What is 'reasonable' return on research?

[Baron Munchausen]

I've been modding weapons and thinking about what the structure of a weapon tech tree 'ought' to look like. Which makes more sense: A fixed increment of improvement for each level, the same amount of improvement at each tech level even thought the cost is higher (diminishing returns)... or gradually increasing improvements as the tech cost increases ('proportional' returns)? I can think of a lot of arguments for each side. Sure, it's 'realistic' to have diminishing returns for effort. But it's also realistic to make sudden breakthroughs giving huge unexpected benefits. Too bad we can't have those in SE, but everyone has the tech tree memorized after a few games.

[ July 12, 2002, 17:31: Message edited by: Baron Munchausen ]

[Trajan]

Baron,

Using diminishing returns seems the most appropriate form. Decreasing marginal gains accompanied with increasing costs as you explore closer to your research possibilities frontier would be a formidable research design.

Keep in mind though, that as new incremental improvements in applied technology are discovered, I think that there have to some kind of tangential creative bursts that lead to new areas of research -- whether applied or theoretical.

Theoretical research paths should likely be very expensive to begin, and have a steep curve for continued research in the particular area.

Finally, I like the idea of a race surpassing their research possibility frontier through long-term research. Their passing of this mark (the "theoretical" limit to their research capabilities) might let them explore new and Extremely radical theories that are outside the boundries for their race. However, these research paths have to be considered within the argument of balance between the designed races.

Cheers!
Trajan

[ July 12, 2002, 17:48: Message edited by: Trajan ]

[Gryphin]

Is it possible to have both?
If so, is there any way to have a "random" event?
Probalby "No" to both questions. In which case, I guess we should put it in the SEV list of "I would really like to have"

[Suicide Junkie]

You can do both if you make multidimensional tech areas.

Have the normal "uninspired" tech area produce a reasonable start, but have diminishing returns.

Add a new tech, the "breakthroughs" area.
At each higher level of the breakthroughs area,
you will open another dimension of research into the original tech.

Your multidimensional tech could be something like the following:

code:

---

   |  uninspired

   |  1   2   3   4   5   6   7   8

---+-------------------------------

B 0|  5  10  15  20  25  30  35  40

T 1| 10  15  20  25  30  35  40  45

# 2| 15  20  25  30  35  40  45  50

1 3| 20  25  30  35  40  45  50  55

---

The uninspired research will provide hefty practical gains, but soon becomes expensive to add levels too.
A breakthrough will allow you to start working on a new dimension to the problem, which happens to be a tech level that is "fresh", and not yet suffering from the diminishing returns.

[Pax]

Originally posted by Suicide Junkie:
You can do both if you make multidimensional tech areas.

Have the normal "uninspired" tech area produce a reasonable start, but have diminishing returns.

Add a new tech, the "breakthroughs" area.
At each higher level of the breakthroughs area,
you will open another dimension of research into the original tech.

Your multidimensional tech could be something like the following:

code:

---

   |  uninspired

   |  1   2   3   4   5   6   7   8

---+-------------------------------

B 0|  5  10  15  20  25  30  35  40

T 1| 10  15  20  25  30  35  40  45

# 2| 15  20  25  30  35  40  45  50

1 3| 20  25  30  35  40  45  50  55

---

The uninspired research will provide hefty practical gains, but soon becomes expensive to add levels too.
A breakthrough will allow you to start working on a new dimension to the problem, which happens to be a tech level that is "fresh", and not yet suffering from the diminishing returns.

The MOD I'm working on can make good use of this concept; thanks for the idea!

As-is, I'll hve a two-dimensional array for technology based on simple-tech-field, and which of several Racial Technology types you have (including a 0-cost-advantage "Standard" type); for example (and only off the top of my head for now):

code:

---

----------------------------------------------------

Simple      |            |            |            |

Propulsion  |     1      |     2      |     3      |

Level --->  |            |            |            |

----------------------------------------------------

w/ Standard | Engine Is  | Engine IIs | Engine IIIs|

            |  speed +0  |  speed +0  |  speed +0  |

            |  cost +0   |  cost -1   |  cost -2   |

----------------------------------------------------

w/ Organic  | Engine Io  | Engine IIo | Engine IIIo|

            |  regen +3  |  regen +3  |  regen +3  |

            |  cost +2   |  cost +1   |  cost +1   |

----------------------------------------------------

W/ Temporal | Engine It  | Engine IIt | Engine IIIt| 

            |  speed +1  |  speed +1  |  speed +1  |

            |  cost +2   |  cost +1   |  cost +0   |

----------------------------------------------------

----------------------------------------------------

Unique      | *at all levels, new engine type w/   |

"Improved   | increased supply storage and damage  |

 Engines"   | resistance becomes available, similar|  

            | to each type above.                  |

----------------------------------------------------

Unique      | *at all levels, gain a +1 speed bonus|

"Inertial   | for Propulstion I-III, +2 for IV-VI, |

 Dampener"  | +3 for VII-IX, and soon, compared to |

            | base engine types above              |

----------------------------------------------------

Unique      | *at all levels, gain a new engine    |

"Compact    | which requires half the space of the |

 Engines"   | normal, base types above             |

----------------------------------------------------

---

... and so on. So "Propulsion 5" will be about as useful for an Organic-tech empire as for a Temporal-tech empire, but what each gets at that level will be slightly different.

Then, perhaps I can add in Unique / Ancient class technologies (as above) which further expand the available tech ... based on what you already have! So instead of getting X component or facility regardless of your own current technology, the Ruins will instead improve your possible research gains, past present and future.

A great inspiration, it'll improve things immensely! Thanks, guys! Of course, this also means scads more work creating components, but hey, what else is MODding for, eh?

Oh, and as you can see, I'm currently figuring on Temporal being slightly -faster- (+1, +2, +4, +6 speed-bonus engines, instead of the current +0, +1, +2, and +3 engines, for example), but I will of course compensate in other ways -- cost is one way, at least.

The neat thing is, with the right work done on the tehcnology setup, I (or others) will be able to add entire new technology-types just by cut-and-psting into the proper data files to set up the racial technology field, and it's components. Even more Unique fields for enhancing what's already there, with similar work.

(edit: UBBcode error, argh)

[ July 12, 2002, 19:55: Message edited by: Pax ]

[Baron Munchausen]

My god... I've started a graduate seminar on 4X game design...

BTW, I agree that Temporal races ought to have a propulsion advantage. I used that nifty 'warp core' component that someone posted ages ago and gave them their own propulsion technology. It's got the 'bonus movement' ability like the solar sail, and it's also a Quantum Reactor so they don't have to research that independently.

As for all these research graphs and trees, this is awfully elaborate for a simple system like SE IV is right now. I was just referring to linear benefits. Does it makes more sense to get a larger and larger degree of benefits for increasing research costs or should benefits be linear, which results in 'diminishing returns' by default.

Yeah, for SE V I'd like to see variable tech research costs, and variable 'connections' across the tree to get to certain techs. This would play havoc with the AI but it would be so much more interesting for human players. But this will require major hardcode revisions. I suspect that the elaborate systems being discussed here will be a real pain to implement and especially to maintain with the current design of SE IV. We need some way of implementing entire series of components with one definition.

[ July 12, 2002, 20:29: Message edited by: Baron Munchausen ]

[Trajan]

I don't think that research can really be thought of in a linear sense.

The two dimensional array's work nicely, though a three dimensional one would be better.

I think I am visualizing it as there being some way to advance through one research matix then make an advance in a particular area that suddenly jumped you to a different level of the matrix where the recent dramatic discovery would have an effect on all further research no matter the direction you take. Could you revert back to the original research matrix? I am not sure. I think that if you found a cheaper/better/faster way to propel your ships, not through better implementation of a know field of advancment, but rather through a never-before imagined breakthrough technology you would leave the old way behind and jump on the new technology.

for ex. If you spent 100 years developing transportation based on steam engine technology, then suddenly split the atom and found that you could make anti-matter propulsion a reality, why would you go back to steam engines in your ships.

---I am just brainstorming here, but hey...it is Firday afternoon in the States, and I am stuck at my desk with little motivation to work.

Cheers!
Trajan

[Gryphin]

Um, cool
If I understand,
You are trying to allow for:
Eureka! = as in the sudden realizaton of how to determine if the crown was pure gold. Making the connection between two seemingly unrelated concepts.
Serindipty = as in the realization that microwaves melted the chocolate in the engineres pocket. Making the connection between cause and effect
Good hard plodding work = As in the cure for cancer. Trial and error, Observing Cause and Effect, and Eurika!
That sounds like one sweet challenge. Hmm, I sound like the mouse who said soneone should put a bell arround the neck of the cat.

[Trajan]

If I understand,
You are trying to allow for:
Eureka!
Serindipty
Good hard plodding work

BINGO!

Basically, you would have a single non-complex matrix for the inital (Good hard plodding work) technological systems research set. If you encounter a "Eureka!" event (I dont know how the cause of that would work) you could jump to a new level of research in your current universe that has some significant multipliers applied to costs, but the payoffs would be equally significant.

Finally, there is the Serendipity event that would suddenly teach you about a whole new avenue of thought and could take your technological research to new levels, with new functions of cost applying to the technological advancements.

I am looking for a picture that would tell the story a little better. Basically I am talking about a three dimensional matrix of technological advancement.

Cheers!

[Gryphin]

oh, I forgot one:
Eureka!
Serindipty
Good hard plodding work
and
BINGO as in:
1) A farm Dog
2) Interpreting peoples thoughts