Research Mechanics in Strategy Games

I’ve always been bugged by how research in 4x style games always ends up being shopping (Civ, Dominions, AoW, etc.) You know exactly what’s going to be available ahead of time, exactly what it will cost, and exactly what you need to do to get what you want. Competive play in such games frequently centers on a plan to get to some “power” technology as soon as possible, requiring extensive research before a game even starts. There’s some variation on the theme (e.g. Moo2) but the trend is pretty firm.

Surely something better (or at least different!) is possible… I’d like to see something which captures the uncertain results of actual research, perhaps with research as a subgame based on puzzle or clue-like mechanics. Are there examples of such games I’m missing? Is anyone else even interested in this other than me? If so, what sorts of research tweaks would you like to see?

Dark Cloud 2 was a step in the right direction, I think, with its photos and item-combining. Can’t imagine how to apply that to a strategy game, though.

I’d missed Dark Cloud 2 (don’t have a PS2). I’m not sure how well that would play though, as it seems great for a game you play once, but on a second run through you’d already know all the combinations (or just look them up on gamefaqs).

Perhaps if the “idea” parts needed for an “invention” varied from game to game? You’d have a general idea of what ideas to pursue to get to some goal, but wouldn’t know for sure. This would be hard to do though, and still have the combinations make sense.

Or perhaps if each step of research let you select one of a set of random ideas? You’d know what was needed to get a specific goal, but the outcome of your research would wander, as unexpected goals might suddenly become nearly finished after a random selection.

You could divide research into various fields, and then let players put their research efforts into a specific field, but not give them control over what specific discoveries they make. So if a player puts research points (or whatever) into weapons technology, there might be a number of techs that he could discover, and the game picks one randomly (alternately, you could have greater or lesser chances to discover certain techs based on discoveries you have made in other fields). That tech could lead to a number of other techs while maybe closing off some others, and so on. But the player doesn’t choose specific things to research, just a general field of study.

The tree would be complicated, but most of that would be hidden from the player. From his perspective, he’s getting a different tech tree with every game, so even if he learns what techs are out there, he doesn’t know what he’s going to discover this time around. It might not always be perfectly balanced, but who cares? What it would do is force players to adapt their strategies to their scientific discoveries, rather than the other way around, which in turn makes research function more like it does in real life.

There was a game called (I think) “Alien Legacy” that had a blind research system. It was a real-time 4x strategy game, as I recall, and you would assign guys to research in general areas, but wouldn’t know how long it would take or what you would get for it. Periodically, they would go “Eureka!” and you would get a report on what new tech you had discovered and what it would allow you to do. Alien Legacy was a bad enough game that I didn’t play many games, but my assumption is that the tech tree was basically static and that once you had played a few times, you would know generally what was where even without a tech tree laid out for you. [Edit - this is very similar to what Ben is proposing]

Ascendancy, IIRC, had a set tech tree, but it was covered with a sort of fog of war so that you wouldn’t know what was next on the tree. Again, though, if you played enough games (although who would with that piece of shit’s crappy AI) you would eventually learn the tech tree anyway.

Which I think is the main problem. Having a set tree that is simply hidden from the player isn’t much of a solution, because it only works the first few times. On top of that, it means the difficulty of the game varies a lot depending on how much you’ve played it (which is always the case, of course, but this makes the problem significantly worse). That makes multiplayer games unbalanced and also makes it somewhat difficult to balance the singleplayer game–if you make it easy enough to win when you don’t know what to expect, it will be too easy once the person learns the tech tree.

The solution is what you suggested in your second post–having some sort of system where the tech tree is random or semirandom. But as you identified, the problem there is making it random but still sensible so that a player will feel immersed in it and will be able to make reasonable guesses about what strategies to pursue. Pretty tough, and most players would probably rather the devs spend time designing better AI or a larger regular tech tree.

Edit: Ben’s idea actually takes that in a way that sounds workable. Someone make a game with that.

Interesting question but I wonder how the mechanic would work.

You could, I suppose, invest research along a couple of chosen lines and the final payoff would be non-specific and random. Maybe no bonus, or longer research times on the negative or a bonus that ranges anywhere from +1 to +n.

In Victoria, specific payoffs from discoveries would not fire immediately. So you could discover the concept machine tools but not figure out how to mass produce them until a specific technology fired.

Civ IV will go the other route - give you multiple roads to the same tech - but the purpose is to create a little more variety. There will still be the predictable end state.

The tech transparency is important for strategy gamers, though, because by and large they like to have their choices laid out for them. To invest time and money into a research path and find that my scientists had discovered magic eggplants instead of a new superweapon would be cute for a while, but would eventually make the game more about the tech race than the game already is.

I’ll think a little more about this, though. Tech in strategy games has reached a default state for the most part with very little in the way of innovation in gameplay.


Wasn’t SMAC like this? It had several areas you could emphasize, and then technologies would be randomly chosen out of that branch. I also thought the branches overlapped slightly as well.

SMAC was (optionally) like what Ben describes, although the variety of what you could get at any one point was limited, and it didn’t have any of the finesse he suggests.

I liked SMAC’s system quite a bit, but in the end it felt like it just traded researching a specific tech research order for researching a statistically optimal research order. If I didn’t get what I wanted, I just tended to keep researching until I did rather than change my plans.

[Edit]Maybe combine this with something like MOO2’s approach? Random low level techs you didn’t get could become higher level selections you’d get later on in the tree, rather than just by repeating the same research? Say for example techs were designed in groups of 3, then randomly sorted into Levels 1,2,3 for each player. The tree links would remain the same from game to game, but what was in each slot would varry.

I’d forgotten about Alien Legacy! The research and the way it tied into the story was alot of fun, although the game itself was dull and tedious enough I never finished it. I think having research tie into a game’s events and in game locations would be cool (and doable).

True, although I think there is something to be said for having some choices laid out and others forced on you. A good strategist adapts to the circumstances. A lot of gamer strategists are just really good at pattern optimization, which is something that’s been around since the dawn of gaming, I guess.

To invest time and money into a research path and find that my scientists had discovered magic eggplants instead of a new superweapon would be cute for a while, but would eventually make the game more about the tech race than the game already is.

I don’t think that would be a good thing. The idea is to give the players at least a general idea of what they might discover–if the player is researching weapons technolgy, then they should discover something weapons-related, not magic eggplants (unless those are a weapon). And I don’t think any of the techs should be useless, though some could be more useful than others. You could even let them nudge the odds a bit; maybe if they put a certain percentage of their research into a field, they have a higher chance of discovering a more powerful tech in that field.

Essentially, the tech tree would be a bit like a hand of cards, and you’d have to play what you were dealt. But it would be less random than cards; the player would have some say over what type of cards he gets, and certain cards would be prerequisites for others. I think it would be a lot more interesting than current, static systems.

I really liked Ascendancy’s tech system the first time I played. I thought it was cool that they tried to add a sense of discovery to making discoveries. As Ry said, however, the hidden tree was pointless after you had played a few times. Unfortunately, so was the rest of the game.

What irks me is that research usually has no relationship to what is happening in-game. What I mean by this is that most research follows the light armour-medium-armour-heavy armour paradigm, when in reality, the first armour created was actually the heaviest. Only battlefield experience would determine if it was practical. In any 4X game, I can have the most awesome techs possible without ever having been in a conflict. (Actually, being in a conflict would divert resources from research in a game, whereas IRL wars tend to be great things to develop technology in). I would like to see technology open up in the areas that you play your game toward. You’ll have to actually fight with your air units before you can develop those jets, mister.

That is an awesome idea.

I definitely agree interaction with the rest of the game is sorely lack from research. Other similar ideas:

  • Research about secrets about the game world – pure knowledge, but with practical value. There’s an omen in the sky – what does it mean? Just who is this lich who came out of the desert dunes, what’s his weakness? Where is Shangri-La?

  • research requiring or aided by specific objects, from alien precuror tech, to books from old ruins, to body parts of magic beasts.

  • Research aided by seeing anothers tech in use first hand, rather than by “trading” technology.

  • Research flavored by the researcher. Perhaps many nations have a “Fireball” spell, but some could be bigger, others could be hotter or longer range, etc.

One interesting angle is research diffusion over the map. Unless you’re talking about guarded, directed, government research most practical inventions come up as a result of private individuals in the general population solving problems. In Crusader Kings you can specify areas you’re interested in improving but far more usefully you can develop roads and markets that improve the chances of technology developed somewhere else travelling to your lands. The Doge of Venice is going to be reeling in technologies faster than any one land can develop them while off in Wallachia you’re not going to have access to that kind of transmission of knowledge. But save your gold and build up clerical and educational facilities and that will help compensate a bit.

I do seem to recall playing some games with random feeling advances which made getting them feel a bit like a pleasant surprise. Might Emperor of The Fading Suns have been along these lines? You could build labs that would get research points and the player could distribute those points across broad categories of research. Which actual subcategories opened up was fairly random, I think, or it could be that your direction wasn’t absolute - more of a suggestion.

At any rate one interesting twist is the Inquisition. Here politics played a role. Factional leaders could petition the Patriarch of the Holy Church on Terra to outlaw certain kinds of technologies as heretical or to change a heretical technology back to something canonically acceptable. There’s a whole mini-game around the different religious factions, proselytizing, and gaining control over the Patriarchate yourself but most competition came simply from bribing an uncontrolled Holy Father or just grabbing as many churchs of the primary sect as one could on different worlds to get the votes needed for control. This was a side game as the main arena was the competition to become Emperor but could effect it. The Inquisition has quite a fleet and control of what’s legal to research can be handy.

Of course heretical tech isn’t just blocked out. You can set up labs and research it anyhow but if the lab is discovered you’re laying open for a big religious smackdown. One, actually useful, tactic was finding a remote or hostile place on a nasty world to build your secret labs. The Inquisition seemed to spend more time searching settled and active planets.

Over time almost all the useful tech becomes heretical. And since you need lab points not only to research but to maintain current (not prerequisite) technologies having the Inquisition burning down your labs can really put one at a big military disadvantage. But those handful of elite or bizarre units you made before you got shut down are still with you and others could be discovered as lost caches of the old Republic here or there.

Try to name one significant technological development that was not government funded and/or directed. The situation is excactly the opposite.

Obviously not talking about tiny little tweaks here and there, but technoligical advances like semiconductors, antibiotics, computers, tool standardization etc. Even private inventors are operating on government funds 99.9% of the time.

So in fact the (perhaps ironically, seeing as how many of them are free-market advocates) the modeling systems used in these gamesfrom an abstact level are actually realistic - although in terms of the gameplay flawed in all the ways jasper explained

Yes, SMAC optionally (and most multiplayer games force that option) have the technology as a “deck of cards” as Ben described it, the only flaw was that it was too realistic (technolgies werent balanced, some discoveries took just as long as others but were only marginally useful - while thats true in real life it makes the game a little too dependent on luck)

I agree with Frazier’s ideas about RPG style technology based on experience particularly for war related techs. That woudl probably solve a lot of the issues raised here

Great thread and btw in related news, SMAC2 may be in development according to the Firaxis website

I really liked the Empire of the Fading Suns research vs. Inquisition setup in principle, unfortunately the AI behind it was broken. :-( Everything from the heretical techs, to the flame thrower toting inquisition, to the Church Battlefleet, to the upkeep for research really exuded flavor. It was such a pity that game was in general so broken; I really hope someday that someone takes another shot and really does Fading Suns justice.

I’m enough of an EFS fan that I’ve put some thought into a research subgame for an EFS remake: Research would work as a puzzle subgame where each technology is made up of tetris like patterns of 4 or so tiles. These would be randomly laid out on a grid face down, such that you could not tell which tile belonged to which pattern. Different parts of the grid would tend to have different kinds of technology. Research would involve peaking at tiles, with new techs becoming available when all their tiles are revealed. This being Fading Suns, you’d also have to pay upkeep for each tile you wanted to retain use of. More advanced techs would be in seperate grids underneath like strata, and you’d have to research the tiles above them before you could drill down to them. There’s alot of possible finesse on top of this, but that’s the basic idea.

Try to name one significant technological development that was not government funded and/or directed. The situation is excactly the opposite.

Obviously not talking about tiny little tweaks here and there, but technoligical advances like semiconductors, antibiotics, computers, tool standardization etc. Even private inventors are operating on government funds 99.9% of the time.[/quote]

The wheel. Indoor plumbing. Traditional remedies and medicinal techinques. Sure, the guy down the street isn’t going to build a Harrier VTOL in his garage but depending on the historical era he might come up with a Conestoga wagon. You don’t think it was great government grants that brought gunpowder or stirrups to Europe do you? Concentric castles? This was diffusion of knowledge, my lad.

In MoO2, there was a racial trait “uncreative” which assigned you a random technology in your chosen research area (i.e. randomly assign you 1 of 3 discoveries).

As mentioned, SMAC’s system was (optionally) another one of these semi-random hybrids, except that people really tried as hard as possible to steer towards a few key technologies like Doctrine: Air Power (aka “beat people up with planes”).

  • Alan

Alright N00blers, I have had the solution for this problem for years and its depressing no game maker has picked up on it yet because I want to play such a game. Originally this was intended for RTS type games, but it can work in any strategic game.

The basic premis is that there are not any fixed techs or units in the game. There are provided techs and units you can use so you do not need to make everything, but you are not limited to what the game ships with.

Now how does this all work? With a point system, say like Hero systems or GURPS. You buy technologies with points, you make units with points, you make improvements with points. Players decide what powers they want in some object, such as armor, hitpoints, speeds, etc. and then the game computes the research cost and production cost for you based on formulas. Similarly you can make ‘parts’ like this, such as heavy lasers or turbo lasers. You reasearch these, and then when you are making a ship you can staple on some reseahed weapon and its just added to the cost.

Certain technologies and power advantages have others as prerequisites. For example, before you research dimensional travel, you need teleport, and to get teleport you need matter-enery tranmutation, etc… This is your tech tree. Also you could include a tech level (to mark technological epochs) which would forbid certain technologies until you reached the preqrequisit epoch. This might also can be used to give discounts on lower-epoch units, or enforce a rule allowing you to make stuff from +1 epoch (you capture some tech, you can make it, but do not understand how it works) at an increased cost and size.

Throw in a graphic editor that featuers some modular graphic design and you are good to go.

Now my favorite part of this whole system is in multi-player when you see an enemy unit and you have NO IDEA what it does. It would add a whole new level of these games. You go into someone’s base, you have no idea what structure does what. You might be attacking a building of minor or cirtical importance. Finally in game, if you need a unit to counter some other guys units, you can always design one, research it, and build it.

Isn’t that actually very similar to SMAC?