#24271 - sgeos - Sat Jul 31, 2004 7:04 am
I think it would be really neat to generate RPG/strategy game style characters using parents and genes. The implementation is simple- full hue spectrum hair and eye color can even be done. I'm looking for an application for these characters.
Here are a few variations on generating new characters:
A) Each turn is a year. Every character is born, grows old, and dies. (Hopefully better units have a bunch of kids before they die.) How big a population would we need for this to work without pulling units from other "tribes"?
B) Generation based. All units are basically the same age. At the end of a generation the new set of characters is generated.
C) Something in the middle. Have multiple generations present at a time- children. youth, adults, and the elderly. Only youth and adults are useful.
I'm not satisfied with my scenarios to date, but here they are:
Play as a diety and lead a group of people to victory. (Seal off demon gates or some such.) The diety can manufacture avatars to lead the populace, introduce new skills and stimulate the gene pool. Only one active avatar can be present at a time- the next avatar is designated as such as a child and takes over when the current one either retires or is killed.
Control a population in a flying fortress. The fortress fights against other flying fortresses. (The survivors of the great accident live on the moon, that is where the fortresses come from.) Your fortress is protected by flying robot things that your population pilots. You also need bridge crew for the fortress itself. If the fortress is heavily damaged it takes years to repair.
I think the main problem I have with the above concepts are:
A) Finding a way to mix with other populations.
B) The scenarios are fairly complex and would be hard to implement.
Better applications anyone?
-Brendan
Last edited by sgeos on Mon Aug 02, 2004 6:54 am; edited 1 time in total
#24286 - sajiimori - Sat Jul 31, 2004 6:07 pm
Besides the mythical elements, there's nothing that keeps your first scenario from otherwise mirroring real life, where populations intermingle all the time. The second scenario seems to go out of its way to seperate populations.
Use a more powerful language to ease implementation. Write only your core engine in C, then use something higher level for the rest. Lua is good; Scheme is nicer but it's hard to find compact, fast implementations.
When using higher-order languages, it's important to start thinking in that language as soon as possible. If you write C code in Lua, all you get is syntactic sugar. If you write C code in Scheme, it's ugly as hell -- the evolutionary paths for their ancestor languages diverged in the 50s.
#24305 - sgeos - Sun Aug 01, 2004 4:52 am
| sajiimori wrote: |
| Besides the mythical elements, there's nothing that keeps your first scenario from otherwise mirroring real life, |
Of course. I'm trying to avoid creating family, or family lineage. No point recreating the sims.
| sajiimori wrote: |
| where populations intermingle all the time. The second scenario seems to go out of its way to seperate populations. |
I think my problem is first scenarios that strike me as interesting are poor applications for the mechanic. The first good application for the mechanic is not something I'd be interested in playing (family).
After I posted (because that is how the universe works) I thought of an application that strikes me as good and interesting- play as an indebted family in a place where debts are inherited. The goal is to get your family out debt over generations. I like mystical stuff, so I'd use a mysttical setting and say that adventuring is a fine way to earn money.
| sajiimori wrote: |
| Use a more powerful language to ease implementation. Write only your core engine in C, then use something higher level for the rest. Lua is good; Scheme is nicer but it's hard to find compact, fast implementations. |
Do these languages have easy ways of putting pretty pictures on a screen, or am I looking at C for that?
| sajiimori wrote: |
| When using higher-order languages, it's important to start thinking in that language as soon as possible. If you write C code in Lua, all you get is syntactic sugar. If you write C code in Scheme, it's ugly as hell -- the evolutionary paths for their ancestor languages diverged in the 50s. |
Thanks for the feedback.
-Brendan
#24324 - tepples - Sun Aug 01, 2004 4:34 pm
You'd probably put pretty pictures on the screen in C and expose callbacks for whatever scripting language you choose.
_________________
-- Where is he?
-- Who?
-- You know, the human.
-- I think he moved to Tilwick.
#24331 - sajiimori - Sun Aug 01, 2004 5:40 pm
Yeah, Lua and Scheme aren't GBA languages, so they don't know anything about how to display pictures on GBA. If you're implementing the language yourself, you can give it support for direct hardware access, but I'd recommend leaving all that to the engine which is presumably written in C.
As an added bonus, you could then rewrite the engine for other platforms and have a portable game. ;)
| Quote: |
Of course. I'm trying to avoid creating family, or family lineage. No point recreating the sims.
|
Reading between the lines, that sounds like "too hard having sims-like complexity," but then you're talking about lineages when you mention debts.
Where does the evolutionary part come in? What's the fitness function? Does the player care about the results?
Getting a population to evolve in an interesting and interactive way is a greater challenge than simulating families and lineages, IMO.
| Quote: |
I think my problem is first scenarios that strike me as interesting are poor applications for the mechanic. The first good application for the mechanic is not something I'd be interested in playing (family).
|
I'd suggest stepping back for a moment and thinking about what your actual core idea is, and what's appealing about it. Chances are, the idea is appealing to a lot of other people too -- you just have to identify it and keep it in focus.
It's really backwards to take something that isn't particularly fun and then say, "Let's make this into a game. What would make it fun?" Find the core of the game first -- the thing that made it interesting to you in the first place.
If you find that your original idea (i.e. the thing that interests you) isn't really a game, I'd say don't try to force it into a game context. If you enjoy programming games, you'll find that writing simulations is not so different.
#24343 - sgeos - Sun Aug 01, 2004 9:32 pm
| sajiimori wrote: |
| Reading between the lines, that sounds like "too hard having sims-like complexity," |
Not at all. My sister is a huge sims fan. The game was originally going to be called doll house. The sims is an electronic doll house. I'm not overly interested in playing doll house. If I am, I'll borrow the sims from my sister.
| Quote: |
| but then you're talking about lineages when you mention debts. |
A game/simulation with genetic unit crossing could be done micro scale, at the individual level- play as a critter, play as one of the critter's offspring, repeat. It could also be done at the macro scale- you have 2000 critters and they move as a group and interact with other groups. I'm more interested in individual critters, so I'm personally looking for something closer to the micro end of the spectrum.
| Quote: |
| Where does the evolutionary part come in? |
If you start with a random gene pool and enough units, eventually poor genes will become rare.
| Quote: |
| What's the fitness function? |
Death. If a unit gets killed it will not be reproducing. The player still has other units to muck around with.
| Quote: |
| Does the player care about the results? |
That depends on what the player can do with the critters. If all we can do with them is cross them and get new critters, nobody will care about anything other than the aesthetics of a given unit.
| Quote: |
| Getting a population to evolve in an interesting and interactive way is a greater challenge than simulating families and lineages, IMO. |
In interesting and interactive ways, yes. I think that is due in part to removing the focus from individual critters- it is harder to relate to a group of 500 island people than it is to Albert, the ferry operator.
| Quote: |
| I'd suggest stepping back for a moment and thinking about what your actual core idea is, and what's appealing about it. Chances are, the idea is appealing to a lot of other people too -- you just have to identify it and keep it in focus. |
I enjoy generating units- be they rolled up or a cross of other units. The problem is that this soon becomes dull if there is nothing to do with them. I could make Pedigree Advance and have dog contests- that keeps the focus on breeding, but again, I'm disinterested in dogs. =P
| Quote: |
| It's really backwards to take something that isn't particularly fun and then say, "Let's make this into a game. What would make it fun?" |
I think it is a case of "this is fun and neat, but gets old fast."
| Quote: |
| If you find that your original idea (i.e. the thing that interests you) isn't really a game, I'd say don't try to force it into a game context. If you enjoy programming games, you'll find that writing simulations is not so different. |
The original idea is a character generation scheme. I'm looking for a game scenario that requires (and forces) generating characters at relatively fixed intervals. I also think it would be neat if the NPCs had lineages as well. The ferry operator's son, the defeated necromancer's great grand daughter, etc.
-Brendan
#24344 - keldon - Sun Aug 01, 2004 9:33 pm
Dragon Warrior Monstors featured inheritance of features, etc; so some of the concepts can make part of a larger game, as is often the case of RPG's. For example a survival war game spanning generations where you meet tribes along the way, the twist is that your women have to pick men otherwise they will be pregnant with a rubbish child; and what a man does makes no difference because you can't carry them with you. These people then take part in command and conquer style warfare, and the grooming part of hte game happens every now and again.
But yes, stick with what Sajimori said about whether it is game worthy or sim worthy.
#24353 - sajiimori - Mon Aug 02, 2004 1:37 am
I guess the question is: Is this an evolution game, or an RPG with evolutionary elements? If it's an evolution game and you think it will get boring quickly if it's not supported by other elements, then it probably won't work as a game at all. If it's an RPG, how much are you willing to invest in evolutionary features, and how much will those features actually contribute to the game? Will they enhance the game, or distract from its core idea?
Ever played Black & White? There's a good example of a game that didn't know what it was trying to be. The designers decided that the core of the game was the experience of being a god and interacting with your people. At some point (apparently years after they started), they realized that the novelty of it gets old pretty fast and they'd have to make an actual game around it. Not suprisingly, the game wasn't very good.
#24363 - sgeos - Mon Aug 02, 2004 7:07 am
| keldon wrote: |
| Dragon Warrior Monstors featured inheritance of features, etc; so some of the concepts can make part of a larger game, as is often the case of RPG's. |
Of course. I started seriously considering this mechanic after messing around with breeding pokemon. The game mechanics in pokemon are great, but I can only take so much of parody kiddie monsters.
| keldon wrote: |
| For example a survival war game spanning generations where you meet tribes along the way, ... These people then take part in command and conquer style warfare, and the grooming part of hte game happens every now and again. |
What is your tribe's goal? If it is survival and the game does not get harder over time, it sounds like an open ended war sim. If it does get harder over time it sounds like it wants a scoring system.
| keldon wrote: |
| the twist is that your women have to pick men otherwise they will be pregnant with a rubbish child; |
An interesting mechanic. Still, I think that requiring a set of units with different genders is a better way of doing things for both a simulation and a game.
| keldon wrote: |
| and what a man does makes no difference because you can't carry them with you. |
Can you carry women to the next generation? It doesn't sound like they can be carried either.
ramble/
I think that the system we use where families taking the male's family name is broken. The mother ought to and probably does know who the father is, but everyone else has to trust her on that one. One always knows who the mother is at birth (and basically forever after if the child is not taken from her). Therefore, it makes more sense to me if children take the mother's name instead of the father's.
Ideally a game would contain the option to operate (in) a matriarchal society, a patriarchal society or a society where a new family unit assumes the family name of the most prominent family, regardless of gender (if you are rich/powerful, your name gets used).
/ramble
| sajiimori wrote: |
| If it's an evolution game and you think it will get boring quickly if it's not supported by other elements, then it probably won't work as a game at all. |
I may agree with you. An evolutionary game wants critters and generations. A fitness function bridges this generation with the next generation. I think that all of the game elements should be part of the fitness function. Anything else does not belong in the game. If the (more than likely complex) fitness function is fun, the game will probably be fun.
| sajiimori wrote: |
| If it's an RPG, how much are you willing to invest in evolutionary features, |
A basic gene based attribute system is actually very cheap to implement in programmer time. It will cost at least double the RAM needed to store gene based attributes.
| sajiimori wrote: |
| and how much will those features actually contribute to the game? Will they enhance the game, or distract from its core idea? |
This depends on how you define enhance, detract and distract. Keep in mind that an enhancement to one person can be a detraction for another. If nothing is done with the gene based system, it will not change the core game is imposed on (wasted programmer time). The more the core game is changed to suit the mechanic, the further the will get from what may be expected. This alone could be considered a distraction.
If we take an RPG and slap a generation system on it, players can not create a final uber party. This will frustrate and turn off players that enjoy doing that and are used to doing it. Romancing Saga 2 was not gene based, but it was generation based. They circumvented this by limiting the number of generations- there was a final generation. The game was not level based, individual stats grew at the end of battles. A new generation started with units somewhat more powerful than those that ended the previous generation.
Allowing the PC party to run into NPC decendants could be costly in programmer time. At any rate, the game could not be based on characterization. Any story would have to be about the world. (RS2 has a non character based story. My friend argues that is doesn't really have a story. The gameplay is good, although atypical.)
I suspect that you are more or less saying 'proceed with caution- maintain focus'.
| Quote: |
| Ever played Black & White? There's a good example of a game that didn't know what it was trying to be. The designers decided that the core of the game was the experience of being a god and interacting with your people. At some point (apparently years after they started), they realized that the novelty of it gets old pretty fast and they'd have to make an actual game around it. Not suprisingly, the game wasn't very good. |
Never played it. After reading a couple reviews on gamefaqs, it looks to me like it was an electronic toy/sim with too few features that they tried to make a game out of. This is actually the reason I'm trying to steer clear of a sim. It is easy to say "it would be neat if the shrubs were genetically generated and created different flowers!" Sure, but that is interesting for all of about five minutes (if noticed) and adds basically nothing to the game. time_wasted++;
Would it be interesting and challenging to add affairs to a game where be breed tokens representing humans. Sure. Would it add anything? Not that I can think of, so it strikes me as a waste of time.
Games need goals. Passing the fitness function strikes me as a fine goal. If there is not a limited number of generations, the game will be open ended. I reasonably complex generation based game probably has multiple fitness functions.
Thank you for the feedback so far. Reading and responding to it has forced me clarify my thoughts. I have also added a poll regarding choice of scale. I voted lineage.
-Brendan