Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Reelya]

By the way, do you think this would fit better in Creative Projects?

Uh yeah, Gornova, i'd heavily suggest shifting the thread there. Click the bottom left of the page "move thread" thingy, and turn off the option to leave a redirection post, for clutter reasons.

[Gornova]

By the way, do you think this would fit better in Creative Projects?

As I've said in the first post, if the moderators think is better, for me is okay

[Gornova]

thanks @Toady One!

[Reelya]

There are a couple of advantages. Creative Projects is a slower board, so your game will stay on the front page for up to a month instead of a day.

Also, you can go back and edit the thread title on the original post, to give the thread as a whole a more informative title. Something like "Infiltration - Game v1.0"

That way makes it clear what type of thing it is, you can put the version number in the title, and update links in the first post, and people can see that a new version has come out when the thread title also changes.

[Gornova]

There are a couple of advantages. Creative Projects is a slower board, so your game will stay on the front page for up to a month instead of a day.

Also, you can go back and edit the thread title on the original post, to give the thread as a whole a more informative title. Something like "Infiltration - Game v1.0"

That way makes it clear what type of thing it is, you can put the version number in the title, and update links in the first post, and people can see that a new version has come out when the thread title also changes.

nice suggestion Reelya, done!

[Reelya]

by the way, have you considered implementing a procedural map generator for this?

"Seed" values can let people create any world they like, for greater replayability.

However, there's one designer who has an interesting procedural game I heard about. He created a ton of levels via seeds, played them and found a set of interesting ones, then graded them from easiest to hardest, and those become the "official" levels of the game. That way, you get the replayability of procedurally generated levels, but you also have a clear number of "official" levels that people can talk about and use to compare their strategies and results.

[Gornova]

by the way, have you considered implementing a procedural map generator for this?

"Seed" values can let people create any world they like, for greater replayability.

However, there's one designer who has an interesting procedural game I heard about. He created a ton of levels via seeds, played them and found a set of interesting ones, then graded them from easiest to hardest, and those become the "official" levels of the game. That way, you get the replayability of procedurally generated levels, but you also have a clear number of "official" levels that people can talk about and use to compare their strategies and results.

Yes I have a lot of ideas for "procedurally generate" content in the game and I will for sure consider seed generation, something I wanted to explore first by myself, but first I need to assest base mechanics and focus on gameplay and then move on content, generated or not.

[Rowanas]

Hah, it's no good having random seed generation until you know all the variables that your seed has to alter.  Oh, and Gornova, if you get round to exploring seed generation I'd like to know how that works, since my flimsy knowledge of coding would only allow me to brute force such a thing, which seems like definitely not the right solution.

As a more general note, you might generate some interest if you take some time to go through your process, turning this into a two-in-one Infiltration and design/mechanics thread.

[Reelya]

Hah, it's no good having random seed generation until you know all the variables that your seed has to alter. 

That's not really how it works. It's an iterative process. A seed is just the initial value for a random number generator, and you grab a stream of random numbers for whatever you need. You start with a seed system, then you find uses for the numbers it generates.

It's impossible (or just really really dumb) to avoid having the seed system until you "know all the variables" you have to alter in your generator system. That's just not a practically workable way to go about it. You start with the seed system then see what things need random rolls and pull numbers from the stream for that. It's the only physical way to build the thing, you need to see what comes out of different seeds during the design process.

Sure, guy could build the final dynamic game, then try and retrofit a world-generator to it, to get initial worlds, but that would basically stink as a design process. What makes more sense is to break down the world into rules, generate worlds from the rules and have those be the same rules that you manipulate during play. So "history" made the world from a random seed, and those same rules are what you manipulate to change the future of the world as the gameplay. Thus, if the world is going to have procedural generation at some point, that's going to profoundly change how the gameplay works, so it would be far more efficient in terms of developer effort to make the seed system first, have it generate worlds according to the normal rules of play, and adapt those rules of play iteratively in ways that make fun worlds that are fun to interact with.

[Rowanas]

Hah, it's no good having random seed generation until you know all the variables that your seed has to alter. 

That's not really how it works. It's an iterative process. A seed is just the initial value for a random number generator, and you grab a stream of random numbers for whatever you need. You start with a seed system, then you find uses for the numbers it generates.

It's impossible (or just really really dumb) to avoid having the seed system until you "know all the variables" you have to alter in your generator system. That's just not a practically workable way to go about it. You start with the seed system then see what things need random rolls and pull numbers from the stream for that. It's the only physical way to build the thing, you need to see what comes out of different seeds during the design process.

Sure, guy could build the final dynamic game, then try and retrofit a world-generator to it, to get initial worlds, but that would basically stink as a design process. What makes more sense is to break down the world into rules, generate worlds from the rules and have those be the same rules that you manipulate during play. So "history" made the world from a random seed, and those same rules are what you manipulate to change the future of the world as the gameplay. Thus, if the world is going to have procedural generation at some point, that's going to profoundly change how the gameplay works, so it would be far more efficient in terms of developer effort to make the seed system first, have it generate worlds according to the normal rules of play, and adapt those rules of play iteratively in ways that make fun worlds that are fun to interact with.

Oh, but doesn't the length of the string need to account for the number of variables you'll be pulling from, or do you expand the possible results by pulling noncontiguous numbers (or letters or whatever's standing in for a value here) from a relatively limited string length?  I also know that there are games that will generate specific universes for certain seeds as Easter Eggs, but are those by-and-large "false" seeds that bypass the generation system to give a pre-designed world? Seed generation and utilisation is one of the bits of game design coding that I know the least about while also being tremendously interested in it.

[Reelya]

The seed is a set length, usually 32-bit but can be 64-bit.

For example No Man's Sky claims to have 2^64 possible planets, that's because their seed is 64 bit. Everything about the planet is generated from that one value, and you can add new randomized features without needing to change the seed value.

The term "seed" means that you use that value to seed a pseudo-random number generator. The RNG then generates an infinite sequence of other numbers for you, and you pull from that sequence to get as many or as few variables as you need. Note how Dwarf Fortress world seeds work. There is a set length seed number, yet Toady is able to add entire new features into worldgen without adding a bigger seed.

However, one big caveat is that if you modify the algorithm for generating worlds, old seeds won't generate the same world as before. This happens whenever Dwarf Fortress is updated for example. To illustrate this, imagine a one-digit seed, say 7, that generates numbers 6,2,6,9,2,1,0,5... Now, you've assigned each value in the stream to be used for some different random roll, but then you add another roll in there somewhere. Say the nations had hard-coded colours, but now you decide to make each nation roll their own colour randomly. Sure, now they have different colours, but the entire history of the world will probably be different too.

However, changes in how you use the RNG numbers only affect subsequent calls. For example, if Toady generates terrain first, then civilizations, a change in how the RNG seed is used for civilizations won't change what terrain is generated, the seed will still be the same "world" but with a new history.

[Rowanas]

Ahhh, so the seed is only the first step in generating further random numbers, rather than the whole random number generation. In the first example you gave, if each use of the random number string is a separate call ,then you could add the flag colours without ruining the world generation, right?  And wouldn't you make each individual element that you want a separate call anyway? Or is that then a waste of time?

[Reelya]

Hmm, you're not really getting this.

To picture it, an rng is a function that you feed a seed value in to prime, and after that it outputs a single stream of random numbers. You then call the function when you want a dice roll. The rng guarantees that you'll get the same sequence of dice rolls across a run of the program, however it does not guarantee what those dice are used for, that's the programmer's job.

So, if you make 10 rolls you'll get predictable results for each roll. However, if you roll an extra dice first, then the next 10 rolls will be shifted along by one roll, since you already used up the first roll for whatever the new thing was. So, the old first roll of the new generator now gets the second roll of the rng, the old second roll gets the third roll and so on.

So, that makes it tricky for a systems designer, since if you share interesting seeds with beta-testers then make even small changes to worldgen, then those old seeds no longer even remotely lead to the same worlds. The choices are to completely ignore this or to find ways to mitigate it. Ignoring it is less work for the developer and less resources needed for the computer so almost all people just ignore this problem.

[Gornova]

As a more general note, you might generate some interest if you take some time to go through your process, turning this into a two-in-one Infiltration and design/mechanics thread.

Sure! I'm trying to write down ideas here when I have time (bonus today, tag system design )

I'm in the middle of the "tag system" work, something I believe will add more content and replayability on the game (and more options to play with it for me too!).
With tag I refer to a simple property that an npc can have, for example "greedy" and a bonus associated to this property, for example -10% on social challenges. My plan is to use these tags when npc have to resist a challenge (a player action, like focus/calling/manipulation or quests), now splitted in three categories (like Vampire the masquerade and old Wod games ):

- physical, related to combat, fighting with swords, resist to poisons, etc..
- mental, related to willpower, mental ability to understand something and knowledge in general
- social, for ability to understand, manipulate and advance on social ranks

So for example good old Barbarian Uzbar could have a tag like "robustness" (+10% physical challenges) and "scar" (-20% on social challenges) and so will be a good fighter, but a not so good social manipulator, so quests that require social ability will be harder for old Uzbar, but fight goblins with an axe more easy. Keep in mind that a quest can have only one type, so it's a matter to find a balance between quests right now, design them so in every city for every character there is the possibility to conquer a district and to be a valid target for player.

Thinking about this, something I want to consider right now is to add a low probability for unique tags, like "chosen by gods" with good bonuses. This will provide some interesting choices for players: a focus challenge can have low probability to corrupt a powerful npc, but on the other side, could provide a powerful ally, right ?

There is only an option for a npc to acquire a tag (remember, a tag could be a bonus, but also a malus!) when a quest is solved or failed: so now for example our good Barbarian Uzbar fail to stop goblins and acquire tag "epic failure" or in case of a good outcome, gets "hero" tag.

What do you think about tags? In my mind could provide more personality and variability for npcs and on the other side for player too !


About seed generation, like I've said, is something I want to explore and think after go throught this prototype. For now random generation is all about the following elements:

- npc stats (followers, money, etc..), name (from an internal list)
- number of npc in a city (at least 3, max 10 more or less)
- combat between npcs (roll a dice in this case)
- challenges (roll a dice with tags bonuses)

So not so much to explore for now on random generation side!

[Rowanas]

Well, my opinion on tags is a double thumbs-up.  Having resolutions be influenced by randomised or earnt tags opens up a lot of opportunities for interesting emergent gameplay, and it has the bonus of being easy to change and add to.  The trick to making it interesting will be ensuring that the tags interact often enough and in a meaningful way to bring about situations that the player can work out the likely outcomes without being entirely predictable.