I asked ChatGPT to fix our long-lasting conceptual issues

[Hume2]

There are a few conceptual issues, some of which were already discussed here, which prevent developers from creating a rich-experience minetest_game-like game. I used mostly MINECR~1 as a reference, because I think that it may give more meaningful answers, but they are easily transferrable. Here are some highlights, feel free to inspire!

1. A story involving NPCs either requires them to be invincible, or the player may get easily soft-locked.
Solution: Make NPCs renewable and allow them to take roles of dead NPCs.

• When a player harms a village, the villagers remember it and refuse to continue the quest, until the damage is repaired (at least roughly) and a compensation fee is paid.
• Damage can be detected by assigning metadata to the blocks owned by villagers. Repairs are not checked for continuously, but only when the player "asks for it".
• A village may posses a central indestructible "shrine", which can be used for respawning NPCs and may store information about the village.

2. Obstacles in dungeons can be easily bypassed by digging around. Making the dungeon walls indestructible feels like a cheat, especially when a player encounters it randomly when mining.
Solution: Allow digging around obstacles, but add penalties for doing so.

• Examples: Spawn mobs behind the wall, reduce rewards in chests if a wall was broken, apply bad effects to players digging through walls.
• The walls can be "cursed". Digging them may result in turning off torches, earthquakes and similar effects.
• The dungeon can be organised so the final room is in the core of the dungeon, so it is impossible to dig straightly to the final room without encountering the other rooms.
• There can be also trap rooms, which are accessible only by breaking walls.

3. There is no real reason, why a player should build a proper base. How to disadvantage rogue-like gameplay?
Solution 1: Add more advanced crafting machines, so moving them all is too problematic.
Solution 2: Make monsters harder to avoid, make them able to break through wider variety of blocks

• Bed shall skip time, but not risk. If you go to bed, the strength of your base is evaluated. And if it is too weak, then you can be woken up by a monster breach.
• Some monsters may target at stealing items, so protect your chests too!
• In some nights, the monsters can be especially tough. Use your time wisely for preparation.

4. Most of the challenges in caves can be skipped entirely by digging shafts through bulk rock.
Solution: Don't forbid that, but add its own challenges to it. Encourage the player to reinforce the shaft, else it may crumble at random.

• Do you see droplets of water from the ceiling? Be careful, it may flood your shaft!
• Is the wall hot? Lava may melt it soon.
• Do you see dust crumbling from the ceiling? It may collapse soon.
• There can be also acid, poisonous gasses, or maybe the player can suffocate from not having enough oxygen down there.

5. Minecarts are practically useless. (MINECR~1 only: Coal-powered minecarts are too slow, powered rails are too expensive) It is practically impossible to set routes dynamically. They are also unreliable because mobs may interfere with them, unless I prevent them from reaching the track. And, most importantly, I can mine way faster than I can build railways.
Solution: All these issues can be addressed.

• MINECR~1 only: Add a special track which isn't as expensive as the powered rail, but is slower.
• When a minecart hits a mob, it pushes the mob away, dealing adequate damage.
• Make it possible to link up minecarts reliably, so they can be used for bulk item transport across bases.
• Players can send signals LEFT or RIGHT from a passenger minecart. A special routing track sends a signal to the left or right respectively, when the player holds the respective arrow while passing the rail.
• There can be special tracks which send information to players riding along.
• Players may be encouraged to build more bases (for example one base deep underground, for easier mining). In such a case, the minecarts can be used for bulk transport across player bases.

6. Nether portals are too cheap and can be placed mostly anywhere.
Solution: Instead of adding a portal to a realm which is said to be deep underground, make players actually dig that deep.

• It adds more to the world lore.
• It encourages the players to invest in transport systems.
• End portals are fine the way they are, as they must be found in specific dungeons and cannot be transported.

7. High degree of automation encourages player to pile up exponentially many items, causing a huge portion of lag in public servers.
Solution 1: Make automatic machines wear out, especially when they are used too frequently or improperly.

• Machines can be still repaired, but it requires manual intervention.
• Players are encouraged to optimise their machinery and also to turn of machines they don't really need.
• Forceloaded machines may get higher chance to break.
• The damage may slowly revert, so if a small machine is used only occasionally, it never breaks.
• Improper usage (e.g. trying to put an item into a full container) may damage the machines more than proper usage (e.g. putting an item into a slot where it can be put).
• If a machine is triggered too often, it may overheat, leading to higher damage per use.

Solution 2: Limit the count of machines which can be placed within an area.

[Blockhead]

While some of the ideas aren't bad in isolation, they seem all unrelated and don't form a cohesive design. Did you actually get something out of the answers that will help you in your own modding? or the design of a game? or are you just proffering ideas out of the LLM that you thought were decent answers to a variety of questions?

[Hume2]

The original conversation is too long to be posted here, so I just chose the most important highlights. And I don't even see the point of doing so, as anybody can ask similar questions to ChatGPT on their own. It didn't just flush out these ideas at random, it always gave enough context, so a mod/game could be implemented according to it. I didn't ask for much specific answers, I was mostly examining possible ideas to deal with the given issues. Specific decisions can be made according to the theme of the game.

[Astrobe]

1. A story involving NPCs either requires them to be invincible, or the player may get easily soft-locked.

A variant of this one often sees in RPGs (e.g. shopkeepers in Nethack) is to make NPCs very strong and able to fight back. A generalization would be to have the NPCs remember the aggression and chase away the player. A "anger" cooldown eventually resets them to normal.
An extension or alternative would be to give the means for the player to attract NPCs - an idea I am using in my game to provide an incentive to build larger houses and more "bases" (NPCs don't carry stories in my game, though).

2. Obstacles in dungeons can be easily bypassed by digging around. Making the dungeon walls indestructible feels like a cheat, especially when a player encounters it randomly when mining.

The solutions are even more "cheat-y". Just make the walls very expensive to destroy - for instance they can only be blasted by an explosion if TNT or other ka-boom devices are expensive.

However IMO the main issue with dungeons (on servers), with regard to loot, is "first come, first served". And if you make the chests "repop" you either have an economy problem or worthless loot. I don't see a good solution other than having a GM refill the dungeons.

3. There is no real reason, why a player should build a proper base. How to disadvantage rogue-like gameplay?

Why do you dislike rogue-like gameplay anyway. Or you don't like nomads and explorers?

Anyway, this is solved by making it so that building more attracts more NPCs (see above), and when NPCs are the sole source of specific materials or items, you have your motivations.

There is an old thread about this topic. I think the title was "reasons to build more" - it certainly in my post history.

[*]Bed shall skip time, but not risk.

Disable night-skipping beds entirely. This is obviously a lazy and bad solution from a game design perspective. Nights are boring ? Create activities that are best done during nights. For instance in my game: smelting ores, repairing armor because they take time.

4. Most of the challenges in caves can be skipped entirely by digging shafts through bulk rock.

Make caves more attractive. You can do that by increasing the wear and/or dig time on/of the pickaxe, making "blind" digging an unattractive option.

5. Minecarts are practically useless.

Yes. So drop them. They are weird and buggy anyway.

Separate the logistics (transport of goods from where they are produced to where they are needed or transformed) from player travel if needed. Maybe find some way to let players hire other players as carriers. This can be a starting job for newbies. That's one less reason to use NPCs with ridiculous quests.

Players may be encouraged to build more bases (for example one base deep underground, for easier mining). In such a case, the minecarts can be used for bulk transport across player bases.

Yes, my dungeons feature teleportation beacons for that purpose. I also have "void chests", which can be used for bulk goods transport.

One can also provide supplementary purpose for caves once the player have mined-out them. A basic idea are large mushroom farms, but one could elaborate e.g. with specific large, underground-only crops, like giant luminous mushrooms. This way players can check on their underground farms, harvest, replant, while on their way to their digging site etc.

6. Nether portals are too cheap and can be placed mostly anywhere.
Solution: Instead of adding a portal to a realm which is said to be deep underground, make players actually dig that deep.

Yes. You can have different strata with different unique features. And with non-OP tools (see above), a good strategy for players is to "cave-hop" to them. Mapgen v7 with the right settings is pretty good for that.

It encourages the players to invest in transport systems.

One shouldn't need to "encourage" anything. Players have "problems" (most of the times its the job of the game designer to create them), you provide solutions to these problems.

If you have a solution that needs a problem to solve, it is not a solution; it is dead weight.

It makes me think about those people who absolutely want to put AI in their products but don't have an issue that an AI would tackle well.

7. High degree of automation encourages player to pile up exponentially many items, causing a huge portion of lag in public servers.

Automation is machines playing instead of players. I understand their appeal as a programmer, but they mean more engineering and less action. And then you have players who want night-skipping beds because they are bored at night. Automation should be done very sparingly.

[Hume2]

The solutions are even more "cheat-y". Just make the walls very expensive to destroy - for instance they can only be blasted by an explosion if TNT or other ka-boom devices are expensive.

Interesting solution, just note that the destructive item can now serve as a dungeon looting tool. And there should be also an explanation why the dungeons always have that hard walls, not seen anywhere else.

However IMO the main issue with dungeons (on servers), with regard to loot, is "first come, first served". And if you make the chests "repop" you either have an economy problem or worthless loot. I don't see a good solution other than having a GM refill the dungeons.

Good point, I asked ChatGPT on that, but it didn't offer me any solution which I would really like. Either, the dungeons could be claimable and players could fight over them. The monsters could also try to reclaim them. The dungeons could also procedurally revert to their original state. Or they could even disappear and reappear elsewhere.

This is actually a deeper problem and doesn't happen only with dungeons but with all resources. The early players get a horde of resources and the later players get only a fraction of what used to be interesting. This can be partially resolved by encouraging the players to settle further from spawn. ChatGPT even suggested me that farther areas could host more interesting resources and dungeons.

Why do you dislike rogue-like gameplay anyway. Or you don't like nomads and explorers?

I forgot to mention one thing: The main goal of the game usually turns out to be so simple, so most of the game's depth turns out to be completely optional. And building a base turns out to be one of the things which are rarely necessary. There was a thread about it, indeed.

Anyway, this is solved by making it so that building more attracts more NPCs (see above), and when NPCs are the sole source of specific materials or items, you have your motivations.

Yes, something similar was actually also suggested by ChatGPT, I just didn't find it interesting enough.

Disable night-skipping beds entirely. This is obviously a lazy and bad solution from a game design perspective. Nights are boring ? Create activities that are best done during nights. For instance in my game: smelting ores, repairing armor because they take time.

Actually, in both Minetest and MINECR~1, I didn't need to skip nights at all, once I reached a certain level of advancement. Exactly because of what you said, I had enough activities which could be done in both days and nights. But before that happens, I don't have anything to do. Or maybe I should dig a mineshaft from my house and go mining? Well, that could work.

Make caves more attractive. You can do that by increasing the wear and/or dig time on/of the pickaxe, making "blind" digging an unattractive option.

On the other hand, slower mining times make building tedious.

Yes. So drop them. They are weird and buggy anyway.

Separate the logistics (transport of goods from where they are produced to where they are needed or transformed) from player travel if needed. Maybe find some way to let players hire other players as carriers. This can be a starting job for newbies. That's one less reason to use NPCs with ridiculous quests.

Also a solution, but I like trains :D

One can also provide supplementary purpose for caves once the player have mined-out them. A basic idea are large mushroom farms, but one could elaborate e.g. with specific large, underground-only crops, like giant luminous mushrooms. This way players can check on their underground farms, harvest, replant, while on their way to their digging site etc.

This actually looks like a good idea. Once the caves are mined out, they serve no purpose. So providing a supplementary usage is a good way to go.

One shouldn't need to "encourage" anything. Players have "problems" (most of the times its the job of the game designer to create them), you provide solutions to these problems.

If you have a solution that needs a problem to solve, it is not a solution; it is dead weight.

Indeed, but I would argue that some games are already in the situation, that they have too many clever solutions, but no actual problems. Many game mechanism can still stay optional, but they should be at least considerable. If you look at MINECR~1 speedruns, there are so many mechanisms which they don't use and they don't even stay as a considerable option.

Automation is machines playing instead of players. I understand their appeal as a programmer, but they mean more engineering and less action. And then you have players who want night-skipping beds because they are bored at night. Automation should be done very sparingly.

Also a solution and I wouldn't protest against it at all. It actually depends on how much it fits into the theme of the given game. And unless the game is certainly technically-driven, then I see no point in implementing any automatic lines.

[Astrobe]

This is actually a deeper problem and doesn't happen only with dungeons but with all resources. The early players get a horde of resources and the later players get only a fraction of what used to be interesting.

Yes, and this makes it difficult to tune a game both for offline and online. In the online case, to some extent veteran players can help newcomers. Some players do it naturally, not just their friends, but if the game adds a benefit for the helper, it can be a win-win situation. Some activities can become tedious in the long run, like hunting some animal for their hide, but for new players it is new and maybe exciting.

Still, when one stumbles upon a dungeon or a cave and there are signs that someone was there before, it is a little discouraging. I believe the game "balance" is in the opposition of a tendency for players to spread because of resource availability, and the tendency to stay together for safety and trade.

A possible way out is with exploration. I joked earlier about nomads, but actually new players don't have a home, don't have farm plots, etc. so their best option is to roam around (the problem, though, is that they'll have to be patient if they actually just want to play Farmville). A nice place could also be a piece of information they can sell.

A game could create or improve that kind of beginner's advantage. I could do that in my game for instance, as leveling up is the players' decision (provided they have the resources for it). There's even an example of this already: they lose the ability to summon NPC traders to gain one ability of their choice when leveling up for the first time.

I forgot to mention one thing: The main goal of the game usually turns out to be so simple, so most of the game's depth turns out to be completely optional.

In an open world and sandbox game, the game should, rather than having a main goal, have "affordances" of goals - or perhaps more simply put - suggested goals. I often use Bartle taxonomy as a starting point when thinking about the game. In that case, a game should provide at least one goal for each type of player. Preferably the goal should actually not be reachable - or "asymptotically reachable" (the further you progress, the harder it is to progress) and maybe also not "monotonic" (one can fall down and regress - dying is a typical case).

This can look like a scam, though :-). But unless you plan on building a game studio and create an endless stream of extensions like WoW, the game has to be open-ended, that is endless.

Actually, in both Minetest and MINECR~1, I didn't need to skip nights at all, once I reached a certain level of advancement. Exactly because of what you said, I had enough activities which could be done in both days and nights. But before that happens, I don't have anything to do. Or maybe I should dig a mineshaft from my house and go mining? Well, that could work.

Yes. This reminds me that in the early stages of my game one has sometimes to AFK because one has to wait for things to grow etc. Probably a solution using "grinding" can be found - a way to get some resource which is so slow that it is not interesting to do past some point in the game. My next release should introduce something like that - weeding out elephant grass to make paper or cheap fuel. It was intended as an online newbie job but could be used by singleplayer. But it is too context specific, I need more of this type of thing.

Also a solution, but I like trains :D

Then may I suggest BVE or OpenBVE? :D As a flightsim fan, I sometimes dream Luanti was better fit for that. Ironically, the "infinite world" project that I don't see as necessary would probably be a sine qua non feature in this case.

[Nininik]

is this gonna be another chatgpt dump similar to what eredberre did or whatever his username was

[Wuzzy]

This ChatGPT nonsense really annoys me.

If you need ChatGPT for good game ideas, I think you're missing the point of game development. Everyone has ideas, but the important part is to actually playtest them and see what works and what doesn’t.

And it keeps TRIGGERING me people keep pretending LLMs are like a magic bullet. The generated text LOOKS believable, but you can't draw any conclusions regarding the real world from it. They're language models, but they cannot reason. LLMs do have use cases, but I think they’re WAY too overhyped.

Playtesting is something you simply can never automate away because the target audience of computer games are real people, not chatbots. This is why I also believe the whole idea of automating away the entire creative process won't work, at the end of the line, there is always a real human.

Also, the irony of using a proprietary tool to generate ideas for a game for a free software engine is not lost on me. It’s actually very sad and such an admission of defeat.

[Hume2]

Yes, and this makes it difficult to tune a game both for offline and online. In the online case, to some extent veteran players can help newcomers. Some players do it naturally, not just their friends, but if the game adds a benefit for the helper, it can be a win-win situation. Some activities can become tedious in the long run, like hunting some animal for their hide, but for new players it is new and maybe exciting.

Still, when one stumbles upon a dungeon or a cave and there are signs that someone was there before, it is a little discouraging. I believe the game "balance" is in the opposition of a tendency for players to spread because of resource availability, and the tendency to stay together for safety and trade.

A possible way out is with exploration. I joked earlier about nomads, but actually new players don't have a home, don't have farm plots, etc. so their best option is to roam around (the problem, though, is that they'll have to be patient if they actually just want to play Farmville). A nice place could also be a piece of information they can sell.

A game could create or improve that kind of beginner's advantage. I could do that in my game for instance, as leveling up is the players' decision (provided they have the resources for it). There's even an example of this already: they lose the ability to summon NPC traders to gain one ability of their choice when leveling up for the first time.

Indeed. Maybe the problem actually lies in that we tend to treat singleplayer and multiplayer equally.

In an open world and sandbox game, the game should, rather than having a main goal, have "affordances" of goals - or perhaps more simply put - suggested goals. I often use Bartle taxonomy as a starting point when thinking about the game. In that case, a game should provide at least one goal for each type of player. Preferably the goal should actually not be reachable - or "asymptotically reachable" (the further you progress, the harder it is to progress) and maybe also not "monotonic" (one can fall down and regress - dying is a typical case).

This can look like a scam, though :-). But unless you plan on building a game studio and create an endless stream of extensions like WoW, the game has to be open-ended, that is endless.

Yes, an open-world game can have more different goals and it is a good feature IMO. I just think that there should be at least one greater goal, which is given by the game itself, and the path to it isn't straightforward. Also, this goal should somehow utilise a large portion of the game mechanics in a meaningful way. It shouldn't need the player to master them all, but these mechanics should serve at least as considerable options.

is this gonna be another chatgpt dump similar to what eredberre did or whatever his username was

Don't worry, if I dumped here whole the conversation with ChatGPT, it would be way longer. I just found some of its ideas interesting and I didn't see them mentioned in different threads. Most of its ideas weren't really usable, so I didn't post them here. If you don't want to inspire, simply ignore it, all is fine.