Is this game idea do-able?

[Vicktor Karloff]

Well, to get some insight or advice from the game development people.
I have checked out a few games to get a feel of what people are doing out there. I have noticed that most seem to be on rails and that is something I would like to avoid in the game I would like to make. So here is my idea and I would like to hear your feedback on this and if it's something one person can do?
I like Corruption so it will be that type of game. I envision an "open world" this to me seems the easiest part as I have built massive maps before in NWN when that was a thing.
Where I do see a problem for me is the programming part. Is this even do-able thing? I want the player to create a character (Race, male or female) and the stats would determine how people would interact with them. A high Corruption stat would help an NPC do a thing that is "darker" once seduced, but it would creep people out and make it harder to seduce them. Items they wear (Cloths, shirt, etc.) would help, but not that much. They would be able to choose a career like a teacher, Police officer, priest, businessperson, Plummer, etc. This would allow them to discover opportunities to do stuff.
I do not want to limit the player either. if they are playing a male and they want to seduce a Female, that is fine but also allow them, if they want, to seduce a male. I want to give freedom to the player to explore all "acceptable" fetishes if they want. They want to seduce a married woman, sure. And then try to have her include her husband? Fine. Or corrupt a young adult (+18) to play with the family, used drugs or your pet? Ok, who am I to judge?. If they want to kidnap someone and take them to a basement as a pet or "hurt" them they can do that. But, that and all actions must have to some negative repercussions. Like they get caught and go to jail and spend the rest of the game in jail, fighting or fending off becoming the prom queen of cell block 4. Or the husband kicks their ass. They don't use protection? They get an STD, get someone or become pregnant themselves.
The player would have to work to get $$$ and seduction, affection, corruption progression would be slow. Forcing the player to focus on a few Storylines. If they ignore an NPC the NPC Stats for the player would go down.
I know this got way long so I will stop here. I would like to hear your thoughts on this.

 

[SeventhVixen]

I'm guessing that you are talking about an open world where the player can do a lot of kind of actions, to any npc, so the actions are scripted in general, or no character by character basis.

I see it as a quite easy project to do in Unity, having some basis on scripting an relying in Assets from the Store. It's not even heavy on the procedural stuff. Yo only have to program every action, and that action is targeted to X target, and any npc can be the target, an everyone has a fuck ton of variables/stats. There are good assets out there, like... Love/Hate I think is called, that calculates relations by emergent gameplay in a mass scale of characters.

Is perfectly doable to make a prototype with limited interactiivity in some months, but if you are not fond of programming and asset jamming, let's say at least a year to make something little and nice.

 

[Vicktor Karloff]

Thank you for the feedback. what other programs would I need other than Unity?
And what else would I need?

 

[SeventhVixen]

Well, in a simple point of view, all you is need is Unity (8), it's quite a standalone application for game making . Graphics are other thing. But with the fuckload of work (and by months I mean working full time), I think that project deserves some quite prototyping cylinders until the logic works..

I can recommend you some Unity Assets, like the Love/Hate. You can use a procedural map generator like Dungen or others that make cities to make fast placement of buildings and npcs. MapMagic to make a infinite procedural 3d map if you want it 3d. Inventory pro to handle inventory/crafting/selling/banking in one go....

It's a complicated project, you'll need to keep the idea very simple if you want to test it out.

 

[HiEv]

I see it as a quite easy project to do in Unity,

OK, maybe I'm missing something, but from what Vicktor Karloff described, this requires levels of AI which are currently unavailable. It couldn't be more the opposite of "quite easy".

The number of examples he gave in his opening post alone would take years to work into a game, and he appears to want to make it even more free than that.

I'm sorry, but I'm going to have to totally disagree here. This game idea, at least as currently described, is totally not do-able.

The reason why most games "seem to be on rails" is because you'd have to build a ton of real-world information into the game before it could even begin to handle things the way the real world would. And that says nothing about making it entertaining as well. At this point, having a narrative that leads to maybe a dozen or so significantly different conclusions is about the best you can do.

Think about it. The game would have to be able to estimate the moral judgements that people in the real world would make based on the knowledge and resources available to them. You'd have to simulate the mind, at least to a certain simplified degree, of every character involved, and then have them react based on their personality and knowledge.

This is no small task, and that's the reason why you don't find any games like that. It's simply not possible to do currently.

If you wanted to get something even close to this, you'd need a very limited set of characters in a very limited environment. Say, for example, a couple of people trapped on a small deserted island. That would make the decision matrix small enough that perhaps you could simulate it.

Good luck!

 

[anne O'nymous]

I envision an "open world" [...]
I want the player to create a character (Race, male or female) [...]
Items they wear (Cloths, shirt, etc.) would help, but not that much. [...]
They would be able to choose a career like a teacher, Police officer, priest, businessperson, Plummer, etc. [...]
I do not want to limit the player either. if they are playing a male and they want to seduce a Female, that is fine but also allow them, if they want, to seduce a male. [...]
I want to give freedom to the player to explore all "acceptable" fetishes if they want. [...]

I don't want to discourage you, but either you are describing a text game, or you already have a team of twenty behind you to do the game.
Look at what we actually have on the scene. Globally speaking, when a game give a freedom to the player, it's at the cost of another one. Not because it's impossible to have all what you describe is possible in a single game, but because it's impossible to do it almost alone.

By example, if almost only 2D games offer the possibility to decide of the clothes, it's because doing this with 3D CG imply that you'll have to render each scene one time by possible combination of clothes. So, if you need 10 minutes to render a scene, which is faster than the average time needed, with just 2 shirts and 2 skirts, you'll need 40 minutes for each scenes. Add 1 shirt and 1 skirt, and you'll now need 90 minutes. Add the possibility to choose the gender, and it become 180 minutes. Add two races, and you'll need 360 minutes (6 hours) for each scenes.
All this just talking about the rendering, assuming that you have a computer dedicated to this and optimized for this. Because you'll also have to create the scene, and make few changes because a pose will not looks as well when you'll change this piece of clothing.

To this, you'll have to take count of the interaction between the MC and the characters, as well as the interactions between the characters. Like each character can be corrupted, and that the MC can be male or female, it imply that each one of your characters will have three types of interactions, so 3 sets of dialog ; "corrupted by a male MC", "corrupted by a female MC", "neutral relationship". If you want to have seduction corruption and dominant corruption, it pass to 7 sets of dialog, since you'll need "corrupted by seduction", "dominated", and "in between", this for male MC and for female MC.
On top of this, you'll have to take count that, in an open world, the player can do what he want, when he want, in the order he want. You can't have predefined dialogs, they must be more dynamic and take count, at least, of the major actions of the player.

In the end, take a look at how many times was needed to write a game like a Fallout or an Elder Scrolls, and how many people was needed for this. Even by taking in count that you'll don't need to code an engine, nor to create the models and textures (whatever because you use an Illusion studio, Daz3D or decide to do a 2D game), it still stay a really big project that will need many time and many peoples.

 

[Winterfire]

Yes it is doable and it is not true you'd need a big team either.

However, not to sound harsh but if you need to ask what software you would need other than Unity to do what you have described then I can assume you have never completed a project of a smaller scale... If that is the case, you should start small first, the good old K.I.S.S. rule.

If you start with such a project, you are very likely to stop soon, either because you begin to see how much there is behind every thing you have described or because you begin to see newbie mistakes and feel tempted to start over which often means the project's death.

-edit-
Also, having too much stuff going on is not always a good thing, part of game developing is also giving the player a small dosage of everything in a way so that they feel they are progressing, if you throw everything at once they will feel overwhelmed.

 

[GuyFreely]

I might be repeating a bit, but the reason you don't see a lot of big, open, do whatever you want kind of games is that someone has to make all those scenarios. Anything is doable with enough time and effort, but those are generally limited. There are ways to limit the amount of work required. Using a 3D engine means you don't have to make a whole ton of renders. Creating a set of animations that can be re-used means you can plug and play the actors. Then you're down to writing all the individual dialogue lines, logical branching, stat checks, etc to determine the path of any event.

I think of something like the mod scene in Skyrim. The engine is in place, models, clothing, animations are all available. Individual mods in that game are very often limited to one story line or mechanism. Like you can prostitute random NPCs with a built in set of limited dialogues. Some prostitution mods add specific scripted encounters as well. At the end of the day, you have to install a whole bunch of mods created by a bunch of different people to approach anything like what you're describing. So if you are a one person team, expect a lot of work.

This is why you'll find a lot of games pick one scenario or situation. Some of them might seem on rails, but I think most devs try to add a few paths that lead to different outcomes. Depending on how many activities there are and how many are mutually exclusive or hard to find, you might be spending time making things that very few players even see or experience. In terms of efficiency, you get the most "return" out of making something that everyone playing your game will see. Let's say your game starts with a sex scene, 100% of players see that scene. Let's say you have five ending sex scenes based on choices in the game. If people play once, then 20% of players will see any given one. It's true that some people might give it another go, but how many players are going to see all five scenes? Chances are it's not 100%. So the more convoluted your paths and choices become, the less efficient your work becomes. It's a bit of a cynical view of things, but I think it's something to consider.

 

[SeventhVixen]

OK, maybe I'm missing something, but from what Vicktor Karloff described, this requires levels of AI which are currently unavailable. It couldn't be more the opposite of "quite easy".

Well, I wasn't want to be discoraging, but I think a across my two post I said thre times "hard" and said that with little knowledge it can be a year of work (and I meant full time, as I said in my second post).

AI is not unavaible in Unity, there are some good assets out there like Behaviour Designer (But I think others are better). With that asset in particular, if you are making one generic NPC multiplied by an entire human population where his drives comes from a RNG stat pool of desires and other numbers, you only need to do some gigantics behaviour trees.

As I said, and what I didn't said I make it more clear (because I lack the level of knowledge of the OP but if he did bigass maps on NWN at least knows a bit of level design and scripting quests in a predetermined asset) making such game will require big practice. If I were to throw buckets of water, Years of practice in Unity.

But it you have this practice, or you know what assets to use, if you don't derivate the project from one thing to another constantly... you can do a prototype (A prototype) in a year? Yes. Is easy and many parts don't requiere high levels of programming, just the programming levels of asset-kit-bashing? Yes.

And thats when I said graphics is another thing. If someone tries to do this game, and do also some level of competent graphics, the time requiered is exponentially multiplied.

PS: Easy, but a fuck ton of work, making errors, rinse and repeat. Maybe I'm saying easy because I'm taking apart the human factor of just being desesperatly stuck on a programming issue and just throwing the project trought the window.

 

[Vicktor Karloff]

Hey HiEv, anne O'nymous, Winterfire, GuyFreely, and SeventhVixen.
Thank you for ALL the input. I know it will not be easy and time-consuming but, that is true for all hobbies right? But you did open my eyes a bit. I was under the impression that it would be, with the use these programs, more of a drag and drop, if "this" then "that", or to just give value to the items. I came at this as a, just build the world and let the PC finds their own way. I know, it was a bit nieve of me. What I did take out of all your posts is, yes It can be done. No, it will not be easy. But two guys, with a fuck ton of skill and know-how, did make Cuphead. It did take them some time but, they did it. I will try to answer or comment to each of you below. And, if I can count on the people that took the time to reply to a noobs post, I'm sure I can count on you all to give me the same kind of honest feedback. Who knows, in one or two years you will be playing something I, along with other people from here, created.

HiEv
"If you wanted to get something even close to this, you'd need a very limited set of characters in a very limited environment. Say, for example, a couple of people trapped on a small deserted island. That would make the decision matrix small enough that perhaps you could simulate it."
-F'n awesome Idea! Start in a small fishing village with a few 10 "usable" NPC's-

"Think about it. The game would have to be able to estimate the moral judgments that people in the real world would make based on the knowledge and resources available to them. You'd have to simulate the mind, at least to a certain simplified degree, of every character involved, and then have them react based on their personality and knowledge."
-Could this not be done by "Skill" checks? And randomize the "Personality" stats of the NPC's the moment the PC starts the game for the first time?-

anne O'nymous
Telling an honest truth is not discouraging, it's speaking the truth. Who would get upset at that?
"On top of this, you'll have to take count that, in an open world, the player can do what he wants, when he wants, in the order he wanted. You can't have predefined dialogs, they must be more dynamic and take into account, at least, of the major actions of the player."
-I was thinking that I would address this by, depending on what paths the PC would do, it would cause some storylines NPC's to not be available to the PC. I was going to base this on the carer they choose. Would this not work?-

Winterfire
Very true what you say about noob mistakes and starting over. I can see that happening.

"Also, having too much stuff going on is not always a good thing, part of game developing is also giving the player a small dosage of everything in a way so that they feel they are progressing, if you throw everything at once they will feel overwhelmed."

-That is why I liked HiEv's idea of a small Area to sharpen my teeth with and your Idea of "Keep It Simple Stupid". If I get the NPC personality and PC to interact well... How would one do something like that?
And another quick question, could make a "pool" of speech or texts that NPC's would pull from depending on the situation?

GuyFreely
-Thank you for your input. And you do bring a good point. but the fun part of gaming is finding all the outcomes. I think that the idea of Randomizing the NPC's "personality" and look after the PC creation when you start a new game will make the player want to play again. The girl you met at a bar in your 1st playthrough that you picked up could be a Black male that hates your ass. Still, have to think more on the points you bring up. I would like to hear your thoughts some more-

SeventhVixen
Thank you for all your feedback.
" Easy, but a fuck ton of work, making errors, rinse and repeat. Maybe I'm saying easy because I'm taking apart the human factor of just being desperately stuck on a programming issue and just throwing the project through the window."
-Like I said before I'm sure if I run into something I can't figure out, I could get some help from people here.-

Now, I do plan to, as in any hobby, use some pocket money to outsource some things that I can not do. I did also plan on using RPG maker to make this game and mix it with interactive Animation for the adult scenes. I had seen some games do this. But with this feedback, I think I may have to rethink this.
Again thank you for all this information. I look forward to seeing what other pointers you guys (and gals) add to this.

Should I start by working on the stats and what they will do, build the world or work on the story?

 

[Winterfire]

-That is why I liked HiEv's idea of a small Area to sharpen my teeth with and your Idea of "Keep It Simple Stupid". If I get the NPC personality and PC to interact well... How would one do something like that?
And another quick question, could make a "pool" of speech or texts that NPC's would pull from depending on the situation?

Should I start by working on the stats and what they will do, build the world or work on the story?

"If I get the NPC personality and PC to interact well... How would one do something like that?"
I do not understand the question, do you mean what interactions they could have? I suppose you could have different batches of dialogue, one for each personality.
So an outgoing personality would say "Hello there! How is it going?" whereas a shy personality would simply say "..." or "H-Hello...", much like it happens on AA2, I believe.


"And another quick question, could make a "pool" of speech or texts that NPC's would pull from depending on the situation?"
Yes, you can write different dialogues in different text files: Have them in a list, get the file you need depending on the variable you want (Day, relationship stat, ...) and assign it to the reader on/before interaction.


"Should I start by working on the stats and what they will do, build the world or work on the story?"
You normally do all that in what would be your game design.
Story->World->Stats would be the order I'd do it.

 

[Vicktor Karloff]

"If I get the NPC personality and PC to interact well... How would one do something like that?"
I do not understand the question, do you mean what interactions they could have? I suppose you could have different batches of dialogue, one for each personality.

I meant how would I go about making the stats interact or to give the NPC's personality. I don't have the remotest Idea how to do that.

 

[Winterfire]

Well, that can be done in MANY ways and it highly depends on where you will make your game... Generally speaking (keeping in mind that this might not be the best way to do things) I would make a Class "character" where I would define all the components of a character, personality included.

Personality would be an enum, basically a tidy list of all possible personalities that you want to have in the game, upon character creation the player could choose which personality to have and, as far as NPCs go, you could simply assign them manually or randomly.

I do not understand what you mean by "stats interact", what do you define as stats?
By stats I presume you mean relationship levels, stamina, Hit points, stuff like this... They are normally just integers, they do not really interact but are interacted upon (You get slashed and you lose hitpoints, you get healed and you gain hitpoints, you rest and gain stamina, and so on).
The only interaction I can think of is calculation... Let's assume a prostitute character has a charisma stat, that could affect the money she will receive upon performing her task, something like money = Charisma*Base money.

 

[Vicktor Karloff]

thank you!

 

[Winterfire]

I am speaking thinking about Unity or similar btw, I saw you have mentioned RPG Maker so I am not sure how you would go with that as I am not too much familiar with it.

I can give you some code examples in C# so perhaps you can better understand my point if I was not too clear or you can just begin with whatever engine you have chosen and ask questions about a subject if you meet any difficulty.

 

[SeventhVixen]

In Unity, About the drag and drop thing, in Unity you can use a Visual Script like

You must be registered to see the links

and a ton more of related assets that you can work half the game out of it without "actual" programming language. (But it's more legwork)

Also, there are asset that I've never used,

You must be registered to see the links

, for example, that mix visual script with an already rpg system that ties quests and dialogues and stats. There is a load of assets like this, better or worst. The most important thing that you have to know is how to expand and mix assets between, and that will need precoded integration between assets (there are a lot of integrations already done by the developers), or you have to code your integrations (that is what I usually do and where I find most of the problems as a non programmer).

About random dialogues and other easy randomising pools you may have Array list/hash tables that allow you make dozens of pools of texts so you can randomize easily what people say in determinated moments so conversations are diverse. You can also use those array list/hash tables to randomize stats pools. So you can create characters at runtime and appareance and stats can be randomised, and you don't have to pregenerate hundreds of npcs. (maybe you can manually tailor some, and let the bulk be randomised).

This is a

You must be registered to see the links

with other assets for playmaker. If you are interested in a visual script tool you can check here the possibilities of using diverse assets to take scripting work of your shoulders.

 

[anne O'nymous]

-I was thinking that I would address this by, depending on what paths the PC would do, it would cause some storylines NPC's to not be available to the PC. I was going to base this on the carer they choose. Would this not work?-

It can work, and it's in the same time an illustration of what I said. To be able to offer to the player the freedom of roaming how he want, you remove the freedom to have all the possible path in a single play.
Mostly, it's the problem with games this ambitious. To achieve it, you need to make compromise between what you want and what will be available.

 

[HiEv]

HiEv
"If you wanted to get something even close to this, you'd need a very limited set of characters in a very limited environment. Say, for example, a couple of people trapped on a small deserted island. That would make the decision matrix small enough that perhaps you could simulate it."
-F'n awesome Idea! Start in a small fishing village with a few 10 "usable" NPC's-

I'm doing something similar in the game I'm working on, where you have a mostly isolated group, with a limited number of options for what they can do, in order to make this even remotely possible to simulate.

That said, the main plot of my game is still "on rails", it's merely the (very simple) side-stories which can dynamically develop in limited ways.

Also, 10 may still be a little large. Keep in mind that the number of possible interactions between the PC and NPCs is (2^N)-1 (where N = the total number of NPCs + the PC). So with 10 NPCs that's 2,047 possible combinations ( (2^11)-1 ).

If you have the PC and 3 NPCs then that's 15 combinations of interactions to track/test:
15 = 1, 2, 3, 4, 12, 13, 14, 23, 24, 34, 123, 124, 134, 234, 1234

If you only pay attention to interactions with the PC, and not NPCs with each other, then it becomes a 2^(N-1) number of combinations, which isn't much better. So the PC with 3 NPCs would give 8 combinations, instead of 15.

There are ways to cull those numbers, such as by having some NPCs act the same or not allowing some NPCs to interact with each other, but I think you can still see how quickly the problem will usually double in size for each NPC added.

"Think about it. The game would have to be able to estimate the moral judgments that people in the real world would make based on the knowledge and resources available to them. You'd have to simulate the mind, at least to a certain simplified degree, of every character involved, and then have them react based on their personality and knowledge."
-Could this not be done by "Skill" checks? And randomize the "Personality" stats of the NPC's the moment the PC starts the game for the first time?-

The problem is, how do you know which skill checks to apply? And what outcomes the results provide? And how does the personality interact with that?

For a very simple example. Say you set up a routine where people react badly to stealing stuff from them. Just keep track of who owns what, and set a "theft" flag when the player takes something that belongs to someone else, and then make the NPC react appropriately angry. Easy enough, right?

Now imagine the player gives an NPC something, say by leaving it at their house, and then changes their mind and takes it back before the NPC finds out. You would have to write code so that it would recognize that the NPC wouldn't count that as theft because they never knew that they "owned" the item very briefly. Otherwise you will end up with a "psychic" NPC angry at the PC for stealing from them, when there is no way they could have known that they owned the item, much less that it had been taken from them.

Or, say that you make an aggressive move against an NPC. Typically the general responses are fight, flight, or freeze. What skills determine what reactions? And what things in the environment affect that? And what skills determine what they will do after the confrontation? And how would the computer determine all of those answers?

To give a specific example of that, what if the NPC is being attacked, but they know that their friend is going to be there any moment now? You would have to program them to be able to plan ahead before they could realize that they need to delay the attacker until their friend shows up.

These are all complicated artificial intelligence problems, and, as I said before, that's why you don't see this in any of these games.

(FYI - I double-majored in Computer Science and Cognitive Science/Artificial Intelligence in college, so I have a bit of knowledge in this area.)

"On top of this, you'll have to take count that, in an open world, the player can do what he wants, when he wants, in the order he wanted. You can't have predefined dialogs, they must be more dynamic and take into account, at least, of the major actions of the player."
-I was thinking that I would address this by, depending on what paths the PC would do, it would cause some storylines NPC's to not be available to the PC. I was going to base this on the carer they choose. Would this not work?-

I'm not 100% clear, but I think the answer is no. You'd still have to give the NPCs some way to generate responses to all of these situations.

Also, keep in mind that every career you add is like adding another 50% more work to the amount of work it would take to do just one career, since you have to add all of the additional information for that path to work.

"Also, having too much stuff going on is not always a good thing, part of game developing is also giving the player a small dosage of everything in a way so that they feel they are progressing, if you throw everything at once they will feel overwhelmed."

-That is why I liked HiEv's idea of a small Area to sharpen my teeth with and your Idea of "Keep It Simple Stupid". If I get the NPC personality and PC to interact well... How would one do something like that?

Well, with the game you're describing, the best way to do that would to be to dangle some goals (plot threads) in front of the player, along with an idea of how to head towards those goals. Then the player should be able to quickly narrow down to the goal or goals they'd like to achieve. Having a "quest" list they could look through would help here.

Also, instead of dropping the player in the middle of a world and just saying, "Go!", building a narrative for the PC to achieve helps them feel less lost or overwhelmed in a game, even if it's a very general goal (such as "survive").

And another quick question, could make a "pool" of speech or texts that NPC's would pull from depending on the situation?

Well, yes, but the trick is having enough pools to cover all situations and making the game capable of knowing which pool to pull from. That's the hard part.

Should I start by working on the stats and what they will do, build the world or work on the story?

I think you need to start "backwards" and come up with some scenarios you'd like to be possible in your game. Then look at all of the possible ways which could lead up to that scenario and how it could go based on the personalities and skills of the characters involved. And then try to break the scenario down to determine what information would be needed to make all of those possibilities work.

Then repeat with another scenario, and another, and another, and so on, until you stop hitting new bits of information you would have to deal with. Then you'd need to figure out how you could code something to handle all of that data and for the game make realistic decisions based on the various situations which are possible.

Like I said before, I don't think it's do-able without lots of limitations, because you're getting into some really difficult AI problems which haven't been solved yet.

I think the closest any game comes to doing what you are looking for is

You must be registered to see the links

, so you might want to take a look at that and

You must be registered to see the links

to get an idea of how the NPCs in that game interact. The "

You must be registered to see the links

", where you can play as a PC, is a somewhat recent addition, relatively speaking. (For about the first half of Dwarf Fortress' development, "

You must be registered to see the links

", where you were more of a "disembodied director of dwarfy activity", was the only mode available.)

Keep in mind that this is the work of one developer with some help from his brother, and they've been working on it since 2002, with their first release in 2006. Currently they estimate that the game is around 45% complete.

Good luck!

 

[Vicktor Karloff]

WOW! Ok, I'll need to ponder.

 

[uradamus]

I started to lose interest about half way through this thread, so forgive me if I'm repeating anything here. To make a game open and do it on your own or as a small team you will have to make a lot of sacrifices to make it work. If you don't have programming experience, you are going to suffer through this and most likely either have to stick to really dumbed down mechanics or find premade mechanics (which will likely cost you some money, like in the Unity store or going with something like RPGM that spoon feeds you all the necessary bits). If you don't have art skills things are even more grim, you'll have to take your best stab at making your own and probably doing a terrible job of it, using often reused assets (which will likely cost money), hiring someone to do your art (which will cost even more), or stick with a text based game.

If you decide to go the Daz or Illusion routes, you will be limited to rendered scenes and any variety you want to account for will come at a major price in duplicated efforts, re-rendering the same scenes over and over again with slight variations. If you go the full 3D route, you need some serious skills or afford the work of someone who does. If you go 2D, variety is again your bane, but it's a bit easier to separate backgrounds, characters and clothing, so once the various assets are made you just need to deal with creating the spaghetti code for all the sprite swapping and keeping track of which ones should be used at any given moment.

If you want the dialog to be even remotely interesting, it can potentially take you years to write it all even before considering all the crazy scripting that will go into putting it to good use. You'll likely have to settle for pretty generic and/or scripted interactions. Usually you can do things like inserting variables into dialog to change what gets said depending on the context of the interaction, like changing the pronouns or names being used, or maybe even swapping whole sections of text based on what flags have previously been raised. The more variable you make interactions, the more tedious, bug prone, and time consuming it becomes, even if you can rely on some drop in dialog system.

Now for some practical advice. If you want the most variety with the least effort, consider doing something a bit like

You must be registered to see the links

. It has many fairly short and mostly linear main story paths you can go down, each with a different girl with a unique personality and their own kinks and fetish boxes they check. You have a basic stat system and a bunch of mini-games that decide how successful you can be with any given girl's route. There are also a ton of side characters for one off events you can run into around the map, some good, some bad. They make for a good outlet for a bunch of content that doesn't fit well in one of the main routes.

The movement is basic/classic point and click adventure fare, you click on image buttons to transition between different maps or to interact with npcs or important items in the environment. The NPC interactions are basic VN style for the most part. All of this is fairly simple for even someone with minimal experience to get going in most general purpose engines with a bit of time going through tutorials and asking the relevant communities surrounding your engine of choice for help along the way as needed. Stats aren't that hard to code for either, but do keep them as simple as possible if you don't want to drive yourself insane with accounting for nearly limitless possibilities. At there most simple, you would implement them where you just do if/else checks against the relevant stats to decide if the player gets to progress or whether they get a good or bad outcome if they are allowed to move forward without meeting a stat requirement.

You can feasibly do everything as a text game at the start. All the scriping/logic/dialogs will be pretty much the same with or without graphics, you'd just have to swap a few things out with how interactions are handled and add an extra routine for your main game loop to handle the graphics if/when you add those by whatever means.

You really should back burner grand ideas though. Start simple, get your feet wet with an engine. I suggest Godot, but it really doesn't matter unless you want to do something that fits well with a specialized engine like RPGM or Ren'Py (for jrpgs or VNs respectively). Unity is an awful engine, but if you've got money to burn the asset store makes it an attractive option to quickly get up and going even if you've got no skills. If you are going to cheap out though, it's best not to get caught up in such money pits, as it will slowly drip you dry if you get sucked into that mentality of paying others to get over your shortcomings instead of putting in the effort to learn things for yourself.