Tip:
Highlight text to annotate it
X
I can split my time in game development up into 3 stages. Up till 2004 I made ambitious
games but didn't know what I was doing. From 2011 onwards I made simple but fun flash games
with decent amounts of polish. But between these times was when I made my most exciting
projects. Where I had the ambition to think them up and the ability to pull them off-
even if they were a little rough around the edges. Two of my favourites were made in late
2009 and I've been looking forward to telling you about them for as long as this series
has been going. First was The Almost Heaven Crisis. I feel
this whole series has been leading up to this game. Which is funny, 'cos it's not a 'good'
'game'. The controls are fiddly, people lack animations, their AI is abysmal and the world
is buggy. But it's the most complicated game that I've ever released- and in part because
of this, it's also very creepy. A bit of a back-story into the location.
There's a real place in America where people live up in the mountains, armed to the teeth
with weapons in case the government ever tries to take away their freedom or whatever. The
place is called 'Almost Heaven' and I watched a documentary on it with Comacchio and was
like 'Wow imagine a zombie outbreak there!'. We travelled to Luxulyan valley for inspiration,
since it's the most depressing network of ruins and aquaducts known to mankind and was
a setting for one of the Omen movies. I then started work on the game, first focusing
on giving the game ambience. There's a 24 hour cycle and numerous weather conditions,
all with their own unique sound effects and lighting styles. It's amazing how much of
a difference sound can make. The other citizens help to shoot the zombies, but apart from
that they're mute, which just adds to the sense of being alone in a very hostile place.
It was also the first game of mine to load level layouts. The game itself played out
on a single frame which meant that any changes to the graphics or gameplay only had to be
done on that one level and it would carry across to all of the game's levels. It was
a novelty for me at the time but is a necessity for any complicated game projects. The world
is made up of a dozen or so individual areas which you can jump between by running off
the side of the screen. In one of them starts a hoard of zombies who can quickly infect
other NPCs. Your goal is to kill them all, or just to
survive for as long as possible. Easy, right? You just find a part of the map where there
aren't any zombies. Only... you can't. Because they'll find you. Every character in the game
wanders about between the zones, even if they aren't loaded. This is what makes the game
eerie. You'll be walking through the middle of nowhere, only to come across a guy who
you previously met a few days ago in a town on the other side of the world. Or maybe he'll
be a zombie by now. Who knows. The Almost Heaven Crisis is a very successful
and playable experiment into what is essentially Radiant AI. The world isn't static. It's living,
breathing and changing around the player and makes you feel like a very small and insignificant
cog in a much larger machine. The impact it has on how the game feels is MASSIVE. Even
as the developer, who knows how it works, it still surprises me. This gets me excited
because normally if you make something, the magic is lost since you can see through the
illusion. But not so with this project. The replayability and potential for gameplay is
massive. ...but this comes at a price. Even after dozens
of hours of playtesting, the game is still super buggy. Sometimes, NPCs refuse to shoot
zombies. Other times, it says that there are people in an area when there aren't. These
are because of the leaps of faith I had to make when developing the game. Ideally you
want to add a bit and then to see if it works. But with such a complicated system, where
the whole world and every character is stored across several databases, I'd often have to
work blind for hours before reaching the next testable state. And even then, it can sometimes
be hard to tell if it's working properly or not. I often have to wait for a situation
to happen during actual gameplay. And this is despite the underlying mechanics
being surprisingly simple. All of the NPCs are stored in a database which records their
location. Every 15 seconds, it randomly moves them by 1 zone. They can't even infect people
or die if they're not in the same zone as you are and once they enter your current location,
they don't leave it again. But you can, obviously. These limitations were done deliberately because
I quickly realised that even with hundreds of NPCs roaming the world, if the majority
of the action goes on off-screen, it's boring. And you'd think it would be predictable if
everything happened on the same frame as yours. But it isn't at all. If I hadn't told you
the inner workings, you would never have guessed it. When you enter a new area, the game has
to pretend it knows where everybody within that zone is standing. It places them in clear
areas, deciding where infected and non-infected people should start based on the mixture and
so on. It all works better than I could have imagined- and that's rare in the world of
game design. In fact, this game's success was its downfall.
This 'radiant AI' was so impressive that it totally changed the game into something I
hadn't prepared for. With all of the bugs that I was encountering and the things that
I learned about open-ended gameplay since the start of this project, I decided that
it was best to learn what I could from this project and to start it again from scratch,
designing the game around this key feature instead. Don't get me wrong- the original
game could support a lot of cool features that I didn't get around to properly implementing,
like destructible terrain and unique characters. Every NPC -even the zombies- have their own
unique names and stats, as well as hair, skin and clothing colours. I even wanted to make
it so that you could talk to them like a text-based adventure game, asking them to follow you
and so on but never got around to any of these things once I got the radiant AI working.
But despite all that, the original game is beatable. It always ends with a frustrating
chase as you hunt the last zombie down as he moves about the world.
I have exciting plans for a sequel and have toyed with the idea for years. I have a working
editor where you can quickly and easily construct a world and the game itself handles everything
else, including how NPCs explore it. I had grenades that could blow up walls and it was
going to be a key feature to make the combat feel juicy. A lot of the ideas found their
way into Destruction Darius, even though it's a completely different genre.
I wanted there to be strongholds, patrolled by professional gun-wielding NPCs. I wanted
shipments of new people to fuel the zombie hoard and even wanted action to go on in other
parts of the world, instead relying on THOUSANDS of NPCs to make the place feel alive. I experimented
with the idea of 'crowds' where, instead of randomly moving from place to place, like-minded
NPCs would group up so it would lead to huge battles where civilians clashed with zombies.
So the question is: why haven't I made this? I don't have an answer. I load it up a few
times a year, add some more stuff to it and then leave it alone again as other projects
take priority. If I didn't have other things to do, I would definitely have done something
with it. I know it's all possible. There isn't anything technically standing in the way of
this game working! It's just that the undertaking is so large when compared with anything I've
done before, I assume I'm going to fail, or run out of steam long before the end. It's
one of those things that I'd like to get done before I die though.
Which brings me onto the second game, made just a month afterwards. It's very different,
but is definitely the better 'game'. Bat Sweeper was made in just 3 short days
for a 'Happy Halloween' Game Builder competition. The idea was that your character can blow
and suck things with his hoover to help him to navigate pits, collect items and to protect
himself from enemies. I can't take credit for this, a friend called JosiahCT was the
brains behind a lot of these gameplay elements, so a big thank you to him for making this
project possible. Probably the best thing about making this
game was adding the voice and video files. I dubbed Tommyttk's voice over my narration
and got my friend Ben (Previously Casper from Santasatnas) to do the talking bat... though
I didn't tell him what his role was. I voiced the Bat Sweeper himself and hate myself for
using a voice that I'll never be able to top. I also originally voiced his Mum but Comacchio
did a much better job. The whole thing was a light-hearted joke between friends.
Of course, storylines are very important to me so I ensure that all of my games features
an award-winning script. For the Bat Sweeper, a man who likes sweeping up rubbish sucks
a bat up, so the King of Bats tells him to stop but he doesn't so the King of Bats kidnaps
his Mum. Eventually the Bat Sweeper confronts him again, realises it's all a misunderstanding
so returns the bats... at which point I suddenly realised that there wasn't much time left
before the competition deadline and that the whole Mum kidnapping hadn't been resolved.
So I did this. I also had to cut out a lot of the middle levels for the same reason and
nicely covered it up with this. Do you think anybody noticed?
Until Destruction Darius I'd probably say this was my best game. It's simple and it's
fun, which is the exact opposite of the Sundown Shambles game that I had made the previous
year. I was learning! And the feedback reflected this. After years and years of trying and
failing, I had finally earned my place in the top 100 games list on the Daily Click.
It got Universally positive ratings and is to this day, the 64th highest rated game on
the site. It even beat James Luke's Death Giver 2.