Skip to main content

Console Gamers Aren't PC Gamers

Introversion's Byron Atkinson-Jones discusses the testing process for the company's games, and the lessons learned

As we continue the series of editorial features penned by the popular indie team at Introversion, the spotlight moves onto the testing and feedback process. Producer Byron Atkinson-Jones explains some of the challenges the company has faced in the past, and a few of the lessons learned.

Who Needs Testing?

Working for a small developer like Introversion has some fantastic benefits. The entrepreneurial vibe makes for a fantastically rich and creative environment, you have the opportunity to get involved in all aspects of running a business you wouldn’t usually get a chance to work in, and our flexible work ethos ensures a good work life/balance.

There are of course down sides too, and one of those is that we don’t have a massively large amount of resources to play with, which means that important activities like game testing are down to us and nobody else. This can have a major impact on the playability of the game because as anyone who's developed a game knows, being the developer makes you fundamentally too subjective and biased to be a good play-tester.

When you've taken an idea, nurtured its embryonic form, built it lovingly from scratch into your fully-fledged all-singing, all-dancing next IP, it can be almost impossible to see it from the naive perspective of the first-time player, and yet this is essential if you are to make a game that appeals and makes sense to your audience. Our experiences developing Darwinia+ for XBLA and working with Microsoft have taught us a great deal about the fine art of playtesting and we have come a long way from where we started.

Using Feedback Well

From the beginning, Introversion has always relied on in-house testing and help from a loyal and supportive group of beta testers, most of whom come from the Introversion forums. This approach - although limited - seemed to work well enough for Introversion in the past; since we made PC games it was easy to deliver preview code to third parties and particularly important, it was cheap.

So the first time we really got bitten from using this approach came as something of a nasty shock. It happened last year, during Multiwinia's development. With the PC version of Multiwinia close to release, we trekked down to the Future offices, as is becoming something of an IV tradition, to show preview copies of the game to the likes of PC Gamer.

Once we had got past the usual hassles of network setups and praying hard that the game would actually run at all, the preview seemed to go well - it wasn't until we were out of the offices that Chris revealed that the guys at PC Gamer had not liked the control scheme at all. This had devastated Chris and he then spent the next two weeks locked in a room coming up with a better control scheme based upon the feedback he got.

You Can't Please Everybody, But You Still Have to Try

It's a familiar problem in game development coming up with a gameplay mechanic that will appeal to many players rather than just a few. When we as developers sit down and design a control scheme we don't deliberately set out to design a scheme that is as obscure as possible, but rather one that makes sense. We then implement and test it ourselves and if something doesn't work then we go through the process again and again until we settle on something that does.

The key thing here is that the control scheme appears to work for us - up till this point nobody but the developers have had a chance to play with it. Since the controls tend to be put into place rather early on, the developer gets quickly acclimatised to his chosen control scheme and it's easy to become convinced that what we have now is actually a very good system and therefore best left alone.

This introverted approach to playtesting also backfired for us with Darwinia+ - just before we entered into the Code Complete phase of the project our account manager at Microsoft expressed some concerns over the control scheme we had in place. Our problem was one of coherence between Darwinia and Multiwinia and also the fact that our control scheme made use of context sensitivity which meant that controls had meanings based on the current action.

What had felt automatically intuitive to us was deemed to be confusing by Microsoft and they wanted to suggest alternatives schemes. Many months of iterations ensued before we finally settled on a control scheme that both parties were pleased with. Again, we have no real way of knowing if it will be universally liked but we are a lot more confident this time round because Microsoft had introduced us to the concept of usability labs.

Related topics