Rethinking the Space 4X - Research

Make using stuff related to some area of research speed up the research. If you mine a lot, you will propably get new ideas for how to improve it faster than someone who gets his resources by robbing other people.

To me, the current system is selecting which results I want and devoting resources and time until I acquire them.

And that's how it should be. Research's purpose is not to be enganging, it is to provide strategic choices to the player. The other parts of the game should be enganging, not research as a process. The fun of research comes from agonizing choices of resource allocation and choice of the next technology. Making research process more "interesting" is in conflict with strightforwardness of the decisions available.

Take a look at Sword of the Stars' system. It incorporates a lot of the features you're describing for research.

And that's how it should be. Research's purpose is not to be enganging, it is to provide strategic choices to the player. The other parts of the game should be enganging, not research as a process. The fun of research comes from agonizing choices of resource allocation and choice of the next technology. Making research process more "interesting" is in conflict with strightforwardness of the decisions available.

In a standard 4X game set-up I agree with you completely -- research is not in itself the focus of the game, and should therefore be easy to interact with and unobtrusive rather than engaging. I think the original poster is trying to make a different type of game though, and that's potentially an interesting goal. It is a good point that the rest of the game will probably need to be similarly re-designed to match though -- simply slipping this theoretically more engaging research system into an otherwise normal 4X game will probably result in a game where research detracts from other game-play.


How about some sort of system where the tech-tree is generated procedurally such as to give different choices from game to game and (hopefully) no easily achievable strategy for picking a "most affective" research method to match any given play-style. Both the technologies available and the requirements for unlocking them would differ from game-to-game, and in combination with different ways of boosting research we could possibly have a more interesting system -- although again this would probably detract from a standard 4X experience where players are tested on their ability to effectively master a known/fixed tech tree.

How about some sort of system where the tech-tree is generated procedurally such as to give different choices from game to game and (hopefully) no easily achievable strategy for picking a "most affective" research method to match any given play-style. Both the technologies available and the requirements for unlocking them would differ from game-to-game, and in combination with different ways of boosting research we could possibly have a more interesting system -- although again this would probably detract from a standard 4X experience where players are tested on their ability to effectively master a known/fixed tech tree.

I like this idea. I definitely think that a fixed tech tree progression detracts from the interest and challenge that a dynamic system could generate and preserve from game to game. If I were going to adopt this idea I'd have to be careful though. Linking two areas of research might result in some fridge logic that breaks willing suspension of disbelief. An explanation of how one area relates to another could allay the onset of disbelief, but would reduce the flexibility of a dynamic system somewhat.

Another thing I think I may want to preserve is some measure of predictability. A somewhat predictable system tests the player's predictive capabilities and rewards good prediction. Of course, freshness and predictability are conflicting requirements, so I would have to work hard to achieve a pleasing balance.

Just throwing random ideas out there:

Maybe there could be (probably optional?) mini-games where the player's success could speed (i.e. you get results faster) or provide additional benefits (i.e. you get a larger stat-boost, or higher damage output, or even access to additional items from a given tech) to research? The games could be research themed along the lines of operating or configuring lab equipment, combining different chemicals, collecting sample items (hidden object?), etc.


Some games (Alpha Centauri being one example) have a "blind research" mode, where the game otherwise plays as normal but you can't pick a specific tech to research each time -- instead you choose one area (conquest for example) and the game randomly selects technologies from that area to research. In the implementation in Alpha Centauri you can prioritise each different area so that you will still discover tech in your lower-priority areas but will have a much higher chance of making discoveries the the area you've focussed on.


What if instead of simply providing new units/buildings/whatever a new tech provided you with a number of prototype things you could build, and then only after testing these in the field do you get access to the regular items, which may operate slightly differently to the initial prototypes. Some prototypes might pan out as items that are impractical or useless when applied in the field. Players would then have to balance the risk of wasting resources on faulty or useless items against the potential benefits of getting cool new stuff. I think I've seen something similar to this in a game before, but I can't recall which one.

Alien Legacy had interesting approach to research in that the topics you could research were driven by discovering items during planet exploration and responding to in game events. You also didn’t know what the outcome of the research was until after you did the work and not every piece of research lead to any benefit; it might only allow a future research topic when combined with other tech. So finding a unusual crystal on a planet allowed you increase the power of your weapons.

The other side was the reactionary research in which you could research topics related to in game events. So when you met a race of aliens that used psychic attacks to kill your pilots it enabled a new research topic looking into ways to defend against those attacks.

A more evolved system like that could be a lot of fun in 4x game. Sending off a research mission to black hole might unlock new technology, discovering a planet with crystal formations might allow to build crystal based ships. Trade routes with another species might allow you research a new way of growing crops.

I would have difficulty making good mini-games that are part of a good research system. In general, the only mini-games I remember being fond of are the ones primarily in the Legend of Zelda and Final Fantasy series.

Blind research is an idea, but it seems much like the standard system. The big difference is the reduction via simplification of player decision making.

I'm considering some sort of prototyping system.

I definitely want the system to have emergent properties based on other features of the game. I'm thinking about having a reverse engineering system, and one of the things that has always bugged me about reverse engineering is that scientists can get this alien technology and figure out how to use it. In terms of realism, this doesn't make sense. A human species would have difficulty understanding and reverse engineering a fish species' ships. There are difficulties in terms of physical differences and psychological differences. Of course, there is "good" realism and "bad" realism. I try to avoid the bad kind. I do think there is opportunity in implementing some barriers. I see it as a potential to foment player-player or player-computer interaction. I'd like it if I could make friends with an alien species (or subjugate them) and either get their scientists to reverse engineer tech they know, or get them to operate it on my behalf.

I want to start working on prototype code to try out some ideas. I'm having trouble getting started though because other aspects of a 4x can determine how research works. E.g. in a turn based points are allocated each turn, whereas in a real-time they are applied constantly. I'll either have to create a concrete implementation and refactor if I need to change, or spend time up front to design an abstract research system with a well defined interface.

I've given this some more thought, and if I wanted to be extremely abstract about it research is just a system for participants in the game to change old things or get new things. I know the word "thing" is ill defined, but it will do for now. Research is results and the conditions necessary to obtain them. To that end, I feel like I should keep the logic separate from the data. I'm thinking about having editable data files that define research results and the necessary conditions to obtain them. I have actually seen another game do this. This makes it easy to modify research without mucking with the source. There are problems I need to work out though. The data in the file needs to be interpreted somehow. I'll either have to learn how to parse input and interpret it, or evaluate the data in the file directly. Neither option is palatable for different reasons. Another thing I need to give consideration is when these conditions should be evaluated.

On the plus side, this abstraction should help my other issues. For instance, it wouldn't matter how research points are applied (each turn or continuously) just as long as they meet or exceed the project's point threshold.

sirpengi at irc.freenode.net ##python-friendly has suggested that a markup language may fulfill my requirements. Fortunately, there are a variety of markup language parsers available for my language of choice. I may be implementing this sooner than I thought I would.