Grand Thieves & Tomb Raiders

by Rebecca Levene

Commodore sent them the hardware: a couple of top-of-the-range Amiga 1000s. Eventually the company realised its error, but by then Taurus Impex had started developing a powerful database called Acquisition, which made a good fit with Commdore’s plan to sell the Amiga to businesses. The two firms developed a close relationship during this time, and Commodore kept sending Amigas to Taurus’s tiny loft offices. Eventually Edgar and Molyneux had ten machines, a modestly selling database, and a daily barrage of phone calls from customers who were struggling to make use of it. ‘It was an extremely complicated relational database and it took all our time and effort and money to support it,’ says Edgar. ‘And there were loads of bugs in it – it was a real pain.’

  While they spent their days debugging and fielding calls for an under-performing product, they were running out of cash. It was a chance meeting with Andrew Bailey that led to Taurus’s first games writing work. Bailey, along with brothers Simon and Dean Carter, had produced a fantasy shoot ’em up called Enlightenment: Druid II for the Commodore 64, and they were looking for a conversion to the Amiga. It was a lucky break, and another bluff for the team. ‘What they didn’t know at the time was that we didn’t even know how to get an object across the screen,’ says Edgar. ‘Database work didn’t require that.’ Nonetheless, experienced Amiga programmers were rare, and Molyneux and Edgar secured a deal with Telecomsoft to complete the conversion. They were paid just £4,000, but according to Edgar ‘it kept the beasts from the door’.

  They still needed an artist, though. Instead, they found a programmer called Glenn Corpes. ‘I got an interview which became a three hour casual chat with Peter followed by being informed that they had no programming vacancies,’ Corpes recalls. But during the interview, he had been toying with the Amiga art package Deluxe Paint, and it was enough to secure him the job. ‘Mostly thanks to the complete lack of any artistic ability of everyone else in the room.’

  The Enlightenment port was a moderate success and with it came a new brand name for the company: Bullfrog Productions. Its first original game followed in 1988, another shoot ’em up called Fusion. It troubled neither the critics nor the charts.

  The cash situation had barely eased, and by now Bullfrog was in trouble. ‘We were living hand to mouth,’ recalls Edgar. ‘We got quite a big pay off when we released Acquisition, but were down to the last few pennies.’ But the atmosphere in the company was good – Molyneux could be inspirational – and most months they found the cash to pay their staff. ‘Druid II and Fusion only brought in a fraction of the money needed to pay the wage bill,’ says Corpes. ‘You didn’t need to be a genius to work that out.’

  He was paid throughout that time, though, even when he had stopped being given any work to do. ‘I thought I could see the writing on the wall, so I decided I better brush up my programming skills.’ Corpes brought his Atari ST in from his home and set about porting Fusion to it. It quickly became clear that it wasn’t going to be as easy as some conversions – the Atari machine would need to use tricks to match some of the Amiga’s specialist graphics hardware – and he became distracted by another idea.

  ‘I was fascinated by the isometric 3D graphics that had been huge on the 8-bit machines a few years earlier,’ says Corpes. ‘My favourite of these was Paul Shirley’s Spindizzy.’ It was a game which used the same 3D style that Knightlore had, but this time where the player guided a spinning top around obstacle courses of ramps and pits. The levels were built around sloping hills connecting remote plains of varying heights, and Corpes set about creating a similar set of building blocks that could be used to make up the various parts of an isometric landscape.

  After eighteen hours, half of them spent drawing the blocks, Corpes had a working demonstration. His creation was a matrix of squares, like graph paper, but seen from the side as if it were lying on a table in a three-quarter perspective. Each point of the matrix could have a different height off the table, creating an image that could look like a three-dimensional drawing of hills and valleys. But Corpes’ initial routine chose the height of each point randomly, and the image looked less like a landscape than a crystalline mess. There were thousands of points in the matrix – adjusting each one individually would have taken days. ‘All I had as “level” data was a bunch of random blocks, and I was far too impatient to write a level editor.’

  Corpes’ solution was to write a routine that would do the work for him: ‘I figured a way of generating landscapes using this set of blocks.’ They didn’t look quite right, though: they were more like intersecting pyramids than a natural landscape, with very few flat areas. ‘So I added tools to raise and lower points just to make it look a little nicer. At the time, I had no idea that a whole game would evolve around that mechanic.’

  The entire Bullfrog team were intrigued. ‘It was one of those demos that just made people get excited when they saw it. We talked a little about where it might go next,’ says Corpes. And Molyneux became obsessed with it. He asked Corpes to send him the code, and worked on it for days in a miasma of cigarettes, cola and pizza. He was using such shaky equipment that he had to keep every line of it as short as possible – his monitor screen was prone to warping if any text extended to the right-hand side. And he was by no means an expert coder, as he admitted in his speech: ‘I did go to the pub with David Braben and Jez San – they were proper coders – and they almost laughed me out of the pub for programming in C.’

  But it didn’t matter – he transformed a graphical toy into a living land. Copes describes how, ‘Peter disappeared into the other room for several days, and when he emerged he’d added houses and people.’ It was far more than a simple aesthetic amendment. The ‘people’ were dozens of tiny human beings, only a few pixels high, who clambered over Corpes’ landscape seemingly of their own free will. Indeed, they were independent of the player and endearingly liberated – Molyneux had created a mechanism that encouraged them to travel, to strike out to parts of the virtual world that were uninhabited. But although they could disperse, they stopped when they reached a barrier, and Molyneux didn’t know how to write a ‘wall-hugging routine’ – a set of rules to tell them how to behave naturally when they couldn’t go any further. So instead he exploited the manual height adjustment that Corpes had implemented: he allowed the player to influence the travel of the population with the shape of the landscape. If players wanted some of the tiny people to head somewhere, they could tweak tips of hills to create paths for them. Molyneux, in his twenties and working in an all-male office, christened this process ‘nippling’ the land.

  Other innovations grew from the limitations of both the computer and of Corpes’ design. Molyneux wanted a large population, but the numbers could become overwhelming for both the processor and the user. He added a feature whereby the people would build a house if they stopped on a flat piece of land. And, of course, due to the tendency of the land generator to create weird hills, flat land started out as a scarce resource. So the player had something to do: create plains and shuffle the population towards them. Once in a house, the people would be considered settled, and the headcount would grow. And if, under the player’s guidance, the land under a house was raised, its inhabitants would leave and set off again.

  Corpes had written the original version on his own Atari ST, but it was ported to the Amiga using a cable the team had in place for playing Geoff Crammond’s Stunt Car Racer. Following the example of that game, the landscape was made multiplayer – two people on two computers could each move an army of people on a single landscape. Even in this early form it was very addictive – simply sinking your opponent’s land and people was delicious fun. It burgeoned into a game.

  Over weeks of playtesting, features were added and tweaked to give focus. Spells helped: they were mostly natural disasters such as volcanoes, earthquakes and swamps that could be visited upon the enemy’s people. At first these were available at will, but such unlimited calamity drained the game of its tension. Molyneux had a brainwave: the players would need ‘m
ana’ to deploy them, a currency which they could earn from establishing settlements. With a simple tweak, there was now a gripping purpose to both building houses and destroying your rival’s.

  They called their project Creation. The aim was straightforward: to ensure that your band of settlers prospered, while the tribe led by your opponent found itself driven off its land and dwindling into extinction. But you had no one character to control – instead you had power over the land and the elements, and guided your people as an unseen deity shaping their fate. Messy and unintended as it was, this was the birth of a new genre: the God game.

  Nobody was yet calling it that, though. It was a real-time, two-player strategy game, and while having an opponent made it phenomenally addictive, there were few people who had more than one of the machines required in the same room – it would struggle to pass as a commercial product. So Glen Corpes was given the job of writing some artificial intelligence to enable the computer to run one of the tribes. It was a complicated game, one which had occupied the full attention of its creators. Yet the AI routines used to reproduce their thinking were extraordinarily simple: the computer would look for potential settlements and try to expand them; it saved up for a spell at random; and for combat, the computer’s people would attack the player’s oldest building. These were short cuts, but they worked: the anonymous computer opponent gave a convincing show of a smart adversary marshalling tactics and strategy. ‘Sometimes with AI, especially with big crowds of characters, the whole is more than the sum of its parts,’ says Corpes. ‘People see behaviour that isn’t there.’

  The game was starting to look exceptional: it was novel and very compelling. The Bullfrog team would play a single session for hours, which was a sign of its quality, but also a symptom of its greatest flaw. The very thing that made the game unique – the fact that the people who filled the landscape could only be guided, not directly controlled – meant that forcing a final confrontation was surprisingly difficult.

  Molyneux tried a series of solutions, and each helped a little. He introduced a ‘Papal Magnate’, a bizarre choice of name for the ability to order groups of people to particular points on the map – wherever the player wanted to build settlements or engage in battle. There was also a hero character called the ‘Knight’, formed when dozens of the player’s people combined into one super-powerful being who could take out opponents with a swipe or two of its giant sword. The final piece, though, was an all-or-nothing endgame, the ‘Armageddon’ spell. It was cripplingly expensive for the player, but would only be needed once: it made every house in the land throw out its inhabitants, whereupon they would fly into a final, epic confrontation. Victory, for someone, was assured.

  The whole team became fanatical about Creation. They would work on it during the day, play it after hours until ten in the evening, and then go to the pub and discuss it some more. As the game neared completion, Edgar started showing it to publishers. But the bestselling games on the Amiga and Atari ST played like arcade titles and were showcases of cutting-edge graphics – most publishers weren’t interested in a quirky strategy game without any shooting. ‘Mirrorsoft threw us out laughing,’ recalls Edgar. ‘It reminds me now of the Beatles, but at the time we thought, “Maybe they’re right?”’

  During development, the team had created a Lego version of the game for visualising and playtesting ideas before they were coded, so Edgar tried showing the game to Lego itself, hoping that it would suit the company’s branding. ‘I said: “Look, we could make a really cool Lego game. It’s perfect – the building blocks, the isometric view – it was very Lego-like.” And they said, “No. Because there’s violence in it.”’ Edgar was incredulous – he pointed out that they already had cowboys and spacemen. ‘But they wouldn’t have it’.

  For some reason, Bullfrog’s most obvious port of call was also its last. Electronic Arts, a large US publisher at a time before there were any games-publishing giants, had opened a UK office. It had published Bullfrog’s only other original title, Fusion, but Edgar and Molyneux thought that Creation would be too alternative for EA. They were wrong.

  ‘We showed it to EA and they loved it,’ recalls Edgar. ‘They saw the potential. We didn’t really understand how big it could be, but EA had the vision: that it could be successful worldwide, in that it was non-violent, it was cool, it was new.’ Bullfrog was offered an advance of £20,000, which covered the game’s development costs and was ‘like a new lease of life’. The company had long passed the point where its bank would prop it up.

  As the process neared its end, a few grace notes were added. Edgar hit on the idea of adding the sound of a thumping heart, its rate slowly increasing. ‘I felt it lacked a sense of urgency as the game was progressing,’ says Edgar. They trialled the game with and without the effect, and found it superbly ramped up the tension. ‘It’s one tiny, quirky little thing, but it makes an enormous difference.’ Glenn Corpes turned his hand back to the art, and added a few flourishes to the design. He presented the play area as a scene in the pages of an open book lying on a desk, heightening the idea of an omniscient being watching the story of a minute world unfold. And it was Electronic Arts which chose the final name. Creation had been copyrighted elsewhere, and in any case, this was a game about guiding your people. How about Populous?

  Everyone approved. A marketing image of an island floating in space was devised, and the game would be packaged in a glossy, outsized cardboard box – now the standard form for prestige titles. It looked terrific and would easily hold its own against games from higher-profile developers.

  Soon after, Populous was in the hands of the press. As Bullfrog waited for the reviews of its unconventional game, it received a message that a magazine would be sending a reviewer in person before giving the final score: Bob Wade, from industry favourite ACE magazine. He was a long-standing and well-respected games journalist, and Molyneux regarded him very highly. Too nervous to ask Wade what he thought of Populous, Molyneux took him to the pub, where the two of them became roaring drunk – Molyneux later claimed that he drank fourteen pints in slightly over five hours, and had to excuse himself to throw up. Finally, he summoned the courage to ask this famed journalist, a veteran of hundreds of games reviews, for his opinion. And Wade told him: it was the best game he’d ever played.

  Molyneux was convinced that Wade would change his mind if he ever returned to the office and played the game again, so he detained him in the pub, force-feeding him beer. It worked: although Wade did ask to go back to Bullfrog’s offices for a two-player match, the two of them collapsed into an alley on the way.

  ACE marked games with implausible precision, but was respected for its cutting honesty. Populous received a score of 963 out of 1,000, one of the highest ever. ‘Populous is a terrific game,’ Wade said in his review. ‘Absolutely wonderful stuff that will keep you playing and playing.’ Other magazines lined up to applaud it: ‘All the magazines loved the previews,’ recalls Corpes. ‘It was our third game and we could tell that journalists suddenly weren’t just going through the motions while asking about it.’

  It was a critical success, but an odd game and still difficult to sell. What exactly was this mutant hybrid of a strategy game and a world-builder? How should they describe it? In fact, Wade had already given an answer. He – or perhaps the staff at ACE – coined a phrase to describe the new genre Populous had pioneered: ‘God game’. It’s an ideal name, immediately graspable and hugely appealing. Who wouldn’t want a game that gave you the chance to act like a deity?

  It was released in March 1989 and debuted at the top of the charts, but its fame had spread beyond gaming circles. Although the packaging made no mention of taking on the role of a god, it did talk about deploying the ‘power of light or the force of darkness’. A month earlier, Salman Rushdie had been taken into hiding after Ayatollah Khomeini had placed a bounty on his life for blasphemy, and Britain had broken off diplomatic relations with Iran. At the height of the fervour, the Daily Mail conta
cted Molyneux, fishing for quotes about the possibility of the game earning him a fatwa for daring to play God. If it was a public-spirited concern, it turned out to be unfounded.

  Unannounced, a cheque arrived at Bullfrog’s offices. Bullfrog had negotiated a ten per cent royalty with EA, increasing to fifteen after a million units had been sold, but it’s not uncommon for ancillary costs to swallow the entire amount before it reaches the developer, so any payment at all was a surprise. ‘The first royalty payment was pretty small, I think it was about £13,000,’ recalls Edgar. ‘Which I think Peter and I split, less a thousand in the bank or something. Because we couldn’t believe it, we thought, “It’s never going to happen again.”’ The royalty was so unexpected, they rang EA and asked if it was correct: ‘And they said, “Yeah, but the next one should be a bit bigger.” And it was – a lot bigger.’

  That £13,000 had been the tail end of a quarter. Three months later, they received a full royalty payment. ‘That was substantial,’ recalls Edgar. ‘It was two hundred odd thousand. And it was unannounced as well. They didn’t tell us about either of these – we didn’t know. And they kept coming.’

  It was common for hit games to earn their money within the first few weeks, and then fade away. But Populous kept going – the trade press listed the bestselling games each week, and Populous stayed in the top ten for months, as Corpes remembers: ‘We papered a wall of Les’s office with charts of all the weeks that Populous was at number one.’

  The money transformed the company. ‘It was life-changing,’ says Edgar. ‘We were no longer scrabbling around, robbing Peter to pay Paul, worrying about salaries, wondering whether the tap would be shut off.’ Corpes sensed the change too: ‘I didn’t get my November 1988 pay until January ‘89. But by the end of the year I’d been paid over 200 per cent in bonuses.’

  With success, Bullfrog shifted gear: once a code shop that bluffed to secure computers and contracts, now it was the creative powerhouse that had invented an entirely new genre of computer game. But perhaps that had always been in the company’s DNA: it didn’t work to a corporate schedule or market research, but found an unlikely idea, and pursued and finessed it until it shone.