Swing for the fence - don’t sit on it - Learnings from 18 months of working on an Open world RPG feature

Firstly I’m under no illusions - (or maybe a little)

I’ve made, finished and launched games before (Flash/Android/Unity) and I’ve been making games even if just prototypes for around 20 years.

There are massive up-sides to working alone. No politics, no bureaucracy - no trying to understand someone else’s spaghetti code. No project manager telling you want to do and what features should go in. The freedom is intoxicating.

My GF jokes about me being a team of one but I do see the Unity asset store creators as kind of on my team - even if they don’t know it. (There’s a marketing campaign in there somewhere Unity.)

I'm totally in awe of indie developers who build all their own models as well as doing all the programming. But for this project I decided to use assets for the environment, and most Enemy models. I figured a tree is a tree right? No sense in me spending time building such things when the result wouldn’t be as good. But I will be making the NPCs and some of the enemies myself in Blender.

Technical approach

I’ve never worked in the game industry and expect some of this stuff might come across as naïve. But so far I’m going with the attitude that if an approach is working, keep doing it until it doesn’t and then re-evaluate and try something else. Pragmatic is the word I was fumbling for.

The game so far is just a demo which comprises a single Unity scene and some connected scenes for the underground dungeons/caves. Exploring it in its current state should take somewhere between 30 mins to an hour and I’m constantly adding new content with lots more to come over the next few weeks.

To achieve the open-world Outwith is built from Unity scenes connected with smaller scenes used as connector scenes. I refer to it as the “The Long Dark approach”. Lot’s of other games do this but The Long Dark is the one I have 800+ hours on in Steam so I’m going with that.

Clearly it means there’s some loading in between scenes but I don’t think it would be a major issue - even the best AAA games have lots of loading scenes these days - especially ones set in space :p

I did look at and evaluate world streaming but chose not to use it - kind of didn’t want to bet my game on it. It seemed to bring with it a lot of complexity that I couldn’t control - may end up going that route eventually but for now I'm sticking with connector scenes.

More generally I have enough understanding to achieve the current tasks at hand. If there’s something I don’t know in the past I’ve previously set aside a few weekends to learn enough of that subject to be dangerous and then move on. A Good example of this was rigging in Blender. I didn’t have enough knowledge at the time and had to learn what I needed in order to achieve what I wanted, which in this case was rigging the arms and hands.

Am I crazy to attempt building on open world RPG?

Obviously as feedback happens it may be in the form of “OMG this is rubbish why are you even bothering?”. If that does happen which I’m fully anticipating btw I can read between the lines and see if there is anything that’s fixable. I’m then staring down at a fork in the road: continue to fix and improve or stop and make a different, smaller game.

I’ve been making games long enough to know I’m not going to stop no matter how painful the wider experience can be. I’ve been on the receiving end of negative feedback and even abuse on Steam and I can tell you - being told to “Go to hell” because you didn’t have a sale is not nice.

Fighting entropy and clean code

If you don’t keep doing this the code-base and the project will end up as an unwieldy mess. Over the first year I rebuilt the project from scratch several times just to get organized again.

Also workflows are the key - or at least a fundamental part of it for me As a single dev it’s vital to get these down IMHO.

As an example the First person arms animations spells and weapons. I started out creating the animations in blender then importing them into Unity.

This worked okay but wasn’t ideal. I then discovered an asset which allows animating the arms directly in Unity. This vastly speeded things up. If you think about the longer term many weapons and spells will need to be created - maybe not as many as AAA equivalent games but Outwith still needs a reasonable number to feel rounded and complete in offering the player a choice of combat/play styles. Streamlining the process for creating these was essential.

There’s a separate model for the left arm and the right arm and a separate model/component tree for both arms which handles archery and two handed weapons. I particularly wanted to get the independent arms working as seen in a well known open world RPG we all know and love.

Everything is modular. I have a core prefab that can be dropped into any scene - It contains the player component tree, the UI and some Global game level managers. Obvs there’s dependencies everywhere but I don’t have to figure out other people’s code so it isn’t too bad.

This allows me to play test/ develop any scene and is also used in loading saved games i.e. loading a saved game instantiates the core prefab which contains everything needed

Pragmatism and Reality - final words

At the moment I’m waiting on feedback who knows? I would never say anything like this would be a waste of time but who knows what will transpire?

I hesitate to mention AAA RPGs as I don't want the comparison to be made immediately although it's probably inevitable.

Part of my challenge is how to make the result of that comparison, when it happens, favourable.

On top of this I have to force rank and prioritize based on players' expectations. Perfect is the enemy of good and all that 🙂Is it fun and if now what's stopping it from being fun and how can I fix it?

Expectations may be around beautiful First Person animations - I just fired up a well know open world RPG and yes they’re beautiful - mine will never be so but meh let’s just pursue this passion and just swing for the fence

Remember to download and give it a try out - and maybe if you’re feeling particularly awesome leave a comment.

Until next time