The Basingstoke Post-Mortem, part 1

What is Basingstoke?

Basingstoke is a mediocre post-war town in southern England, in the heart of what passes for Silicon Valley in our wretched island. It's just like California, honestly, only without the sun, or the money. 

It is also a stealth action game that was made by us, being Puppygames, and the game is all about escaping from the eponymous town in the dead of night following an unfortunate scientific accident that triggers an invasion by the Titans, and kills everyone except you, as you were hiding in the toilet at the time.

Basingstoke, the game, is loosely defined as an action roguelike stealth game. The objective is simply to escape Basingstoke (like anyone sensible would), evading the monsters, and picking up useful loot along the way from bins and corpses and the like, and occasionally crafting something useful out of junk, just like real life. 

Why did we choose to make it?

Back in early 2014 we were in some sort of creative doldrums. We couldn't quite seem to get anything together or agree on what to make next. We'd had a number of aborted ideas: DLC for Droid Assault; some sort of multiplayer deathmatch cave kinda game; a swoopy dogfighting game inspired by Time Pilot; eventually settling on Battledroid, the successor to Revenge of the Titans, which was an idea we'd actually started thinking about back in 2011.

But alas! - the perennial problem of the procrastinating programmer pounced, and we ran out of money.

Battledroid was proving way too ambitious to attempt to complete with the meagre funds left. So we decided to have a go at a couple of little arcade projects. They were both meant to last no more than 6 months or so, the plan being to get something - anything! - released and get some income.

One game was the Skies of Titan (or Titan Interceptor - never did agree on a name) - carrying on in our Java tech that we'd developed over the last few years. The other game was Basingstoke, using some newfangled thing called Unity that Alli and Chaz seemed to enjoy using. Both games were in 2D and we were developing them in parallel for a couple of months, with Chaz reckoning he could do the graphics for both games.

But three factors got in the way of our brilliant wheeze. For one thing, in reality, it wasn't really feasible to provide graphics for two games simultaneously. It's a full-time job putting graphics in one game! For another thing, doing stuff in 2D in Unity was (maybe still is) rather painful, fighting all the way against its 3D leanings - it was originally designed as a sort of top-down 3D view with 2D sprites - so we decided to make Basingstoke a fully 3D game with isometric perspective, fully utilising the Unity engine. In hindsight, this switch to 3D may have been the biggest mistake we made, but don't worry, we made (and continue to make!) plenty more mistakes and you can pick and choose your favourite at the end. And finally, we couldn’t quite settle on a style of gameplay in Skies of Titan that we all liked - it felt great when it was fast and swoopy, but unplayable; and it felt turgid when it was slow and precise.

First commit to source control on 6th June 2014. Release date on Steam 27th April 2018, followed by a couple of patches and some extra content a bit after release. That's four years of development, with 2 people working full time on it for the entire time, and in the latter two years, increasing amounts of a third. Let's call it 10 man-years because it makes sums easier.

So, what went right?

Finishing and Reception

Firstly, and most importantly of all, Basingstoke was actually finished and released. Not only was it released, but it was actually really well received, getting great reviews, and a lot of praise from players. At this moment in time it has an 81% approval rating ("Very Positive") which is a fine rating for a game with 10,000 customers. 

Evolution of Concept

Here at Puppygames, every game we start tends to bear almost no resemblance to every game we finish. We find that as we explore each design element to its logical conclusion, adjusting or discarding those that have gone before it, that a game design naturally follows a twisty turny path to perfection. We frequently hit dead ends in design and have to backtrack over weeks of work, which is probably inefficient - but the end results are, in our humble opinions, the very distilled essence of gameplay perfection. I don’t really understand how other studios manage to simply produce awesome games in no time at all - it just seems to take us a long time - but the end results are at least worth it.

Basingstoke in this regard was no different. started out as an almost completely different game. The initial premise was that it was going to be more like Droid Assault or Teleglitch, with a large procgen open world and survival elements. That idea fell by the wayside (pure procgen is hard), and gradually the game morphed into a stealth game and away from run-and-gun. We kept the roguelike aspects of random procgen and exploration, but switched to a level based system with randomly connected hand-designed sections, and kept on iterating on loot ideas.

We really wanted to differentiate the game from being a run-and-gun, and focus on the stealth aspect, so we removed "hitpoints" (instant death) from both player and monsters. If they get you, you die. This meant we didn't have to deal with all the mechanics of hitpoints and damage and healing and all that tedious guff present in so many other games. If you are able to hurt them, they are incapacitated or killed - no repeated battering to eventually bludgeon your way through a problem.

Finally we added crafting as a way to combine simple loot into cool loot, which we wholesomely pinched from the excellent but extremely hard Teleglitch.

We're highly proud of the whole design: it is unique, taking a few ideas from other games, and mixing them up into something that no other game quite does the same. It has beautiful 3D graphics and animation and environments and lighting, and a slick UI rivalling games with ten times the budget and manpower. And the original soundtrack is atmospheric, bespoke, and sets the mood well.

The character unlock, level progression, campsite and backpack design we're particularly pleased with. It gives a meta-meta-game to Basingstoke: the campsites persist between games, gradually building up loot to make the game more varied, or more interesting, or more easier for you, every time you progress and then have to retry.

Accessibility

We implemented a stats tracking system, which in short order convinced everyone that the game was way too hard, with most players not even managing to get past level 1.

An idea from our other games was a sliding difficulty multiplier, which meant that even though we started out easy, we rapidly adapted to giving the player a challenge, and rapidly adapted to them failing to rise to meet that challenge. All borne out in the stats. So that was a good move.

Also related to balance, random number generators are the bane of roguelikes. It just takes a series of crappy rolls for a player to ragequit and never come back.

We took pains to avoid this by using RNG in a different way. Instead of the traditional RPG "dice roll", we took our ideas from card games instead: there is a deck of loot, with filters on it to discard inappropriate cards to the bottom of the deck. Eventually, every game is fair.

Interestingly, this means the contents of lootable items in the game is not known before time - it occurs upon looting, not at level generation time.

Because of our involvement in the past with Special Effect, we paid particular attention to how people of differing ability would be able to enjoy the game. Could we make the game far, far, easier without actually making it any less fun? Yes, we could. Basingstoke was not meant to be a challenge in dexterity.

Controls wise, we settled on a design that meant the entire game can be played with the mouse using just one hand (if it was a gaming mouse with 5 buttons and a scrollwheel, at least). This greatly increased accessibility - turns out a lot of people can't use a keyboard and mouse simultaneously.

We also settled on a stop-the-world inventory system. The game pauses while you're shuffling about in your backpack selecting things to lob or craft. This also greatly increased accessibility: an awful lot of people are simply flat out incapable of controlling their character while at the same time making complex mental gymnastics about what to throw next (and never mind crafting). Including me.

Style and Design

The isometric design was chiefly conceived in order to be easy to make. The characters themselves didn’t have arms and legs to cut down wasted work on animation - you’ll notice that while playing, you don’t actually notice they have no arms and legs, they are simply implied. The whole look and feel, with chubby little cars and bold colours and such, manages to somehow maintain our signature style but transposed into three dimensions. All in all the graphical style saved us considerable effort over more traditional 3rd-person isometric styles.

Toolchain

Unity gets a lot of flak from "real" developers. And indeed, I'm going to pan it in the What Went Wrong section below, for reasons. However, it made a lot of things really, really very easy, saving literally years of development time if we'd had to make our own bespoke engine... which is exactly what we're doing with Battledroid. Without Unity we simply couldn't have made Basingstoke at all.

We used Autodesk 3DSMax for 3D modelling and animation. It's great, and Chaz has 20 years' experience using it - he's totally fluent. It is also, however, ludicrously expensive, and cost us literally thousands of dollars over the course of development. We don't use it any more as we just can’t afford it.

Adobe Photoshop also featured in our toolchain. We don't use it any more. Adobe will never see another penny from us again so long as I live and breathe. The tools themselves we can’t fault, but the sneaky underhanded tricks Adobe pulled to screw us out of money are unforgivable.

We use Subversion for source code control. SVN is great because it works natively perfectly well with large binary files, it's really easy to understand, it's mature, it's plenty fast, and we all intuitively understand what it's doing and how it works. Suck it down, bearded git users.

Early Playtesting

We got Basingstoke in the hands of playtesters pretty early. It spent almost two years being heavily playtested, changing it from a bug riddled, confusing and incoherently designed mess in to the accessible, consistent, much less bug riddled and tightly implemented concept it was at release.

We weren't overly precious about radical game design changes even three years into the cycle. The stop-the-world inventory came very late in the design. The weapon damage system, likewise. Even the stealth mechanics weren't really figured out until late in the design. Crafting only appeared after two years. Being open to radical change in design helped Basingstoke be the very best it could be, and one of the very best in its niche.

Early Marketing Efforts

This time around we announced the game as early as possible - I think we were first mentioned in the news in 2014! We got out an awesome trailer in 2016 (anticipating a timely release, haha). This raised a little bit of awareness, though if I'm honest, though we went through the right motions, I'm not sure we were successful.

Launch (The Good Bit)

Some aspects of the launch went reasonably well. We sent a series of mailshots leading up to and upon release, reaching ~20,000 people each time. We had relatively few serious bugs at launch, and they were rapidly patched.

Well, that's enough of singing our own praises. Look at us, see how l33t our game design skillz are.

£3270 a year. Go and make another cup of tea and think about that number again for a bit, and come back to read part 2 next week...