Empires of the Undergrowth: A game about ants

[JeffNevington]

I have been working on an ant colony management game for a while now and posted around here a few times in relation to that.  We just put out a new demo (with some much overdue performance enhancements) and so it seems like a good time to do an announcement about it in the indie alley.  You can grab the demo (0.064) for Windows here: http://www.indiedb.com/games/empires-of-the-undergrowth/downloads

It's a bit like the original dungeon keeper, crossed with sim ant. I stress "a bit like", I don't know if we can truly do either of those immortal titles justice. 

You start each level with a queen and a couple of worker ants, and you have to build up from there.  You need to gather food, build more workers and soldiers, and most importantly, defend your queen.  Currently in the demo, you can only explore underground maps - but you can see overground in action on our steam "coming soon" page. http://store.steampowered.com/app/463530

This is how the underground looks (after a lot of digging)


And here is a shot of the overground (the beach)


Our team is really small, there are just three of us, and only one gets to work on the project full-time (slowly eating away at our modest kickstarter pile). We have to try and do as much as we can on a tiny budget. That includes making our own sound effects:

If you have any trouble with the demo or have any feedback or anything else, let me know 

[Jubal]

Ah yes, I was doing that sound-generation method for my own game recently!

I'll check it out, been looking forward to test-driving this one for a while

[JeffNevington]

That's great. You seem to give everything a go

One day it would be nice to have mod support too. A few people have already expressed an interest in making their own scenarios.

[Jubal]

OK, I've had a few runs through the three available maps!

Main reaction - kudos, it's a lovely game. It looks pretty - not just "looks the part", but the whole thing is really quite visually pleasing, the sounds are good, the controls are smooth and make a lot of sense. I sense there's a lot more content to come, but what's there is already an actual fun game and I've enjoyed playing through it. The fact it's quite scenario-driven thus far perhaps limits replay value a bit, but the scenarios are good fun; the "critters" and "the nest" took me 3-4 attempts each before getting the knack of them. Surrounded I wanted to play more of to see more other enemy bugs but I also enjoyed it less, I think simply because I don't like the "fight waves until you die" RTS scenario style so much.

Detailed thoughts/notes:
- When I first booted it up, the graphics settings were weird and so fuzzy I could barely see the camera until I tweaked them (I should've got a screenshot but didn't, sorry)
- It definitely makes my fan boost up quite a bit faster/laptop heat up a bit, don't know if this is normal
- I definitely wanted more of an info base - like mousing over enemies to see what they were or something (both in terms of their species and in terms of getting some sort of tactical sense of how tough they were). The research is clearly good behind the game, and it would be good to see more of that coming through.
- Similarly I needed a bit more clarity on my own side - I guess some of this may be covered by the tutorial, but even so it might be good to flag up "worker ants are workers, black ants are soldiers" more clearly within the UI.
- UI stuff in the top right corner might be a bit small, I wasn't using it much & found I preferred to just look at where the scent markers were
- I felt like I wanted more tactical options at times, though I imagine this is something that will be fixed as more content comes in/different maps - the scenarios as they are work well for the game as it is, but Critters and The Nest at least are fairly pure rush-tactics styley, especially the second half of The Nest (though that bit is good fun!)
- The scent-trail mechanic works very nicely, I really like that and the way it pushes towards a non-micromanaged playstyle
- For Critters and The Nest, I found there was a strong weighting towards learning a map with a couple playthroughs before winning it - because it's not always clear when you're going to hit a chamber with critters in it, learning roughly where those are gives one a huge advantage

I'd very definitely be interested in scenario building if the option was available

[JeffNevington]

Awesome feedback, thanks!

What you say about it being "scenario driven" hits a chord with me.  We have been aware of this for a while and although I was working on a full blown bot colony AI, it has taken a back seat for the moment as we ramp up to early access (which will be primarily scenario driven).  The fact is that it is much easier for us to make engaging content in this way than create a balanced bot AI that changes tactics.  The plan is to release 4 new scenario style missions for early access (with a bit more of a story, including narration), and a "pet formicarium" mode, which is a bit more free and the danger is controlled by the player.  After that, I want to pour effort back into the bot AI and developing a nice skirmish mode.

I have noticed all sorts of things happening with graphics on first boot of the game.  We need to look into that prior to release. As for your fan - it is a pretty heavy game. We have been working on getting the CPU load down: the main culprits here are movement, collision and individual ant behaviour AI.  The GPU load is also high; this is a symptom of having free reign of UE4 graphics.  There are still optimizations to be made there, but the main thing I think we want to do is make sure that lower graphics settings still look nice (we haven't paid much attention to anything below maximum so far).

Having more info on the species is really important - as they are real species. Hovering over for their name seems sensible - also there is health bar info if you hold down "alt", perhaps that could include some stats too. We want a creature wiki in the game anyway to give all of this information but that's not so helpful in the heat of battle.

Clarity on how to use your ants is understandably difficult.  When I watch people play on Youtube, I find it incredibly frustrating "NO NOT LIKE THAT, YOU ARE DOING IT WRONG".  Totally our fault as we have no tutorial yet. Distribution of ants in your groups has a massive effect on colony efficiency, and arguably the most important group (the "home" group) is unlabeled.

More tactical options should emerge as we include more ant species.  The next one planned is a type of wood ant, which sprays acid as a defense mechanism (so.. ranged ants  ). The introduction of tile-based upgrades should diversify your army more and effect who you want to send in first.

Again - with the scenario based missions, learning the level is a significant part of it.  Actually I don't like chamber shapes being unknown - I don't think we will be having long elaborate cave-ins in the next missions (unless they don't contain enemies).  You should always know if you are about to dig into danger (at least to a large extent).

Really appreciate you taking the time to give such a thorough breakdown of your experience!

[Jubal]

No probs, I'll probably have more thoughts when I've played a bit more

Additional scenario content sounds good and I definitely think a "campaign mode" would be a decent addition.

Query that's probably quite basic: I've been instinctively putting hatching areas near the queen under the assumption that the workers are supposed to take eggs from the queen to the tiles - is this how it works? Conversely, it doesn't feel like there's any reason to put food tiles near the queen, in "The Nest" I often build small blocks of new ones as I go forwards to make the food-gathering chains shorter and more efficient, don't know if this is how it's intended to work...

[JeffNevington]

Food tiles have actually begun to annoy us as they don't really seem to serve much purpose game-wise. If I was to start again, I would just have the ants bring food to the queen, but they have sort of become woven in and will probably be more trouble than it's worth to remove them.  - You are of course using them as any good strategy player should (putting them near by food as you move further from the queen).  I don't really like this and we are coming up with some quick ways to combat it in future.  Our first plan is to have food tiles be very expensive, but with reduced cost for each neighbouring food tile (so investing in a large food chamber is more sensible than small satellite ones).  This should at least reduce the number of random food tiles that players put wherever they please.  Future plans include having creatures that steal food, and the ability to create food networks (jobs for ants to move food from one storage area to another).

For hatching areas - the workers (primarily those in the far left "home" group) will take eggs from the queen to the tiles, which then hatch at the cost of some food and are then replaced, again by the workers.  Short distance from the queen helps with this, and you start to feel it when you have big nests and are replenishing 30-40 ants after a big battle.  Another thing to consider will be tile upgrades.  Although not available at the moment, if you hold down AltGr, you can see the level of the tiles (graphical representation of level is a WIP).  A tile's level is increased by surrounding tiles and walls; once it reaches a certain level, then you will be allowed to upgrade it (which then in turn increases the level of all those it touches).  This will promote large rooms, enclosed by walls in the final game.  Another thing to be introduced that will effect nest architecture will be that ants will move more slowly over hatchery / food tiles.  This will encourage the player to create corridors with satellite rooms, so that when the ant paths to a location, they avoid built tiles and hence move more quickly.  So even though digging out a massive area will allow ants to travel in straight lines, they wont be as efficient as they would if you had created a network of tunnels.  Hopefully this will organically lead to people building nests that feel more realistic.

[Jubal]

Hmm, I think just having the neighbouring tiles rule would lead to players making forward-base food zones - I think if you want food to be returned to central areas it may need some mechanism whereby it actually has to be eg carried to the queen in order to spawn eggs or something. Food stealing enemies sounds like a good plan though, and I like the sound of the tunnel requirements.

Does being in the home group make ants much more likely to do egg stuff? I always ended up making a group then tasking it around the queen, which may not have been efficient...

[JeffNevington]

The home group is a strange thing that is left over from before you could actually control the ants with pheromone markers. In the first demo of the game, all you could do was build and the ants were entirely autonomous.  There were no groups and all workers would do jobs including building, digging, carrying eggs, and collecting food automatically; soldiers could only fight and would go toward the nearest enemy, wherever it was.

We then introduced 1 group marker just for soldiers as an experiment (and decreased their aggro range from encompassing the whole map, to just a few tiles), people really appreciated this control.  After that, multiple groups appeared and we muddied the waters with regard to who can do what: we let workers fight, and we let soldiers collect food.

As soldiers could collect food, it started to feel like food shouldn't be universally visible to the ants, and so in the same way as we did for combat, we reduced food detection range to just a few tiles - you must now tell ants to go near food with the marker or they wont collect it.

The three other jobs that only worker ants can do: digging, building and egg carrying are still universally visible (wherever the job is, they will do it).  This is controlled by a 'job manager' that calculates the most efficient distribution of jobs within the nest and then assigns them to workers accordingly; updating periodically and swapping jobs if necessary (see the travelling salesman problem). The job manager also checks if the worker is in a group, and has its pheromone marker placed down: if the marker is not near the job, the job manager won't assign it.  This means if all of your workers are in a group, and that group marker is set in location A, and you dig in location B (far away), no one will come to dig.

If you put your workers in the nest group (or any group with its pheromone marker currently disabled) - the job manager will automatically tell them to take eggs to tiles in an efficient manner.

If you want your workers to prioritise taking eggs to a specific room, you can actually use the pheromone marker to send them to that room; the job manager will then only give them nearby jobs, and they will fill up that room first.  Though, this would probably be a very specific scenario emerging possibly in a future build (say you had a small room of highly upgraded tiles, or specialist ants, and you wanted those replenished before any others).

The nice thing about this whole system is that as long as you tell the player to "keep some workers in the home group", even if they don't really understand the mechanisms behind everything, they can play the game fairly well and it feels quite intuitive.  However, when someone wants to really figure out how to do something efficiently, they need all of this explained properly - and the above is going to be difficult to get across in a short tutorial.  We will try though.

[Jubal]

I think the main thing you need to help with that from an intuitive perspective is a marker in the UI to show that the home group is a specific home/nest group, rather than just "the bit before you assign workers". So essentially have some sort of little tab marker above it or something, maybe with an image of the queen or similar?

[JeffNevington]

Definitely, It needs a little queen or a house or something like that yeah.

[Jubal]

Yes, either would work

[WarpDogsVG]

Gonna be following this one, definitely love all the games you list as inspiration, and man, it's crazy there hasn't been a modern version of Sim Ant! Looks great, duder

[Silver Wolf]

This looks very interesting, I'll give it a try.

[Jubal]

Can confirm, is indeed very interesting!