aspect of MoO3 GUI, they got these essential concept right. This is how a MCC should be displayed: it's easy to see where you are in the tree, which techs are mutually exclusive, and what comes next. A very much superior way to display a multiple-choice tech system.
tech columns.jpg
tech columns.jpg (243.99 KiB) Viewed 1263 times

User avatar
Bigjoe5
Designer and Programmer
Posts: 2058
Joined: Tue Aug 14, 2007 6:33 pm
Location: Orion

Re: Tech Tree Discussion: FO vs. MoO2

#22 Post by Bigjoe5 »

eleazar wrote:Well that's sorta true, but in practice it doesn't work.

First of all, the amount of our plain tree that would be unused by a tech-focused player is undefined, but i really doubt it is 2/3rds.

Lets assume the average tech-focused player will be able to research 50 techs before somebody wins. No matter if it is a tree or a column system (or both) we need more than those 50 (or for a 3-choice column 150) because we don't want the above average tech-focused players to research everything possible, and then research will be useless, which would be a let-down. It also helps keep the game more continually engaging if the "ultimate" or "ultimates" are usually out of reach.

So while MCCs could replace the need for alternate choice branches in a tree, it doesn't replace the need for an extra buffer of techs beyond what players will normally reach. To keep exceptional players equally busy, and MCC needs 3x as many techs in that buffer.
I've addressed that issue to a certain extent in my tech tree. The ultimate Biology tech, Evolutionary Mutation, requires RP for its use. Likewise, I mentioned that the ultimate Construction tech, Stargate, might also require RP with each use, to act as a late-game RP sink. In general, I would say that it should not be terribly common for players to research as many techs as they can, but if they do - and some players certainly will regardless - I don't think it makes sense for research to become obsolete at that point.

Basically, the problem you're describing is a wider one, and the solution isn't to make more techs researchable, but to give the player something else to do with his RP once he's done researching all the techs he can.

Bigjoe5 wrote:Not at all similar.
True you can narrow the displayed techs to only those that are immediately researchable, but it's a lousy UI, since it just shows a single column of techs in a large field of black, unorganized in any discern-able way.

As much as i find it disconcerting to praise any aspect of MoO3 GUI, they got these essential concept right. This is how a MCC should be displayed: it's easy to see where you are in the tree, which techs are mutually exclusive, and what comes next. A very much superior way to display a multiple-choice tech system.
If you want to see one category at a time like that, then the tree UI is probably just as good if not better. First of all, the screenshot you posted did not make it all that clear which techs are mutually exclusive (even if techs in MoO3 were mutually exclusive, which they're not) because there are two rows of level 10. In tree view, all applications of a theory would be clearly shown branching off from it, so there would be no doubt about which techs are mutually exclusive (provided the player knows techs are mutually exclusive). Furthermore, in the tree UI, you can turn on both the category you're looking at, and the Cross-Category techs, which shows those techs branching off at the appropriate point on the tree - a feature which MoO3, with its complex category dependencies (though one couldn't tell such dependencies exist from the screenshot you posted), lacks.

Or, if you wanted to see all the categories at once, but not be confused about exactly where you are in the tree, it might be a good idea to turn on only Researched and Researchable techs, which shows clearly in each category where you are, and what your options are for researching next (better code for drawing the tech tree so that it actually looks logical would help with this too... Ideally if there's a string of theories leading one by one to to the most advanced theory in the category, they should be placed in a straight line).

Anyhow, it would probably be possible to implement the sort of tech tree I'm talking about without a lot of changes to the code, and while leaving the remaining functionality intact. What we could do is:

1: Add an optional field in the tech descriptions - "max_dependants" or some such - that defines how many of the immediately dependant techs can be researched by the player, before the tech is considered "exhausted" and no more immediately dependent techs can be researched. If the field is left out, the tech would act normally as it would now, with an arbitrary number of dependent techs being allowed.

2: Modify the researchcost and researchturns fields to accept some value that will make the tech permanently unresearchable.

With these modifications, we could make all theories unresearchable and have any application in the previous tech level unlock the next theory. (Cross-Category techs would be more complicated, but still doable with the current effects system). Then, each theory could be given max_dependants of 1, to make all its applications mutually exclusive.

Basically, it would be a way to let anyone create a tree with mutually exclusive applications, while still retaining all the current functionality if we eventually don't want to use mutual exclusivity (in which case modders can make their own MoO2-style or hybrid trees). And as someone said recently, being able to have it in the game would provide much stronger grounds to critique/praise/insult the system than just seeing a sample. 8)
Warning: Antarans in dimensional portal are closer than they appear.

User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: Tech Tree Discussion: FO vs. MoO2

#23 Post by eleazar »

Bigjoe5 wrote:
eleazar wrote:First of all, the amount of our plain tree that would be unused by a tech-focused player is undefined, but i really doubt it is 2/3rds.

...

So while MCCs could replace the need for alternate choice branches in a tree, it doesn't replace the need for an extra buffer of techs beyond what players will normally reach. To keep exceptional players equally busy, and MCC needs 3x as many techs in that buffer.
I've addressed that issue to a certain extent in my tech tree. The ultimate Biology tech, Evolutionary Mutation, requires RP for its use. Likewise, I mentioned that the ultimate Construction tech, Stargate, might also require RP with each use, to act as a late-game RP sink. In general, I would say that it should not be terribly common for players to research as many techs as they can, but if they do - and some players certainly will regardless - I don't think it makes sense for research to become obsolete at that point.

Basically, the problem you're describing is a wider one, and the solution isn't to make more techs researchable, but to give the player something else to do with his RP once he's done researching all the techs he can.
That sounds like a good idea, but i don't think it's the whole solution. I would still want a buffer to maintain something to strive for, if a smaller one with RP-burning devices.

But it doesn't invalidate my point. A 3-option MCC has more items than an ordinary tree, unless that tree has 200% more techs than tech-focused player is likely to research. And IIRC quite a few of your examples have more than 3 options. And a tree grows increasingly clumsy to use the more techs you have in it. My screen is ~1900x1200 and still i find our current tree a pain to use, generally with a whole lot of panning and zooming to find anything-- even with filters.

I'm interested in ideas that lead to a smaller, simpler tree, (like RP-burning devices), not ideas that lead to a bigger, more complicated tree, even if the increase is small. Getting rid of do-nothing theories would be a great way to simplify, but with your hybrid tech system we can't do that, because they are needed to bind the multiple-choice groups together.

Bigjoe5 wrote:
eleazar wrote:As much as i find it disconcerting to praise any aspect of MoO3 GUI, they got these essential concept right. This is how a MCC should be displayed: it's easy to see where you are in the tree, which techs are mutually exclusive, and what comes next. A very much superior way to display a multiple-choice tech system.
If you want to see one category at a time like that, then the tree UI is probably just as good if not better. First of all, the screenshot you posted did not make it all that clear which techs are mutually exclusive (even if techs in MoO3 were mutually exclusive, which they're not) because there are two rows of level 10...
My point was not that MoO3's tech system was good, but the basic arrangement you see in that screenshot-- for an MCC with no cross-links-- had much superior ease of use, and comprehensibility.

If you want a MCC without branching or cross-linking, i have nothing to say against it. I probably favor it a little now that we've discussed the different types.
Of course there's the additional programming work to consider, though i've gotten the impression that making the tree untangle itself in an intuitive manner was pretty complicated, so a new list GUI might not be any more work than making the tree behave.

Bigjoe5 wrote:And as someone said recently, being able to have it in the game would provide much stronger grounds to critique/praise/insult the system than just seeing a sample. 8)
Well, the difference there is it only took me an hour or so to make the binary stars and upload them, and will only take a couple minutes to remove them, and nothing else was effected. I believe it would require a much larger amount of work to test your idea.

User avatar
Bigjoe5
Designer and Programmer
Posts: 2058
Joined: Tue Aug 14, 2007 6:33 pm
Location: Orion

Re: Tech Tree Discussion: FO vs. MoO2

#24 Post by Bigjoe5 »

eleazar wrote:If you want a MCC without branching or cross-linking, i have nothing to say against it. I probably favor it a little now that we've discussed the different types.
I don't really have any objection to this.
Warning: Antarans in dimensional portal are closer than they appear.

Zireael
Space Dragon
Posts: 429
Joined: Mon Aug 15, 2011 5:33 pm

Re: Tech Tree Discussion: FO vs. MoO2

#25 Post by Zireael »

I think a list GUI like the one in the screenshot would be a good idea!

SowerCleaver
Space Squid
Posts: 54
Joined: Wed Oct 25, 2006 8:59 pm

Re: Tech Tree Discussion: FO vs. MoO2

#26 Post by SowerCleaver »

I loved both MoO2 and Civ IV, and agree with Bigjoe5 that the "mutual exclusivity" feature of MoO2 tech tree/column made the game much more interesting. It differentiated each civ much more than it would with a Civ IV style tech tree. In Civ IV, because of tech trading, everyone on the same continent eventually had the same techs and it was just a matter of who gets the cutting edge techs a little bit faster than everybody else so that they can hold strategic advantage. On the other hand, in MoO2 one civ may have very advanced missle technology but lagging beam tech. All in all, the uniqueness of each civ in MoO2 enhanced replayability.

Also, I note that there was a variation on the "mutual exclusivity". Psilons were "Creative", so they got all 3 "applications" under one "theory" at the same price. The ant race was "Uncreative", so they could not choose which one of the application they will research - it was random for them.

What was the rationale for having two layers of research - "Theory" and "Application"? Is it to permit a player to entirely skip all applications of a lower-ranking theory and shoot for higher-ranking application earlier? If so, I would think the system is also worthwhile to keep.

We can also make the exclusivity not a systemic must but a practical necessity. After you research one application under a theory, it will take x times the normal RP to research another application under the same theory. The value of x can be varied depending whether the race has "Creative" or "Uncreative" pick (there can be more gradation of such a pick). We can also make the researching of sister applications possible only after a higher theory was also researched. That way, a research-focused civilization has a way of obtaining sister applications when no civ in that game chose to research it as their first choice.

User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: Tech Tree Discussion: FO vs. MoO2

#27 Post by eleazar »

SowerCleaver wrote:What was the rationale for having two layers of research - "Theory" and "Application"?
I don't know. That discussion was before my time on the previous forums that died.

I've been working to minimize the number of theories (by combining them with an obvious application) since they really inflate size and complexity of the tech tree.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 13587
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

Re: Tech Tree Discussion: FO vs. MoO2

#28 Post by Geoff the Medio »

SowerCleaver wrote:What was the rationale for having two layers of research - "Theory" and "Application"?
An early designer or designers liked Hearts of Iron or Europa Universalis (or possibly their sequels), which reportedly had a research system with theories and applications as distinct techs or steps in researching techs. Based on limited reading of relevant wikis, the actual structure of research and the meaning of theory and application in those games wasn't much like what FreeOrion has, though.

User avatar
OllyG
Space Kraken
Posts: 151
Joined: Wed Jul 20, 2011 12:03 pm

Re: Tech Tree Discussion: FO vs. MoO2

#29 Post by OllyG »

I like the idea of being able to reach the end of a single category without having to reseach anything in the others at all. Being able to specialise so much will be one way to allow different empires to have very different strategies.
But I do like the cross linking, as it makes reseaching combinations of two or more categories more interesting.

I think that a good compromise would be like Bigjoe5's example. Having every level of the tree include mutually exclusive techs is a bit too much for me though. I do like examples of ship parts or buildings which are mutually exclusive (such as asteriod hulls and neutronium ship parts as in v0.3.17 - but this may be an accident.)
Each category should have a 'trunk' made up of theories, with branches made of Applications. Refinements should lead from some Applications, but should also require a more advanced theory. Cross linking should be done by having some Applications require more than one theory (from different categories - not always at the same level). Theories should only require other theories from the same category.
Each category should in effect be a pillar - as in MoO1&2. but Applications could require two or more pillars. Refinements should be harder to get than Appplications, maybe they should require more categories of theories or even applications.
Keeping the tree simple in appearance (by keeping the theory structure simple) will make it easier to 'read'. When the tree is displayed the theories should be in a straight line (or in a simple zig-zag for aesthetic reasons). Letting Applications and refinements have more complicated prerequsites makes the tree more interesting and lets players have fun working out what routes they need to take through the tree. A player should be able to have a well balanced Empire by only reseaching two or three categories, so each category shoudn't be focused too narrowly.

When the game is completed and balanced the research costs should be adjusted so that it is virtually impossible to research everything, even when trading and espionage are taken into account. I think that trading or stealing theories should be limited or even impossible, so that each empire still has to research along the tree even if they have amazing spies. (Refinements should be harder to steal than Applications, to make them more valuable).

If a single category can be exlusively reseached it could be an easy target to give the AIs before they are fully developed. Each one could choose a (random) top level tech to aim at.

I would like to see categories such as Cybernetics, Biotechnology and Psionics (maybe even Force Fields, Fusion or Gravitronics). These have a theme, but could affect different parts of the game. The learning or knowledge Category now has many psionic technologies.

In conclusion.
The main structure (the theories) of the tech tree should be simple (and linear).
Applications should sometimes be crosslinked.
Refinements should not just be repeated applications with better numbers and they should be harder to get by research or stealing.
Reseaching everything shouldn't be possible.
Using everything should be even less possible!

User avatar
Krikkitone
Creative Contributor
Posts: 1559
Joined: Sat Sep 13, 2003 6:52 pm

Re: Tech Tree Discussion: FO vs. MoO2

#30 Post by Krikkitone »

OllyG wrote:I like the idea of being able to reach the end of a single category without having to reseach anything in the others at all. Being able to specialise so much will be one way to allow different empires to have very different strategies.
But I do like the cross linking, as it makes reseaching combinations of two or more categories more interesting.

I think that a good compromise would be like Bigjoe5's example. Having every level of the tree include mutually exclusive techs is a bit too much for me though. I do like examples of ship parts or buildings which are mutually exclusive (such as asteriod hulls and neutronium ship parts as in v0.3.17 - but this may be an accident.)
Each category should have a 'trunk' made up of theories, with branches made of Applications. Refinements should lead from some Applications, but should also require a more advanced theory. Cross linking should be done by having some Applications require more than one theory (from different categories - not always at the same level). Theories should only require other theories from the same category.
Each category should in effect be a pillar - as in MoO1&2. but Applications could require two or more pillars. Refinements should be harder to get than Appplications, maybe they should require more categories of theories or even applications.
Keeping the tree simple in appearance (by keeping the theory structure simple) will make it easier to 'read'. When the tree is displayed the theories should be in a straight line (or in a simple zig-zag for aesthetic reasons). Letting Applications and refinements have more complicated prerequsites makes the tree more interesting and lets players have fun working out what routes they need to take through the tree. A player should be able to have a well balanced Empire by only reseaching two or three categories, so each category shoudn't be focused too narrowly.

When the game is completed and balanced the research costs should be adjusted so that it is virtually impossible to research everything, even when trading and espionage are taken into account. I think that trading or stealing theories should be limited or even impossible, so that each empire still has to research along the tree even if they have amazing spies. (Refinements should be harder to steal than Applications, to make them more valuable).

If a single category can be exlusively reseached it could be an easy target to give the AIs before they are fully developed. Each one could choose a (random) top level tech to aim at.

I would like to see categories such as Cybernetics, Biotechnology and Psionics (maybe even Force Fields, Fusion or Gravitronics). These have a theme, but could affect different parts of the game. The learning or knowledge Category now has many psionic technologies.

In conclusion.
The main structure (the theories) of the tech tree should be simple (and linear).
Applications should sometimes be crosslinked.
Refinements should not just be repeated applications with better numbers and they should be harder to get by research or stealing.
Reseaching everything shouldn't be possible.
Using everything should be even less possible!
I like that general setup..
"Pillars" of theories with Applications branching off of them.

I'd like to see researching everything Possible, but Ridiculously impractical.

So something like... Theory: Learning Level 3... has 3 applications, The first costs 100 RP, the second will cost 300 RP, the third will cost 1000 RP. (or the first requires Learning Level 3, the second Learning Level 5, the third Learning Level 8. )

Post Reply