Back in 2005, I began to have ideas about how to make very
stripped down, cheap to develop games that would be interesting enough that
they might make some money in the budget market. Of the many “verb game” ideas
we originally had, one of them stood out as the easiest to get under way, and
that was Fireball, the game that
would become Play with Fire. (We weren't able to use the name Fireball owing to an IP conflict with a curmudgeonly fellow on the Isle of Man).
post examines the many things that went wrong with this project – and also the
many things that worked out nicely.
1. No Money
By far the biggest problem that the project faced was the
lack of budget. Now even from the outset, the plan was to develop cheaply – but
I had originally been thinking $50,000 for a budget PS2 title (coming in at the
end of the PS2’s life cycle, when there is the maximum possible installed base and a great
opportunity for budget titles). In fact, we did have a deal with a European
publisher to fund us for the full $50,000 – but for reasons that we will touch
upon shortly, we never saw this money, the PS2 version had to be scrapped, and
at that point the project was somewhat doomed.
The version we made had a budget of about $5,000. On the
whole, it is miraculous that we delivered anything!
2. No Passage to India
We set up a new developer in India – Fantasy Labs – to work
on this project, and began working with Sony in London to find a way to get a
PS2 devkit to India. This turned out to be monumentally difficult! Sony wasn’t
the problem – they were actually tremendously helpful and supportive – Indian customs
were the point of tension.
Perhaps if we’d been a larger company we could have greased
the wheels better, and found a way, but without getting the devkit into India
we couldn’t get the budget for the PS2 version, and eventually the battle had
gone on for so long that the publisher in question had to break the news that
it was too late to develop a budget PS2 title.
We were disconsolate, but at least we could ship the PC
3. No Tweaking Phase
Yet there was a major problem with the PC version: we didn’t
have the money to finish it. The most important stage in any game is tweaking
and blind testing, when you sit it down with players and observe their
reactions to the game. This phase allows you to eliminate any confusion on the
part of the player, smooth any rough edges, and generally streamline the play
of the game.
AAA games get months and months of tweaking time – in Nintendo
and a few other places, games have the luxury of not being released until they
are just right.
We had a zero-length tweaking phase. I conducted three blind
trials with the game – but without the budget to pay for the programming work
to fix the problems found, it was pointless continuing. Despite knowing we had
problems that needed to fix, there was nothing that could be done about it.
(One particular problem haunts me: the Puzzle Path, which
many players will struggle to enjoy, is directly in front when the game begins.
Most players end up trying this Path first, and get very confused. I really
wanted to make the Fun Path be directly in front – this Path is easiest to
understand, and comes with essentially no risk of failure. But without the
money, we couldn’t get even this simple fix made).
4. Spiraling Code Base
Additionally, we had a problem with the code base. Fantasy Labs had
another development contract at the time, which was helping meet the bills. But
this project required a much more complicated code base than Play with Fire, and sadly a decision was
made somewhere to keep both projects on the same code base. This meant that the
project was constantly being remade as the code base for the other project was
refined, giving no end of development problems.
5. Too High Spec
By far the most devastating of the problems resulting from
the spiraling code base was that the final game turned out to be too high spec
for players in our target audience to be able to play the game! Ironically, the
development tool set ran on any more machines than the final game – some people
who worked on the project have never seen the finished game running!
Depressingly, this has had the most negative effect on the
project. While we have had many downloads from Manifesto Games, the vast majority
were unable to get the game to run.
1. Verb Game Method
The basic idea of the “verb game” methodology was to
identify some unusual verbs with playful aspects, and develop these in
isolation into a game. I believe this worked brilliantly, and I’d do it again
if we had the budget.
The original plan – to be developing games in less than six
months, and to have a portfolio so we were not dependent upon the success or
failure of any one title – was completely sound. But of course, without the
money we couldn’t follow through.
Nonetheless, I believe the strategy was sound, it was simply
the implementation which let us down.
2. Arson IS fun!
Burning things to the ground in Play with Fire is tremendously entertaining once the player knows
what they are doing. I believe there’s potential for a future game to build on
this mechanic, although I very much doubt it will be done by us.
The play of the game is really quite original. Although some of the platforming elements work similarly to conventional
platform games, this is the only game I know of that has serious consequences
when you jump around indiscriminately, since the platforms you touch can burn
to the ground shortly afterwards, along with any others that are connected to them!
Just watching the flames spread through a big object like
the Trojan Horse in ‘Helen’, or the Houses of Parliament, or a WWI biplane is
3. Multiple Paths
The multiple paths to allow for players to play in different
ways was a sound idea – but alas no information is provided to the player to
judge how to approach this. This was a product of the lack of tweaking phase,
already mentioned. Despite this, the basic idea of structuring the game to be
played in different ways while using the same engine emerges as a sound
Anyone can play the Fun Path and get some enjoyment from it,
while Fiero-seeking players have the Challenge Path, and players who prefer
cerebral riddles have the Puzzle Path. Despite the problems, I’m still pleased
with how this worked out, although doing it again I might be tempted to offer a more explicit mode select screen.
4. Open Pool
The reason the multiple paths approach worked to some extent
was the tremendous talent of the Open Pool – a group of non-professional level
designers who worked on the project in return for a share of revenue (at least
on paper – in practice, without the PS2 version we were never going to see good
The open pool method worked something like this:
- Firstly, we invited people to get involved,
specifically looking for people who had not worked professionally in videogames
before. About two dozen people registered, signed the license agreement, and
were given the tools.
- Of these people, only about a quarter actually
produced anything like a working level. But those that did get to grips with
the tools immediately started producing interesting material.
- Just three or four of the members of the Open
Pool contributed 60% of the material in the final game (I made 40% of the
levels myself), but the quality of the best levels submitted were far better
than anything I could have made.
- We gave letters to the biggest contributors
promising them a share of revenue proportional to their contribution to the
This basic method is a brilliantly simple (and cheap!) way
of developing game materials, and one that I believe has further potential. Patrick
Dugan (who was one of the most productive members of the pool) dubbed it “the
Bateman Method”, which as a Brit I find slightly embarrassing, but it
underlines the confidence we have that this approach could be used successfully.
I want to take this opportunity to thank Maurizio Pozzobon, Patrick Dugan, Ian Tyrrell, Marc Majcher, Wil Evans and Toby Everett once again for their contributions to the project – you were the best part of the development process, by far, and I loved seeing your ideas come out in the game! I also want to thank ihobo troubleshooter Neil Bundy for getting involved in the level design late in the game, and also my wife, Adria, for making a simple level!
5. Recognised as Art!
And finally, it is a cause for minor celebration that Play with Fire was featured at IndieCade
in Sheffield. We may have lost money making the project, but it’s nice that
what we made has been recognised as being artistically interesting.
Although there were many problems that dragged this project
down to the point of commercial obscurity, I don’t regret pursuing it for a
second. This was the first time I have controlled every aspect of game
development, from production and design through to marketing and legal, on a
single project – and the first time I’ve used my own money to fund a project. I
learned a lot from having this opportunity, and more than that, I am actually
really pleased with Play with Fire
despite its problems.
There is a great game, buried under the hardware problems
and the confusing initial experience, one that I hope will be fondly remembered
by the small audience the game has received. If you have a machine that has the
necessary spec, and you haven’t tried Play
with Fire, why not give it a go? The demo is available from the Manifesto
Games site, and is free. (It’s $20 to unlock the full game). You really need to
see the game running to appreciate it – screenshots don’t capture the carnage
of watching the flames spread! (Just remember to try the Fun Path first - it's on the left when you start!)
I’m glad to say that we set out to make a game about burning
things to the ground, and with the help of the Open Pool, the good people at
Manifesto Games and a bit of luck, we succeeded.
My infinite thanks to everyone who helped on the project!