Print

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

[Asgarus]

There could be "ticks" in each turn. Each tick is an action event, or you just take events instead of ticks^^. Which action gets executed first, depends on the skill and swiftness and whatever of the fighters. An action would be the direct reaction on the action from the last tick, so you can still raise your shield against an incoming blow, or counter it with your weapon, duck away, whatever)

To determine the order in case of several actions from one fighter, we could do something like thi:
(this could be the advantage, calculated from different factors, character stats and skills, terrain, and so on)
A has an advantage of 90
X has an advantage of 85
B has an advantage of 60
C has an advantage of 40

Turn 1:
tick 1: A makes a powerful, high, horizontal swing (-90, since he used everything on one attack, 0 left)
tick 2: X raises his shield, blocking A's swing (-28, since it was one of 3 attacks. 57 left.)
tick 3: (B's value is greater than X's now) B makes a powerful, high slash (-60, 0 left))
tick 4: (X's value is higher than C's) X ducks down, avoiding B's attack (-29, 28 left)
tick 5: (C's value is higher than X's) C makes his downward swing (-40, single action)
tick 6: X makes a low swing towards B and gets hit on the helmet by C's swing (-28, 0 left)
tick 7: B gets hit by X's low swing

the advantage could be influenced positively and negatively by wounds and buffs and stuff, so that the second turn would start with different values.

[mastahcheese]

Unless yo combined the advantage of your system with the advantage in the ideas before, in which case pushing an opponent back and gaining ground over them would actually net you more actions.
Which would certainly be an interesting combat system to play with, particularly when facing multiple opponents, as it would make fighting in unfair fights really dangerous.

[Clownmite]

Thanks for the continued feedback everyone! For the first iteration, I think I'm going to go with something like the following:

• On a turn, a combatant chooses 2 moves, an "opening" and a "closing"
• The "opening" moves of the 2 combatants are directly compared, and the winner is chosen in a weighted-random method based on the sums of both combatants openers/bonuses
• The winner of the opening move gets a chance to strike at the opponent with their closing move and deal damage; although the opponent has a chance to (passively) dodge, block or parry the attack
• Combatants who attack an opponent who doesn't attack them back get 2 chances to strike the opponent and deal damage

I'll have to see how everything feels once the system's in place, and it will likely change before long. The goal is to add a bit of player thought/skill that can go in to fights; however, for the most part, the player should be able to let the AI choose moves for them without much worry. So ideally, you can pretty much "bump" into regular bandits without too much worry, but you may have to choose your moves a little more carefully when fighting the ringleader. And of course, the UI will need to be super clear and intuitive, making choosing moves as simple and quick as possible.

[mastahcheese]

Sounds nice to me.

[Clownmite]

I realized today was a weird day when I generated a world, ran into the wilderness, took off my pants, and started attacking them. The crash I was getting when attacking non-creatures is gone though, so that's a success.

The simultaneous turns are working out OK so far. There are 7 types of attack to choose from, but you can't pick any of the attacks you chose last round so you can't just stick with a winning combo the whole fight. Attacks will only land at logical areas, so a low swing can only hit a leg and a high swing can only hit the head. There will need to be a little bit of predictability so that you can try to gauge your opponent and counter their moves, but that's not in place yet. I picture most people not caring enough to carefully plan each round, so you can still just bump into an opponent and the game will pick some moves for you.

[mastahcheese]

Sounds good so far!

Except for the pants, of course.

[Clownmite]

Still working on this, slowly but surely, whenever I get the chance. I intentionally left combat off in a mostly incomplete state, giving me a lot of time to think about how to improve it.

Since the last post, I've implemented the first part of battles which the player isn't around to witness. It all ties into the same combat code, so there's not much abstraction going on (although there probably will be when fights get to army vs army scale). This has tied in with a better system to track which factions are openly hostile to each other.

Some historical events are now tracked and recorded, and knowledge is able to be spread between creatures. The knowledge can be about particular events, the existence / location of sites, and the location / existence of other people. This part is very interesting to me because I'd like to make some of the focus in the game on recovering lost artifacts by discovering their locations. Some of this information can be discovered by reading books or maps.

Tying into that, sentient creatures now speak one or more languages, and may also be able to read / write in those languages. I hope to have this lead to situations where you uncover a book or note that is written in an ancient script that you can't read, and must find someone who can translate it to convey the information to you.

Here's a couple of screenshots, mostly showing some of the new events and some (placeholder) conversations you can have with people about them.

Spoiler (click to show/hide)

[Graknorke]

Do you have any kind of devlog where I can look up how you've done what? The bit that interests me most is the economic system and the kind of ideas you've used to make it work, but going through the whole thread trying to find what's relevant is a daunting task.

[Clownmite]

Do you have any kind of devlog where I can look up how you've done what? The bit that interests me most is the economic system and the kind of ideas you've used to make it work, but going through the whole thread trying to find what's relevant is a daunting task.

This thread is the closest thing I have to a devlog, unfortunately. Do you have any questions in particular?

[SolidSlater]

This looks amazing, how long do you think until it will be playable?

[Clownmite]

This looks amazing, how long do you think until it will be playable?

Thanks! It's really tough to say. I thought I would have had one already, but some of the old code was a complete mess and I've cut a lot of it out recently. It's a hobby project so I can't dedicate myself to it full-time.

There's definitely some cool elements in place, but there's not really anything to do yet. There aren't yet any shops you can buy items from, or reasons / items that would drive one towards anything resembling an adventure. Think of it like URR or Cult/Empyrea - you can walk around and see the world, but there's not anything yet driving gameplay.

Of course I hope to drive towards something playable soon, but I can't give a much more concrete estimate than that.


Some notable recent developments:

- I've fine-tuned the language generator just a little bit to fix a few minor issues, but the result is even better languages / words being generated.
- I started implementing an AI system (goal-oriented action planning) that should allow for some cool stuff, pending a ton of future tweaking and debugging 
- I made some tweaks to the economic simulation, which looks like it's allowing economies to be stable indefinitely.

[lemmily]

I'd just like to add that it's always awesome when i come back to the forums and see some new posts on this thread - I always check!

[Graknorke]

This thread is the closest thing I have to a devlog, unfortunately. Do you have any questions in particular?

Well, how you've abstracted it out in general. I've gathered that you've got a number of agents each of whom have some things, need some things and produce some things. I'm curious in particular with how they decide what to buy/sell/make, though. It's the stumbling block I consistently hit when trying to simulate an economy, which is unfortunate because it's the most important part.

[Clownmite]

I'd just like to add that it's always awesome when i come back to the forums and see some new posts on this thread - I always check!

Thanks!

Well, how you've abstracted it out in general. I've gathered that you've got a number of agents each of whom have some things, need some things and produce some things. I'm curious in particular with how they decide what to buy/sell/make, though. It's the stumbling block I consistently hit when trying to simulate an economy, which is unfortunate because it's the most important part.

I used this paper as the basis for the economy stuff: https://larc.unt.edu/techreports/LARC-2010-03.pdf. In short:

Each agent can only produce one type of thing - some gather raw resources, and others convert that to finished goods.
Each agent needs certain things to do their jobs.
Each thing in an agent's inventory has a certain chance of breaking each turn.
Food must be consumed every turn.

So:
Agents will always produce as much as they can each turn - but if they don't have the tools they need to perform their job, they are much less effective.
Agents will always attempt buy anything which they need to do their jobs which is broken.
Agents will try to buy food if they don't have enough in their inventory to feed them for more than 2 turns.
If they have a certain amount of money and there are no items of the type in their inventory, they will try to buy "luxury" items that aren't strictly necessary for them to perform their jobs.


Each agent has their own belief about the price of items, and they will post their buy / sell offers based on this belief. They'll adjust their beliefs based on whether their offers were successful or not. When a lot of agents need an item, there will be more overall bids, meaning the high bidders will get the items and the other agents realize they need to bid higher next round - this simulates high demand. The same applies when a lot of agents try to sell an item - those who post low sell offers will make the sale, so the others will adjust their price downwards so that they can compete next round.


Just for fun, here's how the economies of 4 cities in the world I just generated look, after about 15 years: http://imgur.com/a/Xrt0i

Spoiler (click to show/hide)

[Graknorke]

-economy snip-

Ah, thanks! The link and your summary were both helpful for me setting out how to create a similar system.
In your simulation, what do you do about the number of each type of agent? Is the number of them constant for each, do they switch roles but the overall number of agents stay the same, or does the actual number of agents change? The first would of course be easier, but in both the report linked and your own data the supply of commodities changes over time.