|
|
Post by echocdelta on Jan 4, 2016 11:56:51 GMT
Ok guys! Some interesting developments on this front. We've had our first round of meetings and things are going much better than expected. In order to curate the feedback properly, we're going to open up our Trello board for public viewing (or for people here) so that you guys can see/vote on features and we can discuss them here. I'm actually hoping to also (by the end of Friday) start opening up my design OneNote for public viewing but I need permission to do that and such.
Firstly, based on a lot of discussions on technical stuff from a fair amount of feedback from Lord Ba'al, morsealworth, Aynin, undeadoverpromise, Deth, Spiderweb, Qetesh and everyone else we've actually scoped out how the first round of development will go. Initially we'll be building systems that aim for availability, iteration and reaction to design/QA/testing.
As best able, I'll try to describe the first leg of the 'gameplay' design, implementation and development. The scope is to first create one gameplay 'loop' - game starts, player engages in tasks/objectives, receives feedback, tasks/objectives are completed and the game 'ends' or objectives are 'completed'. morsealworth, definitely a special mention, because a lot of the initial scope ideas came from your feedback. We hope to initially have the player's avatar and NPC settlement clusters as separate technical tracks that intersect at specific interaction points - our tasks will include to create NPC settlements that essentially work in their own interests, such as meeting needs, assigning pawns to do certain things (like move pawn to forest/area for resource) - essentially focusing on creating NPC ecology for a single village. The goal here is to essentially be able to prototype how AI manages NPC settlements in vacuum. Things like physical idols, NPC's doing their own thing, settlements having their own pagan gods/holy sites are all definite design goals. The next step will be to replicate settlement clusters across a specific landmass and have them all working in their own designated area, essentially prototyping the idea of 'they're going to do stuff without you'.
Why? Because a lot of the feedback here is exceedingly overlapping on removal of player interaction to 'order', or essentially substitute, the NPC AI or task allocation. Settlers, Tropico, Crusaders ie. are all design inspirations at this point. Our main aim for settlements is to get the groundwork to function as a platform and begin quickly setting up the ability to QA/Bug test/iterate on AI reactions to player interaction.
Our goals on interaction is scoped mainly, this is with consultation on tech/AI programmers, building systems to track how miracles are cast or how player actions are perceived and augmenting behavior sets accordingly - we feel pretty strongly that we want the NPC behaviors to change according to how you, as a God, act in their presence. It is likely that we're going to, on purpose, try to remove as much control over NPC pawns as possible and instead have them do things, like send out crusades/missionaries/armies/scholars based on your actions. Again, to stress, our main scope is first build the systems that track these things and make possible actions having AI behavior changes (and persistently tracking those changes).
Now the nitty gritty of it is that we have a lot of high level concepts from here that need to be crushed down into tech tasks, but more so, we're scoping for the following legs of development: prototyping single village NPC AI, multiple village NPC AI on same gameplay area, interaction between NPC settlements/behaviors on points of intersection, high-level AI cluster management (invade this, convert that etc.), pathfinding QA/bugs, prototyping feasibility of terrain manipulation (through miracles, to avoid 'flat land' repetitive stress), tracking of actions/miracles/interactions influences on AI and closed game loops - ie. rewards, disincentives, expectations from player.
The initial leg of the game's main goal is that you will spawn and be able to watch a village do it's thing without you needing to tell people to cut trees, click on cards or bubbles. There are challenges here on emergent AI systems, or 'how emergent', and simulation of AI etc. We've got an in-house AI specialist (who's doing a PhD on AI systems) and we have another AI specialist on the way, they'll hopefully be able to provide more feedback on this.
The downsides!
- At this point, I don't know how terrain manipulation is going to work. Changes to the landscape/mesh are things well within what we can do, we haven't done the prototype on it so I can't determine the outlook and the effects it'll have on pathfinding/AI.
- Idols are in, temples are in, pagan holy sites are in, factions are in, customization of God/Avatar is in (we want you to 'pick' your God hands at the start), but I'm making the call on limiting how low level the NPC pawn tracking will go, as in, I don't know that having individual level of detail on pawn descriptions/decision making trees is going to be the best option over having clusters of NPCs of a settlement react to you.
- Miracle casting will probably only be gesture based because we've decided hotkey-ing or menu selecting miracles is not an immersive experience. We know that we can tie the strength of miracles to how well gestures are 'drawn' by the player.
- I'm pushing for each hand being able to 'hold' a miracle and mixing miracles, but I've been told the main issue with this is design and particle system interaction. I really want a lightning-fire tornado to be a thing.
- Balancing is going to be a nightmare, as in, balancing of NPC AI so it doesn't go mental at conflicting behaviors etc.
- The fidelity of the game will remain low-fidelity/low-poly stylized art, because we can really do a fair bit on saving GPU processing and limiting drawcalls. The benefit of this is that we did performance testing on 400-500 pieces of physics debris flying around from smashed houses and the FPS (non-optimized mind you) only dropped a tiny bit.
- Decals are up in the air, like scorching or scarring the land. Requires feasibility and performance testing.
- Commandments might be relegated to a visual indicator per settlement, such as the town center having flags or banners that reflect NPC behaviors. Maybe a warlike (your fault) NPC settlement will have weapons and stuff etc. We're still looking at representation of behaviors visually and how they'll be tracked/communicated to the player.
- Complexity of NPC pawn assets are being debated, however swapping assets like hats, changing colors and stuff are all looking more than feasible.
- There will be no 'ages', instead we're going to incrementally build up the prototypes of NPC settlements to respond their visual assets to behavior sets and the 'size' of settlements, maybe swapping wooden/mud huts with stone buildings as settlements get bigger.
- War is going to be a method of NPC interaction based on behavior, we'll be making a call that combat is basically for RTS games and it should be a consequence as opposed to a primary method of interaction. God doesn't go to war, God inspires war.
- Miracle unlocking/abilities/how they're accessed or balanced will be further down the track.
- Lighting and particle systems will be prototyped, however we're currently running our prototypes with dynamic lighting.
- We're completely ruling out any low-spec machine, the minimum spec we're aiming for is higher-end PC machines. We've decided that if we cater for low-end machines or mobile, we have to cut a bunch of stuff. A lot of stuff.
- Currently scoping for sandbox singleplayer, but we will be essentially looking to scope out to 1 vs 1 multiplayer.
- Any persistent or account tracking is sidelined for now, mainly because I can't comment on server/client side hosting, how that will impact systems etc.
- We currently have no idea about timelines but plan on having transparent development, so hopefully with a public Trello board you can see what we're working on as we're working on it, what is next etc.
- Currently the project is yet to be formally adopted but end of next week we'll know 100%, as we'll know then too.
- Building destruction will remain but will not be 100 pieces of debris per building (maybe like 5-10), I'm an idiot for doing that and I was intentionally trying to break the engine (and it didn't break, but I'm still an idiot).
|
|
|
|
Post by Aynen on Jan 4, 2016 13:04:47 GMT
Thanks for the update! It's really nice to see how many of the ideas made here are finding their way into the game. For the nitty gritty of the design, will it be possible for us to play builds of the game, to get a feel for how a system plays out in the game? P.S., you're not an idiot for wanting to see awesome explosions...  |
|
|
|
Post by morsealworth on Jan 4, 2016 17:30:35 GMT
Why, thank you for your kind words, echocdelta . It even make me feel like a productive member of society for a moment. And that was a very good moment in my personal ranking. Also, I love the plans for public Trello boards. I'll never stop to use Subnautica as an example: It is well-organised, neat, hype-inducing (and, what's important, it gives adequate hype, the opposite of Molyneux hype) and is not PR showoff despite the previous part. And hey, if it lets me see where the heck the development is going, it sure as hell helps developers visualize that as well. Which, in turn, helps organise the project into something both possible and meaningful, as opposed to the mismanaged mess Universim is becoming (the guys are enthusiastic and do have potential, but their management side is severely lacking). And, like I said before, if you want anything to consult about, you can always come to me/us. You even already have my skype ID in case you ever need it (everyone here does, in fact). Not that you will ever need it, though. |
|
Lord Ba'al
Supreme Deity
Posts: 6,260
Pledge level: Half a Partner
I like: Cats; single malt Scotch; Stargate; Amiga; fried potatoes; retro gaming; cheese; snickers; sticky tape.
I don't like: Dimples in the bottom of scotch bottles; Facebook games masquerading as godgames.
Steam: stonelesscutter
GOG: stonelesscutter
|
Post by Lord Ba'al on Jan 4, 2016 17:52:05 GMT
I think it's a good call to aim for higher end pc's as the people who can afford to invest in a VR headset probably already have such a system or will at least have the common sense to realize that they would need one in order to enjoy VR.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jan 4, 2016 18:11:07 GMT
|
|
|
|
Post by echocdelta on Jan 5, 2016 12:32:01 GMT
Thanks for the update! It's really nice to see how many of the ideas made here are finding their way into the game. For the nitty gritty of the design, will it be possible for us to play builds of the game, to get a feel for how a system plays out in the game? P.S., you're not an idiot for wanting to see awesome explosions... It's my intention, and probably my call, that if the game is officially picked up next week that everyone on this thread gets a key and preview branch builds. My main apprehension on this is hardware availability, I'm going to have some discussions this week on how we can do non-VR builds for mouse-keyboard/free-flight WASD with mouse whilst we develop these prototypes for QA/play test feedback. I always liked the idea of preview builds on iterative development, and we might get to spot issues with compatibility really early on. I'll get back to you on this. Safe to say that once that kicks off, we'll probably just limit it to everyone here as opposed to having an open door on it. |
|
|
|
Post by echocdelta on Jan 5, 2016 12:38:36 GMT
Why, thank you for your kind words, echocdelta . It even make me feel like a productive member of society for a moment. And that was a very good moment in my personal ranking. Also, I love the plans for public Trello boards. I'll never stop to use Subnautica as an example: It is well-organised, neat, hype-inducing (and, what's important, it gives adequate hype, the opposite of Molyneux hype) and is not PR showoff despite the previous part. And hey, if it lets me see where the heck the development is going, it sure as hell helps developers visualize that as well. Which, in turn, helps organise the project into something both possible and meaningful, as opposed to the mismanaged mess Universim is becoming (the guys are enthusiastic and do have potential, but their management side is severely lacking). And, like I said before, if you want anything to consult about, you can always come to me/us. You even already have my skype ID in case you ever need it (everyone here does, in fact). Not that you will ever need it, though.Hey no stress, you helped out a ton with pumping some great ideas/concepts into this whole thing! Subnautica and UE4 development boards are my prime examples. We're really organized when it comes to Trello (I'm a bit obsessive about logging my cards) so it'll be a smooth transition I hope. Also you guys can see where things are hard, easy, developed in phases and for anyone doing game-dev it'll provide good insight into how systems are developed. I hope to open the board up by our Friday and share a link here. My development style is one of always under-scoping things and trying to build base platforms/mechanics. Everything we have right now is based on 6 full days of development time by two people, when the initial scope/budget was for 12 days with 3-4 people, or a 6 + 6 days of multiple phases of development. Instead, we have a number of base systems already working surprisingly (suspiciously) well. I have no idea about what Universim is doing at the moment, I was going to back it but I was fresh burned from Godus and thought I'd leave it and see how it went. I have friends that share your sentiment on their development but they're really taking on a massive scope project. It'll be cool to maybe have Skype development chats with everyone once a week, so I might hold you to that. So far, the crowd-sourced development thing is working really well, so I think it'll be a good thing to try. |
|
|
|
Post by echocdelta on Jan 5, 2016 12:41:50 GMT
I think it's a good call to aim for higher end pc's as the people who can afford to invest in a VR headset probably already have such a system or will at least have the common sense to realize that they would need one in order to enjoy VR. We do have other stuff in the bag in terms of hardware but I'm under NDA on that (because it's partner hardware). Low-end PC and mobile is totally out, but we will run performance testing on VR over multiple platforms, which might mean that everyone from PlaystationVR to HTC Vive will have access. Thankfully, we've got many Vives here already and a lot of other devkits for other platforms. I'll hit up our people on how much I can say about that stuff and elaborate as time goes on. |
|
|
|
Post by echocdelta on Jan 5, 2016 12:49:27 GMT
Yeah that was great news for us, it'll open the market up for consumer VR before touch controllers start coming through. Now, to highlight that stuff appropriately, I was a VR skeptic until I tried touch/motion controllers in VR. It's absolutely insane. The Vive DoTA and Bow demo was mind-melting. We're going to closely watch how the pre-orders and consumer reaction will go for those pre-orders, but we're anticipating a solid run on PlaystationVR consumer release and SteamVR is going to push pretty hard too. Hardware wars, I think, is an apt term. Really excited to see what GDC in March is going to unveil on the VR front as well! |
|
|
|
Post by Aynen on Jan 5, 2016 12:51:29 GMT
Thanks for the update! It's really nice to see how many of the ideas made here are finding their way into the game. For the nitty gritty of the design, will it be possible for us to play builds of the game, to get a feel for how a system plays out in the game? P.S., you're not an idiot for wanting to see awesome explosions... It's my intention, and probably my call, that if the game is officially picked up next week that everyone on this thread gets a key and preview branch builds. My main apprehension on this is hardware availability, I'm going to have some discussions this week on how we can do non-VR builds for mouse-keyboard/free-flight WASD with mouse whilst we develop these prototypes for QA/play test feedback. I always liked the idea of preview builds on iterative development, and we might get to spot issues with compatibility really early on. I'll get back to you on this. Safe to say that once that kicks off, we'll probably just limit it to everyone here as opposed to having an open door on it. I'm also thinking it would be very very handy to be able to share ideas visually, through drawings and the likes. I'm very much a visual thinker, so drawing up something in paint to explain something and tossing it over on skype could really help. |
|
|
|
Post by Gmr Leon on Jan 5, 2016 22:52:00 GMT
The scope is to first create one gameplay 'loop' - game starts, player engages in tasks/objectives, receives feedback, tasks/objectives are completed and the game 'ends' or objectives are 'completed'. morsealworth, definitely a special mention, because a lot of the initial scope ideas came from your feedback. We hope to initially have the player's avatar and NPC settlement clusters as separate technical tracks that intersect at specific interaction points: -Our tasks will include to create NPC settlements that essentially work in their own interests, such as meeting needs, assigning pawns to do certain things (like move pawn to forest/area for resource) -Essentially focusing on creating NPC ecology for a single village. -The goal here is to essentially be able to prototype how AI manages NPC settlements in vacuum. Things like physical idols, NPC's doing their own thing, settlements having their own pagan gods/holy sites are all definite design goals. The next step will be to replicate settlement clusters across a specific landmass and have them all working in their own designated area, essentially prototyping the idea of 'they're going to do stuff without you'. Why? Because a lot of the feedback here is exceedingly overlapping on removal of player interaction to 'order', or essentially substitute, the NPC AI or task allocation. Settlers, Tropico, Crusaders ie. are all design inspirations at this point. Our main aim for settlements is to get the groundwork to function as a platform and begin quickly setting up the ability to QA/Bug test/iterate on AI reactions to player interaction. Our goals on interaction is scoped mainly, this is with consultation on tech/AI programmers: -Building systems to track how miracles are cast or how player actions are perceived and augmenting behavior sets accordingly -We feel pretty strongly that we want the NPC behaviors to change according to how you, as a God, act in their presence. -It is likely that we're going to, on purpose, try to remove as much control over NPC pawns as possible and instead have them do things, like send out crusades/missionaries/armies/scholars based on your actions. Again, to stress, our main scope is first build the systems that track these things and make possible actions having AI behavior changes (and persistently tracking those changes). Now the nitty gritty of it is that we have a lot of high level concepts from here that need to be crushed down into tech tasks, but more so, we're scoping for the following legs of development: -Prototyping single village NPC AI, -Multiple village NPC AI on same gameplay area, -Interaction between NPC settlements/behaviors on points of intersection, -High-level AI cluster management (invade this, convert that etc.), -Pathfinding QA/bugs, -Prototyping feasibility of terrain manipulation (through miracles, to avoid 'flat land' repetitive stress), -Tracking of actions/miracles/interactions influences on AI and closed game loops - ie. rewards, disincentives, expectations from player. The initial leg of the game's main goal is that you will spawn and be able to watch a village do it's thing without you needing to tell people to cut trees, click on cards or bubbles. There are challenges here on emergent AI systems, or 'how emergent', and simulation of AI etc. We've got an in-house AI specialist (who's doing a PhD on AI systems) and we have another AI specialist on the way, they'll hopefully be able to provide more feedback on this. All this sounds pretty neat (minus terrain manipulation, anyway). Minor concern is: how do you guide the player behavior without seeming too prohibitive and retaining a sense of playfulness? I recall that being an awkward pitfall of Black & White 2. After assigning people to basic tasks, you could let them gather and do their own thing, but you were always directed towards tasks that felt somewhat out of place, and the map designs unfortunately felt all the more like levels with the zoomed out perspective. I feel like whatever tasks that are produced should, and I suspect this has been tossed back and forth in prior posts, only inadvertently involve the people of the world. At least, if sticking with the only faintly interventionist deity perspective. At the same time, it does seem like unless there's a world association with your deity, a tenuous connection with them should exist to justify directing any attention towards them. Otherwise you've made yourself something vaguely more like a mystical Cities/SimCity with even less control, which I doubt is the aim. Perhaps worse though, without sufficient relation to the world/people, you've also almost made something of an ant snowglobe, you watch and watch, occasionally shake or drip some food in some places to change how they make the ant colony, but once you've exhausted your options, it risks growing kind of dull. Think a curious idea might be this, to reduce that: -Create different AI sets once you figure out the base of what you want to do with it. -Try to adjust it so that recurring sets (or shared sets in the same match) are few, so each match will feel a little different. -Alternatively, create a few interesting, distinct sets that, when shared in the same match, have unique patterns from overlap (if possible). -Limit the reaction of AIs in response to your power use (e.g. through proximity to miracle use, "perfect" miracle casts*, etc.), basically a disbelief system. -This reaction instead of being solely a negative, serves to produce differing powers for yourself and, potentially, an opposing deity. -Disbelief powers would enable you to be more sly against opponents, yet at the same time reinforcing opposing deities' powers in other ways. *Perfect miracle casts would be something to be determined. Precise gestures could prove too frustrating, so perhaps instead an effective combo chain? Whatever it might be, were it to be implemented, it would have to be extremely clear, otherwise it would prove far too unenjoyable. Anyway, just some offhand thoughts. Sorry if any of this is tired repetition, haven't taken time to catch up with everything here. Very much imagine much of it may be. Nevertheless, really curious to see how this may pan out. |
|
|
|
Post by echocdelta on Jan 6, 2016 3:34:46 GMT
All this sounds pretty neat (minus terrain manipulation, anyway). Minor concern is: how do you guide the player behavior without seeming too prohibitive and retaining a sense of playfulness? I recall that being an awkward pitfall of Black & White 2. After assigning people to basic tasks, you could let them gather and do their own thing, but you were always directed towards tasks that felt somewhat out of place, and the map designs unfortunately felt all the more like levels with the zoomed out perspective. I feel like whatever tasks that are produced should, and I suspect this has been tossed back and forth in prior posts, only inadvertently involve the people of the world. At least, if sticking with the only faintly interventionist deity perspective. At the same time, it does seem like unless there's a world association with your deity, a tenuous connection with them should exist to justify directing any attention towards them. Otherwise you've made yourself something vaguely more like a mystical Cities/SimCity with even less control, which I doubt is the aim. Perhaps worse though, without sufficient relation to the world/people, you've also almost made something of an ant snowglobe, you watch and watch, occasionally shake or drip some food in some places to change how they make the ant colony, but once you've exhausted your options, it risks growing kind of dull. Think a curious idea might be this, to reduce that: -Create different AI sets once you figure out the base of what you want to do with it. -Try to adjust it so that recurring sets (or shared sets in the same match) are few, so each match will feel a little different. -Alternatively, create a few interesting, distinct sets that, when shared in the same match, have unique patterns from overlap (if possible). -Limit the reaction of AIs in response to your power use (e.g. through proximity to miracle use, "perfect" miracle casts*, etc.), basically a disbelief system. -This reaction instead of being solely a negative, serves to produce differing powers for yourself and, potentially, an opposing deity. -Disbelief powers would enable you to be more sly against opponents, yet at the same time reinforcing opposing deities' powers in other ways. *Perfect miracle casts would be something to be determined. Precise gestures could prove too frustrating, so perhaps instead an effective combo chain? Whatever it might be, were it to be implemented, it would have to be extremely clear, otherwise it would prove far too unenjoyable. Anyway, just some offhand thoughts. Sorry if any of this is tired repetition, haven't taken time to catch up with everything here. Very much imagine much of it may be. Nevertheless, really curious to see how this may pan out. I think this highlights some really important parts of engaging gameplay design - firstly the core elements need to be identified and then the secondary/tertiary loops need to be fleshed out. Then, we can distinctly branch tasks/objectives allocations to multiple tracks. For example, tasks are broken down to assignment and selection - player selects a task/objective to resolve or is given a task/objective to resolve. Currently I'm concerned about how these tasks are 'framed' instead of how they are generate or what they generate. Actually all of your concerns are major, not minor, gameplay concerns. But in terms of framing, we need to have stable core loops - if I shake the snowglobe, the snow will whirl, the snow acts like that everytime, what kind of tasks can I assign or select based on that, what kind of delivery of content can I design around those tasks etc. It's one of the things that Godus did, deviously, right (because it was using templates of mobile short gameplay loops) - click, feedback, reward, click etc. Sculpting became a process of those mobile design driven systems - sculpt, feedback, reward; see chests, sculpt, need faith etc. But, I think the 'base' point is to actually have the AI play like an antfarm and then begin expanding on those systems, so if we drop food, the ants flock to food, if we drop food in between two colonies, both colonies flock to food, if both colonies flock to the same area of food, what types of interactions will occur, if those interactions can occur, what kind of gameplay can we create for the player based on that etc. Expanding loops implies depth, the main concern (and one I share) is that we have core loops and we stay there, Godus being the single best example of this. Primary mobile game loops, no depth of gameplay past that. -Create different AI sets once you figure out the base of what you want to do with it.
This is chicken and egg in some respects, we'll probably need to work out the core mechanics and then look at how AI is, how it can be and how your actions can change that. -Try to adjust it so that recurring sets (or shared sets in the same match) are few, so each match will feel a little different.
Replay and avoiding repetition will be a challenge for any game designer, which is why linear narrative/progression is an attractive/cheap option. Too premature to speculate on this until we know what the core mechanics are/feasible. -Alternatively, create a few interesting, distinct sets that, when shared in the same match, have unique patterns from overlap (if possible). This is something I'm inclined to gravitate towards. -Limit the reaction of AIs in response to your power use (e.g. through proximity to miracle use, "perfect" miracle casts*, etc.), basically a disbelief system. - If AI can 'react' to miracles, having that change to a spectrum as opposed to a binary is a lot less problematic than one would assume. I imagine that we can do attrition of value on miracles from over-exposure (BW1 style), or even disbelief. -This reaction instead of being solely a negative, serves to produce differing powers for yourself and, potentially, an opposing deity.
- This idea is straight up cool, with the core mechanics in play I actually think there is a lot of merit in exploring this for design purpose. -Disbelief powers would enable you to be more sly against opponents, yet at the same time reinforcing opposing deities' powers in other ways.
- I actually can see a number of emergent gameplay opportunities coming out from this, which begins to feed into how we can tackle gameplay replay value, in all the good ways too. Note: I've attached the core loop I was talking about. Also because Aynin mentioned he is a visual person so this might help clear up the idea of game mechanic loops. So iteration on this will be to have a third core mechanic, 'cast', and a second loop for NPC core mechanics. An experimental plan I have is to develop two separate core loops for you and the NPC, bridging them over/intersecting them at interactivity mechanics. That might be a key way of tackling how NPC's will flesh themselves out and provide enough of a contrast to your agency. Attachments:
|
|
|
|
Post by Gmr Leon on Jan 6, 2016 5:16:45 GMT
Much of this has struck me as a little chicken and egg, for sure. Despite that, it's cool to see we're on the same page regarding the loops and the need to discern where it begins to descend into more depth, which is roughly what I've been prodding towards a little. I think once the AI performance is hammered out a little more, figuring out the influences of the player on them will become far clearer and easier to build around, as right now it does sound like a fair amount of it remains up in the air.
That said, I think striking at it along certain points (few distinct AI sets, spectrum of reactions, etc.) is perhaps the better approach until then. Especially as so much of it hinges on that as essentially the core user interface, albeit acting more of its own accord than preset repeating animations. With more complex AI behaviors, you potentially create more depth for the player to engage with as their interactions with each other or the environment become more complex.
Interestingly enough, this faintly reminds me of Reus, if you're familiar. That game is practically the epitome of hands off, indirect interaction with the people of the world, and fascinatingly enough handles the increased abundance of resources, and thus complex environments, pretty strikingly. It's not disbelief, but greed which begins to alter the behaviors of their AI, leading to infighting amongst them. However as it is so hands off, that's about all there is to it, there's no opposing deity so the greedy bastards don't have much of a counter to you, and those just getting by, despite your moderate aid, don't use that to develop any major defensive measures of their own, instead relying on you more or less.
It's not the best system by any means, but it's an absolutely interesting one to play about with, and one that's heavily influenced some of my own ideas towards god games (hence all my environmental shaping hubbub earlier).
Another one I only briefly tried recently also sort of dabbles with the differing ability set idea I mentioned, called Gods vs. Humans (free version is apparently available on Android, huh). You have to stop humans from building a tower of Babel, basically, while simultaneously preserving their respect for you as you crush their hopes and dreams. You have a completely different ability set from them, with divine powers and the like that are entirely dependent on their respect for you, which is earned based on how happy they are, something decreased every time you use an ability directly against them. Their abilities are all based around improving their work speed, power blocking you, reinforcing their constructions, and the like.
Not exactly what I had in mind, obviously, but a somewhat decent example of how if instead you duplicated the respect system for the enemy and instead of locking out powers with lost respect, unlocked different ones for each side, you could have a more interesting and curious system to experiment with.
|
|
|
|
Post by echocdelta on Jan 10, 2016 5:49:07 GMT
Much of this has struck me as a little chicken and egg, for sure. Despite that, it's cool to see we're on the same page regarding the loops and the need to discern where it begins to descend into more depth, which is roughly what I've been prodding towards a little. I think once the AI performance is hammered out a little more, figuring out the influences of the player on them will become far clearer and easier to build around, as right now it does sound like a fair amount of it remains up in the air. That said, I think striking at it along certain points (few distinct AI sets, spectrum of reactions, etc.) is perhaps the better approach until then. Especially as so much of it hinges on that as essentially the core user interface, albeit acting more of its own accord than preset repeating animations. With more complex AI behaviors, you potentially create more depth for the player to engage with as their interactions with each other or the environment become more complex. Interestingly enough, this faintly reminds me of Reus, if you're familiar. That game is practically the epitome of hands off, indirect interaction with the people of the world, and fascinatingly enough handles the increased abundance of resources, and thus complex environments, pretty strikingly. It's not disbelief, but greed which begins to alter the behaviors of their AI, leading to infighting amongst them. However as it is so hands off, that's about all there is to it, there's no opposing deity so the greedy bastards don't have much of a counter to you, and those just getting by, despite your moderate aid, don't use that to develop any major defensive measures of their own, instead relying on you more or less. It's not the best system by any means, but it's an absolutely interesting one to play about with, and one that's heavily influenced some of my own ideas towards god games (hence all my environmental shaping hubbub earlier). Another one I only briefly tried recently also sort of dabbles with the differing ability set idea I mentioned, called Gods vs. Humans (free version is apparently available on Android, huh). You have to stop humans from building a tower of Babel, basically, while simultaneously preserving their respect for you as you crush their hopes and dreams. You have a completely different ability set from them, with divine powers and the like that are entirely dependent on their respect for you, which is earned based on how happy they are, something decreased every time you use an ability directly against them. Their abilities are all based around improving their work speed, power blocking you, reinforcing their constructions, and the like. Not exactly what I had in mind, obviously, but a somewhat decent example of how if instead you duplicated the respect system for the enemy and instead of locking out powers with lost respect, unlocked different ones for each side, you could have a more interesting and curious system to experiment with. We spent a fair bit of time last week hammering out the loops now and starting to do paper prototyping, or white-board prototyping, of how different things will consistently aim to fit into the core gameplay loops. Currently we have a good idea of the starting point. I've been running the design loops in conjunction with our lead programmer/engineer and trying to see what types of stuff raise warning flags - currently, based on his comments, nothing so far is unrealistic or blown-scope. The main challenges are ensuring we don't have critical paths that bottleneck those iterations on development and we're heavily leaning on having a track for things that you do and things that NPC villages do. One of the best pieces of advice I've ever received on game design is that if you can't test it on paper or as a board game, your game is probably not going to work. The process was basically to first create a core NPC loop with secondary loops as deviations for unmet conditions to proceed to loop - village needs food, manager assigns pawn to get food, food is 'gathered', village grows, village needs food updated, manager assigns pawn etc. The idea is that we aim to build a system that can interchange food with resource A, B, C etc. The rationale for this, contextually speaking in terms of game concept, is that humans are particularly uninhibited at self-managing growth. We grow, breed and spread until we hit that 'loop' where food requirement is not met, which causes stagnation, which cycles that loop. If we remove existing food in the loop, on the next cycle, the system has to then say 'oh crap no food, start killing pawns' in order to correct itself. This then introduces an organic (as best as organic is, probably with a world of technical bugs) point to have a point of 'interactivity' for player; player causes famine, village begins to die out, or oppositely. On purpose I'm underscoping these systems as best able, so it might scarily sound like it is over-simplified, but our goal is to start implementing these loops into an NPC manager so that we can start building depth - food, wood, stone etc. The next step/phase is that we place two NPC managers side by side and have them 'share' a single resource. Then to flesh out how the AI will continue following closed loops on how two managers interact - maybe red villages just kills yellow village on contact over a forest because both managers have sent pawns to pathfind to that forest to gather wood. One of the major things that popped up here was that 'ages' seems to be a HUGE design nightmare. I don't know how Universim or original Godus design plan took that into account - how to add new loops, new mechanics and new interaction methods as technology was represented as progressing in game. It is super intimidating as a designer to even consider that as a viable option that doesn't blow up the scope of the game on an atomic level. Without worrying about that tidal wave of features, we can focus on exactly what you mentioned - the AI-AI and AI-player interaction. One of the design pillars I'm looking at is drivers of conflict and resolution. That is the essence of how AI interaction will be represented, including how you intersect into those loops as AI behaviors play their actions out. In paper prototyping we play-tested one scenario; Green/Blue/Red factions spawn in parts of gameplay area and begin growing and seeking resources. Player observes as they start pushing into neutral population centers, notices that blue has ran out of forest land and has started intersecting with red's forest land. Red starts killing blue, and you jump into scorch the red pawns by divine intervention. Blue, whom you aligned with because they're nice missionary-fishermen, have +10 to their militancy because X-number of reds were burned in holy fire from player triggered event. Blue starts sending crusades instead of missionaries/trade caravans, convinced that their vengeful Blue-Goat-God is the right god. You cast heal on reds as your blue massacre their smaller villages, because this is not what you wanted, but now Red's are convinced they're on the right path too. etc. etc. We did a bunch of these rounds to think about how the game rounds would play out, trying to immediately cut away anything that seemed like a deviation from the idea - like trying to 'send' blue to red. I think there is merit in trying to build systems that remove your direct control, instead make your NPCs free-willed little puzzles that might misinterpret what you did. Because we can relate to that in the real-world context of the subject matter, the whole 'reinterpret God's actions for Holy War' etc. I attached an example of one of the diagrams we filled out as we were discussing these mechanics - keeping in mind that they in turn inform/generate technical tasks. On the topic of environment shaping, I did a bit of a premature firing on that. I assumed it'd be stupid hard, turns out our programmers know how to work with Navmesh and basically said it's an extension of scope, like RTS pawns can pathfind around updated navigational mesh changes based on landscape change. I'd move away from fine sculpting the land, but it might bring in the option to cause broad brush changes, like land bridges etc. Hopefully we can start prototyping that soon. I know about Reus but I've never played it, I might give it a whir this week and have a good look at the mechanics. Fantastic art style. I spent some time this week also looking at Settlers and Banished, as I was doing art tests on NPCs, but I'll report back how that progresses. Attachments:
|
|
|
|
Post by echocdelta on Jan 10, 2016 5:50:25 GMT
Oh and an art test shot from just turning some sketches into 3D mock-ups. This isn't anything past exploration into art styles, but I thought it might be fun to share here.  |
|
|
|
Post by Aynen on Jan 10, 2016 10:31:44 GMT
I'm really liking your green/blue/red paper prototype! It's already very close to the kind of god-game I'd love to play!
In terms of art style I'd recommend you find a style that doesn't scream 'miniature'. Try and set yourself apart from that style. Even consider abstract styles if you have to. You'll really want to set yourself apart visually.
|
|
Lord Ba'al
Supreme Deity
Posts: 6,260
Pledge level: Half a Partner
I like: Cats; single malt Scotch; Stargate; Amiga; fried potatoes; retro gaming; cheese; snickers; sticky tape.
I don't like: Dimples in the bottom of scotch bottles; Facebook games masquerading as godgames.
Steam: stonelesscutter
GOG: stonelesscutter
|
Post by Lord Ba'al on Jan 10, 2016 10:46:24 GMT
I'm really liking your green/blue/red paper prototype! It's already very close to the kind of god-game I'd love to play! In terms of art style I'd recommend you find a style that doesn't scream 'miniature'. Try and set yourself apart from that style. Even consider abstract styles if you have to. You'll really want to set yourself apart visually. Not sure I get what you mean by miniature. Wouldn't a zoom option take care of that? |
|
|
|
Post by echocdelta on Jan 10, 2016 10:46:38 GMT
I'm really liking your green/blue/red paper prototype! It's already very close to the kind of god-game I'd love to play! In terms of art style I'd recommend you find a style that doesn't scream 'miniature'. Try and set yourself apart from that style. Even consider abstract styles if you have to. You'll really want to set yourself apart visually. I think it'll be clearer on those prototypes once I can open up the trello board for public view. The interesting thing was how they highlight the points of complexity - what happens when pawns meet each other from villages. The stuff inside a village is far less complicated in response. We're hoping to prototype those interactions being driven by each cluster's manager, instead of outcome being driven by the encounter itself. Yeah, my background is in much more complicated texture/material design than vertex colors. I'm looking at Banished at the moment because I need limited draw calls - maybe have a texture atlas and more complicated geometry (geometry on meshes on profiling is coming up pretty cheap). I'll keep doing art tests, it's a process I'd like to see people give feedback on. I'll definitely do a new round on something a bit more grounded into less flat-shaded/low-fi route, to see how that looks too. |
|
|
|
Post by echocdelta on Jan 10, 2016 10:52:25 GMT
I'm really liking your green/blue/red paper prototype! It's already very close to the kind of god-game I'd love to play! In terms of art style I'd recommend you find a style that doesn't scream 'miniature'. Try and set yourself apart from that style. Even consider abstract styles if you have to. You'll really want to set yourself apart visually. Not sure I get what you mean by miniature. Wouldn't a zoom option take care of that? Nah it's DOP (depth of field), we won't have that in VR (I hope/maybe/pending art testing), but the design of the meshes themselves can influence those outcomes. Buildings can look small by play of proportions and shapes, like tiny doors or windows, weird angles or low-fidelity geometry. Zooming in VR is actually insanely immersive, you lean down and look at a house, it makes you feel like a titan bearing down onto this one little hut. Post-processing effects, like DOP, can be just disorientating, even City Skylines I feel has it too strongly. Regardless, a note on art design (quick one) - it's easier to do realistic low poly than stylized low-fidelity, because one is about decimation of real-world reference whilst the other is interpretation into interesting shapes. I think branching out to both options and being able to compare is important. Remember, you'll always have the chance to 'backhand' a building... My favorite anecdote so far is that I was trying to juggle houses and dropped one. I picked up the rubble of the smashed house and sprayed it onto another village. You don't need to code in meteorite miracle, you can just lob house rubble like a homeless crazy person. |
|
|
|
Post by echocdelta on Jan 10, 2016 11:07:44 GMT
Oh and quick note, on paper prototyping the question came up of 'well, how the hell do we choose a people?'
Great question that, because how do we do that? If you impress one village enough, and they believe in you, isn't it weird that they somehow know who/what you are? We tend to make Gods from mythology or perceived signs.
Two tracks of solutions came up-
Player selects chosen people and thus selects 'avatar' or faction based on miracle casting
-OR-
Player's idol comets down into the world at start of play - player grabs their idol, walks over to the town they want to adopt and smashes their idol on the top of the old idol. Everyone in town takes that as a sign of 'TRUE GOD MOTHAF---RS' and immediately become your people. They get really into being your people, hence the interactions always being 'have you heard the word, friend, of our lord and savior Purple Monkey God?'
The second system is something we could prototype immediately after NPC managers, even rudimentary, are in. Also we can scope that out for things like progression on your 'idol' and your believers building around your idol to represent their character etc - like dumping weapons around your Purple Monkey God idol. It's basically NPC cluster updates state because an object is brought to an area. Boom, banner colors update (I assume maybe an RGB node value) and everyone in little peaceful village is now Purple Monkey fo' lyfe.
|
|
