next »

[Pineapple]

you might not want to make it too unpredictable, though. as a player, I expect to, given some forethought, be able to control the outcome of something that is entirely due to my own efforts with a significant degree of accuracy. when I tell my platformer character to jump, if I can't predict where I'm going to land then I'm going to fail quickly and become frustrated very quickly. just make sure there isn't so much to consider that predictability becomes impossible, is all.

[sublinimal]

I absolutely agree, it should always be possible to see a mistake coming or explain it afterwards. Roguelikes sometimes rely on randomness in the wrong places, especially when it comes to risk and reward - you may or may not find the almighty +5 sword of manslaughter lying on a floor, and when you do, it's like "What'd I do to find this? Why was I rewarded with this?". Of course the point is often to adapt to the cards you're given, but I'm not going to go as far as to obscure cause and effect. My philosophy would be to randomly spawn a boss enemy that drops such a valuable item, so you'd have the choice of fighting it for a reward or avoiding it for less risk.

But there's no swords in this game. Some degree of randomness might be necessary for plant mutation and evolution. Still, there should always be a model for what you're doing, and a way to influence the results.

So let's try factoring in age a bit more. A plant's age has to be within certain limits for it to count neighbours and be counted as a neighbour.



Here, rapidly aging plants only have one shot at breeding during their lifetime. No more glider guns, everything seems more organic. This one's probably my favorite from a couple of experiments. It comes to a halt pretty fast, but that's only good gameplay-wise: farms are supposed to require a lot of care from the player rather than being just something to stare at.

That is, at least until they figure out some infinite growth pattern. But even then, you'd still have to worry about selective breeding and protection from invasions. And it's not like you'll have farms this big any time soon.

[Pineapple]

cool then, I think that looks like a good way to go about it 

[sublinimal]

Code:

def wall_of_text(i,y,t):
    def u():m=len(t)%2;p=(h-len(t)/2+1-m)*'-'+'= ';return n+m*'-'+p+t+p[::-1]
    h=a/2-3;n='\n'*2;o=u()+n
    for r in range(y,y+s+1):o+=(a-w)/2*' '+i[r]+'\n'
    t=str(y+1)+' of '+str(len(i)-s);o+=u();print o

I can get carried away with code golfing sometimes.

[kamac]

Looks cool!

I fancy that farm placing 

[Pineapple]

I can get carried away with code golfing sometimes.

you should see what my code looks like when I don't expect to be looking at it again.

[sublinimal]

I've talked about keeping things predictable, and it usually involves displaying stats. The roguelike format means there's simply no way to have everything on the screen at once, so I'll have to come up with a couple of creative solutions.

First things first, can it be embedded in gameplay? I'm thinking of a "biology" skill that would let you see more information in more convenient ways as it levels up. In the start of the game, you'd have to do everything by hand: walk around a lot to view what's planted and where, memorize some numbers and do simple arithmetic in your head. Later on, you'd have several "layers" you could scroll through with the keys 1-9, quickly displaying the climate and plants' attributes, automatically pointing out maximum values, predicting what's going to be born and where. Seeds and plants in your inventory could also be sorted better by a certain attribute.

[Pineapple]

that sounds like a good idea, but I imagine the game would earn a certain charm if instead of being a skill, this information could be viewed in books or somesuch that the player could view and perhaps buy, and would be encouraged to write the information down on paper for reference. (I'm thinking of the library in what harvest moon I've played)

[eclectocrat]

Code:

def wall_of_text(i,y,t):
    def u():m=len(t)%2;p=(h-len(t)/2+1-m)*'-'+'= ';return n+m*'-'+p+t+p[::-1]
    h=a/2-3;n='\n'*2;o=u()+n
    for r in range(y,y+s+1):o+=(a-w)/2*' '+i[r]+'\n'
    t=str(y+1)+' of '+str(len(i)-s);o+=u();print o

I can get carried away with code golfing sometimes.

OMG! That's gonna hurt in a few weeks...

[sublinimal]

that sounds like a good idea, but I imagine the game would earn a certain charm if instead of being a skill, this information could be viewed in books or somesuch that the player could view and perhaps buy, and would be encouraged to write the information down on paper for reference. (I'm thinking of the library in what harvest moon I've played)

I was thinking potion recipes could be acquired that way. Since they actually unlock new game mechanics, it makes sense for them to be tied to your progression. Meanwhile biology would simply speed up something you already can do, and it would be trained by farming, so there's a justification for it being a skill.

Player skills should be a matter of specialization. It fits the nature of the game, high freedom but also high responsibility. No classes that determine your playing style right from the start, just skills you'd choose to train during the game. Farming increases Biology, trading increases Haggling, crafting increases Alchemy, attacking increases Dexterity (movement speed/potion throwing range), being attacked increases Stamina (the more the closer you come to death)...

[Pineapple]

can I recommend against the getting near death makes you less vulnerable to it? I feel like a fool every time I play a game with that mechanic and realize that I am subjecting myself to injury on purpose. maybe it could be tied to not getting damaged? I dunno. I am extremely enthused by all the other ideas, though.

[sublinimal]

I'm not convinced it's a bad idea. The difficulty level will be high enough that you won't need to deliberately hurt yourself to have NDEs. It's simply supposed to offer more reward for surviving tricky situations - "what doesn't kill you, makes you stronger" and all that.

But we'll see for ourselves how well it works when I start working on combat. Which is soon, considering my todo list for the first step is getting short. Just some uninteresting tinkering going on (more object-orientation and some details in the mechanics).

Not much actual content though, nothing worth releasing yet.

[sublinimal]

In fact, I think I'll just start working on combat right now. I need to take a break from optimization and do some creative stuff for a change.

I'd like different types of enemies to have distinguishable behavior. There'd always be a certain pattern to defeat them, and you shouldn't have to grind through cannon fodder. RPGs sometimes feel like you're fighting the same enemy over and over, just with different HP and attack power. While I'm no knight of AI myself, I'm at least going to include tactical elements & rock-paper-scissors mechanics to keep things interesting. I've already got a couple of enemy ideas:

Trickster - Small, fast, weak and harmless, but comes in packs. They'll try to swarm you, block your throws and steal your potions, so they're often your primary target. Especially if you're armed with potions that deal splash damage.
Assassin - Lurks in the shadows and throws poisonous potions from afar. You'll want to check the nearby hallways for these before stepping into open areas.
Zombie - Carries a couple of damage potions stolen from its victims. It can both drink and throw them, since damage potions heal it and healing potions damage it.
Gnome - Intelligent little creatures that can craft powerful potions to protect their underground farms. However, due to their small size, they only have the strength to throw potions after a couple of turns. Which means they'll attack rarely but deal heavy damage if you don't get rid of them during the countdown.
Dweller - The "anti-gnome", both because of their fight for territory and their contrasting attack patterns. A primitive but nimble creature, carries a heap of crudely made potions but throws them rapidly. The grudge between gnomes and dwellers will be a part of the plot at one point.

Running from enemies is a viable tactic (as long as you can outrun them), but you're going to miss their loot when you run out of your limited potions. Some of them even hold keys to unlock nearby doors, behind which you'll find their personal farms, often with brand new types of plants. The final screen of each cave will have a tougher monster guarding the exit.

Cave generation needs to fit gameplay, not the other way around. Since fighting revolves around ranged attacks (=potion throwing), the common approach of rooms connected with narrow hallways wouldn't be very appropriate. I want to favor open and cavernous spaces.



It's pretty amazing how usable and "organic" caves a cheap random walk algorithm can generate (with slight trimming):



There's some nice variation between open and narrow spaces, small walls to hide behind for cover, and connections to surrounding screens. Fun to explore too:



Nothing's terribly wrong with that kind of raycasting, but a large field of vision makes everything look that much more dramatic, and probably helps reduce bullshit deaths because you'll really want to see the enemies before they can attack you.

[sublinimal]

This thread's view count jumped by 50% while I was asleep. One more lighting style: dynamic to the point of being chaotic, but keep in mind this animation doesn't describe the pace of the actual gameplay.



One possibility is that's what you'd see, and the places you've been to would be recorded in a separate map. Wouldn't be as convenient but all the more immersive.



There is horizontal space left as well, but it should be reserved for messages.

Nonetheless, I'm going to work on enemies now. Gameplay shall be the judge of these visual styles.

[sublinimal]

A pack of tricksters. They can be quite persistent. Probably should be faster than the player to make them a more immediate threat.



I hope that gave nobody seizures. Basically, they wander around until they can see the player, then start approaching him. They refresh their path while they keep seeing him, but if they lose track, they'll carry on until the end of the path before switching back to wandering. I don't even need A* for pathfinding here, I can just use the raycasted line that hit the player. I've noticed some inaccuracies in the raycasting, though.

[Pineapple]

that's looking pretty cool. do you think there will be any enemies that are more tactics or squad based in behavior? that could be interesting

[sublinimal]

Not sure how it'd translate into roguelike combat. Said tricksters are probably the closest I'll come to that. They should have some simple method of "communication" within their packs, such as each member of a pack agreeing to target a different tile neighbouring the player. Maybe the first to see you would announce your location to the others so they'd go for a coordinated rush. Anything they can to trap you more efficiently.

[sublinimal]

Tricksters are becoming interesting. I'm now leaning towards stealth elements with them. You can see them before they see you, and if you're within their range but there's an obstacle in the way, you get a message to the effect of "you hear excited little footsteps". Being slower than them, if one sees you, you'll eventually be caught if you can't hide quickly. Then it proceeds to steal some of your potions and run away with them.

Tricksters die easily, but there are only downsides to fighting them, so they're best left alone. While they do drop your stolen items, they go down with a screech that alerts the rest of the pack to your location. And those damage potions are more needed against other enemies.

You could think they're not dangerous since they can't even damage you, but since potions are scarce, they might leave you defenseless against monsters that can. They definitely increase the "oh shit" factor and I like that.

[sublinimal]

Had to try what caves would look like rendered Wolf3D style.



I'm sticking with ASCII, because it conveys more information in less space and leaves more room for imagination. But I'll keep an eye on this aesthetic for some other projects.

[sublinimal]

Got a couple of enemies in, ironing out their AI at the moment. It's kind of starting to resemble a game.







The cave system is almost like a separate project from the town files, so I'm going to have to spend some time merging files and generalizing functions. Organization is really one of my weaker points.

The empty area of the screen obviously lacks stats. And rather than a separate message window, there should also be a box for messages you'd scroll through with +/-, since they're such an important part of what's going on.