next »

[Reilly]

Thanks for the suggestion, we're still pondering everything.
New map, DoubleYou

[Reilly]

So I have to clean up my tileset and I was going to take this as an opportunity to add some more terrain tile types.

Right now we have:
Plains - Very little defense. Very plain.   
Forest - Good defense, difficult for mech units to travel over, and provides cover in fog of war.
Mountain - Can only be traveled by infantry and air, great defense, infantry gain vision in fog of war.
Water - Can only be traveled by air units
Road/Bridge - Similar to plains but zero defense.
Dirt/Mud - Units take extra damage on this terrain, difficult for mech units to travel on.

What I'd like to add:
Wall - Some sort of wall or object that can't be crossed by any unit.
Cloud - Acts as a forest but for air units, can only be placed over water... so maybe a Squall?

If anyone feels like sharing ideas/thoughts I'd be super grateful.

[danieru]

To be honest the wall tiles (pipelines) in Advance Wars felt like a cop-out for map designers. It gave the designer a strong method for controlling the flow but robbed the player of possibility space.

Instead another option is to use rivers and allow the player to construct bridges. This goes in the opposite direction by stealing some of the designer's control and expanding the player's options. The designer can still guide the flow and slow a mechanized advance but timing is in the player's hands. This is what I've been planning to do in our game. It does require a somewhat flexible code-side representation for terrain so it might not be possible with your existing codebase.

But if you do have the flexibility it should open options for campaigns, like your villains destroying cities 

[Reilly]

@ danieru - Hey, just checked out your project, lookin' good!
I get what you mean about limiting options, but our game only has land and air, so I really think there needs to be a way to wrangle in the air units. Obviously they would need to be used sparingly and in clever ways.
We have discussed a bridge building mechanic in the past, so we may revisit that down the road.
Thanks!

[danieru]

Thanks Reilly! To be honest I'm downright impressed by Warbits, it has incredible polish and I hope it will bring more fans to the genre.

That is an interesting problem and I guess walls are a fine solution. I have to apologize for not knowing this, but does Warbits have a weather mechanic? If so would a weather type which increases air units' fuel consumption work? If there was a land based supply unit this could restrict air units on a leash around the terrain restricted land units. Yet still give room for inventive players to subvert the map designer.

A squall does sound like an interesting tile. On a similar theme what about a tile which eliminates vision during fog of war? Perhaps an EM field, building on the robotic theme. This would create a trade off between fast movement and safety. The player can decide between racing across a field without seeing the enemy while the enemy can see and fire upon them, or take the safer long route.

[Photon]

I'd like to say I really like the idea of having cloud terrain, and having it only over water makes sense too.

Perhaps the clouds could have a 2 or 3 space movement cost for air units too? That's at least something. But danieru did mention weather; didn't you say earlier that there will be some sort of weather system? Perhaps rain/storm weather could make the clouds harder to traverse, or snow could cause air units to consume more fuel in clouds. Or maybe that's all just getting too complicated? HMM.

But as for the bridge building, I'm not so sure about that. I feel like being able to alter the terrain and set-up of a map (chokepoints, defensive positions, etc.) would be a little overpowered. I suppose it would depend on the climate of the game in the first place. I feel you have to consider how it changes player tactics and how strongly it stands to swing momentum. Again, depends on the desired game climate.

[Chicha]

Clouds and walls are both very good idea, I like them!  Clouds with additional movement costs (mentioned by photon) may not be a good idea though. I think that would make things too complicated and kind of ruin the nature of air units.

As a map designer, I would actually like to see more terrain with homothetic purpose. For instance, shoals in advance wars are passable for ground units, in a case there is no naval unit, shoals are very similar to roads. Even though their properties are similar, adding shoal as an element of design offers more visual variety to the maps, which make maps look more fun to play (in other words, not that bland.) Here are some tiles that i think may be good to add in:

Shoal - Already mentioned above.
River - Similar to mountains, but with lower defense. (can be passed by probes?)
Hill - Similar to plains, units on it gain +1 vision (and +1 attack range for indirect units, if you want more.)
Ruin - Provides cover in fog but with lower defense than forests.
Swamp - Hard to pass for infantry, while same as plains for mech units.

Hope that would help. Keep up the good work!

[rek]

This talk of clouds makes me think of a mobile fog-of-war-like cloud, one that slowly drifts across the map and temporarily blinds (and immobilizes?) any unit encompassed.

[Reilly]

Some great suggestions, thanks guys!

Made some cloud tiles the other day.

I'm thinking they'll hide air units in fog of war, give a defense buff, but cause air units to consume more fuel.

[Photon]

Why so surprised, Mr. Air Unit?

I like it. I feel like its a nice aesthetic addition as well.

To add a little bit more to what I was saying earlier about clouds, walls and bridge building, I think it may boil down to the balance you want between tweaking the formula and building around the formula. By "building around," I mean taking the variables of the formula (terrain defense and move costs, fuel consumption, etc.) and building new elements using those; by "tweaking," I mean adding new variables (such as terrain alteration with the bridge building) and essentially expanding or modifying the core of the formula.

For me, my impression was that Warbits wanted to capitalize on the tried and true formula of past strategy games and work primarily around that (more building around than tweaking,) so that's why bridge building sounds a bit too "ambitious," if you will. I could be wrong though.

[Bombini]

Looks great!

[Reilly]

We found a bug awhile back that lets carriers resupply units if there is an APC inside. It's fun to play with so it's staying in.
I made a little visual mod for the carrier when it has an APC.

I went the R2-D2 route.

[Reilly]

Here's a video of the last few turns of a CPU game to show the current state of the AI. Video is kinda washed out for some reason

Let us know what you think!

[danieru]

Nice job on the AI! It looks fun to play against. The only thing I noticed was the last air lift dropped an AA further south than I would as a player. The AI still won in the same round so it might have been a randomization choice?

How did you approach the AI? I've seen some games which use a single pass through tactics -> long-term movements -> buying, performing the best decision at each stage for the current round. Then there is the chess engine like approach, but I'm not sure how practical the latter is. I'm curious how you approached the problem since its clear you have a high quality AI. That it finished on 31st day is pretty awesome, I'd expect a human to play just a bit faster.

[Vanhail]

Nice job on the AI! It looks fun to play against. The only thing I noticed was the last air lift dropped an AA further south than I would as a player. The AI still won in the same round so it might have been a randomization choice?

The transport AI is brand new so it often has tunnel vision, I need to work on making it more flexible. Right now a full Carrier looks for an end tile where both units can attack from, regardless of gas consumption or the length of the path. 

[Reilly]

@danieru, thanks! Joe's been slaving away and it really shows.

I'm currently plugging away at an old school instruction booklet.

WarbitsManual.pdf (2Mb)

Right now I'm trying to get all the basic mechanics in. I'll be adding unit, terrain, and structure info soon.

[Vanhail]

How did you approach the AI? I've seen some games which use a single pass through tactics -> long-term movements -> buying, performing the best decision at each stage for the current round. Then there is the chess engine like approach, but I'm not sure how practical the latter is. I'm curious how you approached the problem since its clear you have a high quality AI. That it finished on 31st day is pretty awesome, I'd expect a human to play just a bit faster.

How to approach the AI has been a constant internal debate for me. This is my first experience trying to develop AI so it's been a challenge, a mix of frustrating and rewarding development moments. It's interesting because it's easy enough to think through the data and logic required to gather a general sense of intelligence, but how to organize and execute it in practice has been the real chore.

Knowing how smart to make it in some areas has been important. It's tempting to go overboard calculating every possible scenario, but sometimes a heuristic approach works well enough.

In general I'd say I'm taking the GOAP approach (goal oriented action planning). I have different modules that handle various phases of decisions in each turn. All goals are weighed and the highest rated goal executed. Every time a goal is executed the module calculates the best goal again, since it can often change as a result of the precious execution. Ex: A units health is brought down to a killable threshold for another unit.

I'm happy to talk more if you have any other questions.

[Belimoth]

Do you have a fuel system? Could two apc-carriers fly forever?

[Reilly]

Yes, yes they could. But making 2 APC Carriers would eat up 30,000 of funds, so I think it evens out.

[Vanhail]

Did we just solve the energy crisis!?