Free Radical’s story continues as Haze development stumbles, and a deal with Lucasarts marks the beginning of the end for the studio. VG247’s Dave Cook speaks with Steve Ellis to set records straight.
If you missed part one of my discussion with Free Radical co-founder Steve Ellis, you can play catch-up here. It covers everything that made the studio such a recognisable name in the console market, having first proven that the team’s work on GoldenEye 64 while employed by Rare was no fluke, and that yes, shooters had a comfortable home on boxed consoles.
While notoriety and critical acclaim came naturally to the Nottingham outfit, the company wasn’t shifting the kind of units it needed to break into the big time. EA had promised runaway sales of TimeSplitters: Future Perfect in its publisher pitch meeting with Free Radical, but a shift of focus to single-player narrative saw the game fall short of sale expectations.
“We weren’t going to be able to raise funding for another TimeSplitters game but increasing dev budgets were causing publishers to become much more risk-averse,” Ellis explained. “We wouldn’t be able to raise funding for a game that wasn’t a shooter and in some way similar to something that had been already successful either.
“The days of agreeing a deal for a new IP after one short meeting were long gone. We needed to make a shooter that wasn’t like TimeSplitters. This wasn’t a terrible development because many of us were ready to do something different.
“At the time there had been a lot of what we felt were fairly lazy ‘me too’ military shooters. We liked the idea of doing something more realistic, but didn’t want to make yet another WWII shooter. We came up with Haze.”
The key to Haze’s success, Ellis and his team decided, would be in state-of-the-art enemy and companion AI. The game focused on a dictatorship future governed by the ruthless Mantel corporation and ‘Nectar’, the drug it had engineered to give its personal army an edge in battling rebel group, ‘The Promised Hand’.
The drug would see enemies, friendlies and the player-character Shane Carpenter hallucinate, act crazy and have difficulty distinguishing friend from foe. It was an ambitious AI mechanic that would indeed require a strong scripting backbone to realise, but Free Radical was confident it could deliver.
The studio was so confident in fact, that it signed a publishing deal with Ubisoft in September of 2005 based almost entirely on the promises of its revolutionary AI system. It was a promise that the studio would be unable to keep.
Ellis explained the pitch, “Rather than us writing scripts to determine what each character – or class of characters – would do from a limited set of predetermined possibilities, we would give them the ability to decide for themselves. They would be able to construct complex multi-step plans to achieve their goals.
“They would co-operate. Behaviour would emerge. It was going to be revolutionary. In many ways, everything else about the game was irrelevant – the AI alone would distinguish it from every other game, and would reinforce our position as a premier FPS developer.”
This reliance on a single mechanic goes back to what Ellis stated previously in part one of this article – that betting the farm on a single publisher could prove disastrous if a game deal fell flat. While Free Radical had signed up with Ubisoft on the promise of a single mechanic, Ellis explained that it simply didn’t work.
“This came alongside another problem,” Ellis explained. “The engine that we had used for the previous games – the engine that I had originally created – was written in C, not C++ and relied heavily on things being done using code rather than data. This was the right decision at the time as it was a good fit for our abilities and experience, allowing us to hit the PS2 launch.
“However as the company had grown I had long since stopped being responsible for technical direction, and after we finished TimeSplitters 3 and started gearing up for next-gen development, all of the senior programmers – whom I trusted and respected greatly, and still do – said that they wanted to ditch the old engine and completely rewrite it in C++.
“It was felt that the limitations of the old engine would prove to be a hindrance and result in worse games, and that a ‘modern’, data-driven C++ engine would be better, more flexible, scalable and allow us to achieve more, quickly. We would be changing to do things ‘how the pro’s do it’. I agreed and gave them the go ahead.”
While Free Radical’s decision to switch to C++ was intended to speed up Haze’s development, Ellis confirmed that it served to slow things down. The switch also caused the game to become dogged by rampant performance issues, memory fragmentation and difficulty when working between teams. Many staff hated the change.
“Despite these problems,” Ellis continued, “We were contractually committed to delivering a particular game according to a schedule. Milestones had to be passed in order for us to get paid. If not, we wouldn’t be able to pay the staff.
“It was our only source of income, so we couldn’t afford for the money to stop or we would go out of business. We couldn’t insist on going back and starting again, so the only way was to go forward and try to fix the problems along the way – but it would definitely require more time.
“We bit the bullet and raised the issue with Ubisoft, who eventually agreed to give us a small amount of extra time and money to cover the time, coupled with new obligations, new penalties and extra monitoring. From that point on things only seemed to get worse.
“Where the programmers needed time to go back and fix the underlying problems, all we were able to give them was a new milestone schedule that still required them to deliver things that we weren’t yet ready to deliver convincingly. This meant that we entered into a cycle of “hack it in now, fix it after the milestone”, with the post-milestone fixes obviously having a similar impact on the next milestone.
“Nobody thought it was a good idea but there wasn’t an alternative. Soon we had to go back to Ubisoft again and try to get a meaningful extension that would allow us to properly address the issues. Again we got a short extension that gave us a better chance of hitting the milestones while still not allowing us to properly address the underlying problems.”
Ellis conceded that the time extensions had caused Ubisoft’s confidence in Free Radical to waver, and that they were only prepared to give the studio subsequent extensions if it agreed to give up creative control of the project to Ubisoft’s in-house producers. Ellis explained that the team could no longer make the game it set out to achieve, and that development was being micromanaged by people less experienced than Free Radical’s coders.
Despite disagreeing with the new change of approach, Free Radical no longer had purchase to say ‘no’, given that it had deadlines to meet and staff to pay. The atmosphere at the studio grew desperate as its employees were forced to work on a game they had lost faith in, assured in the knowledge that it wouldn’t turn out great.
Ellis explained the events that followed, “At some point – I forget exactly when – Ubisoft told us that there was an opportunity for Haze to be a PS3 exclusive, which would mean significant promotion by Sony. I wouldn’t say that this change was a major factor in the quality of the finished game – other events had already seen to that – but it did add a few new hoops that we had to jump through.
“When the game came out it wasn’t what any of us wanted it to be. However I always felt that it was judged fairly harshly. Rather than being judged on it’s merits it seemed to be judged by many based on their expectations of what we were capable of.
“Our first review was from IGN and they gave it something like 4 out of 10. It may have fallen short of our goals, but it’s not a 4 out of 10 game. However, many followed suit and we ended with a Metacritic of 55.”
I have a bad feeling about this
With Haze out in the wild garnering a poor reaction among the gaming press, Free Radical needed a comeback, and TimeSplitters 4 seemed like an obvious choice. In 2007 Ellis commissioned a small team to look at the possibilities of a fourth game, and they set about creating design and pitch documents, teaser trailers and some concept art to show off to publishers. Although spirits were regrouping at Free Radical, publishers simply weren’t interested.
“We put this down to the fact that we were asking for a fair amount of money,” Ellis recalled. “Budgets had grown since the previous console generation, but we only had a PowerPoint presentation to show – there was no prototype at that stage. So, in 2008 we decided to develop a prototype.
“We actually used an amalgamation of the old engine and the new engine, and fairly quickly we were able to put together something that looked great. It ran on PS3 at 60 FPS, hopefully eliminating any perceived technical risk – the experience with Haze had made us extra cautious in this regard – and it played just like the old TimeSplitters games.
“It was a great foundation for building a new game. We showed this to numerous publishers through the second half of 2008 and always hit the same two questions: first, what happened with Haze? They had seen us release a game that didn’t reach our usual standards and that made them nervous.
“Second, the marketing guy in the room would always say something along the lines of ‘I don’t know how to sell this’. The pillars of TimeSplitters are variety, freedom, choice. It is a multiplayer sandbox, first and foremost. The feeling amongst everyone we spoke to was that it isn’t easy to sell that kind of experience. For these reasons, one by one they all declined.”
At the time, Free Radical was working with Lucasarts on two games – one we now know as Star Wars: Battlefront III – and it was a pairing that had the power to save the company. However, the collaboration didn’t bear fruit, and although the history of Free Radical’s Star Wars title has been muddied by conflicting claims on both sites, Ellis was happy to provide his version of events.
“In summary, we were working on 2 games for LucasArts,” Ellis explained. “Things were going so well with the first game that we had agreed to put all of our eggs into one basket and made ourselves completely dependent on them – until or unless TimeSplitters got signed.
“Then various things happened, new people suddenly replaced everyone who we had been working with, their internal development halved in size and we went from two projects to zero projects within a few months. We had over 200 staff and no income.”
Times grew desperate at Free Radical, and the only viable option left to Ellis and his team was to downsize significantly and quickly to curtail any additional loss. However, thanks to UK employment law, the company had to first enter a 90 day consultation period before making staff cuts, and this was time that the business couldn’t afford.
Regardless, Free Radical entered a consultation period as soon as the first LucasArts title was canned, confident that the second project would come to pass and recoup the studio’s losses. It never happened, and with it the demise of the renegade studio loomed precariously on the horizon.
In a separate article, Ellis claimed that Star Wars: Battlefront III was 99.9% complete when it was canned. We’re not here to debate the claim, but if true, it would have come as a killer blow to Free Radical’s workforce. This leaked gameplay footage suggests that Free Radical had invested a great deal of time in it.
Just when things looked their bleakest, Free Radical was handed a lifeline: Activision wanted to discuss GoldenEye.
Four no more
Free Radical was essentially the team that made GoldenEye 64 a reality back at Rare, and showed the world that multiplayer shooters had a place on consoles, even if the rest of the industry had tapped into the trend in a big way. Call of Duty: World at War had just launched that year, showing that the first Modern Warfare was no fluke. Multiplayer shooters were here to stay, so why shouldn’t the team that helped popularise the notion have one last crack at the whip?
“Out of the blue I got a call from Activision asking if we would be interested in making a GoldenEye game for them,” recalled Ellis. “Yes please.” Ellis also receive another offer, “a credible acquisition offer from one of the publishers who we had pitched TimeSplitters to. Both of these interested us – suddenly things looked quite positive again – but they both eventually didn’t happen for reasons that we never discovered.”
With the company’s financial standing dangling by a thread, and its employee’s morale in tatters, Free Radical decided to go back, right back to the beginning, and pitch TimeSplitters 4 to Eidos. The company had helped the studio become a reality back in 1999 – almost ten years earlier – and now it had the power to give it a second wind.
They declined the deal.
“That afternoon I spoke to the lawyers and an insolvency practitioner about putting the company into administration,” Ellis explained, “and how we might be able to at least protect some jobs. We came up with a rough plan, which I had to keep quiet about for a few days while we put things in motion.
“Later that evening I attended the company Christmas party – all arranged and paid for months before, in happier times. Despite the situation, I was surrounded by happy faces, staff enjoying themselves, saying how much they enjoyed working for the company, not knowing what was coming.”
Just after the joyous staff party, Ellis herded 143 employees into a nearby hotel conference room and told them that they were now out of a job. In the next room, the remaining 39 employees – Ellis included – had survived the fallout, but he explained that they would remain employed for an uncertain amount of time while the administrators sought a buyer for Free Radical.
The studio had grown accustomed to depicting the future through its TimeSplitters games. But now, for the first time, their own future was unclear, completely devoid of direction or stability. Anything could happen.
Friends in the East
Christmas and New Year were far from the enjoyable holidays they should have been for Free Radical and the recently laid-off employees. The dream of splintering off from Rare, and to carve out an independent legacy had failed, shot down by publisher intervention, crushing deadlines and the difficult process of adjusting to PS3 technology. Shortly after the holiday season, Ellis received an email from Crytek CEO Cevat Yerli. He wanted to buy what remained of Free Radical.
“We spoke on the phone,” Ellis explained. “[Crytek] spoke to the administrators and they quickly visited to discuss it further. They believed that we had similar culture and abilities, and that our console experience might help them make their transition from PC to console.
“It sounded like an amazing possibility but a deal had to be done very quickly – the administrators could only continue to operate the company while they believed that it might achieve a better overall return for the creditors – the redundant staff, the landlords and so on. The administrators set a deadline and received a number of bids from interested parties and ultimately Crytek was the winning bidder.”
This 11th hour purchase kept the remaining Free Radical staff in jobs as a new unit called Crytek UK, and it secured their future to this day. However, Ellis didn’t join them. He resigned, opting to stay independent and to form a new studio instead. Looking back at the deal, Ellis is thankful for Crytek’s intervention, so that the studio could live on in some form or other.
“I was obviously pleased that Crytek were successful in their bid,” Ellis stated. “Becoming part of Crytek gave the remaining staff the best opportunities to work on the kind of games that they wanted to be making, and we were able to immediately set about re-hiring some of the recently-redundant staff. For me though, I had done what I could to protect peoples jobs and it was time to move on.
“I didn’t have much interest in becoming an employee again. The culture of Crytek would require a mentality that didn’t come naturally to me. I don’t like team-building exercises and I would find it hard to refer to my colleagues as ‘heroes’ without feeling like I was patronising them. It wasn’t a good fit for me.
“But mainly, the past year had taken its toll and I was ready to do something different, something that involved less people, less gambling other people’s money and jobs. So once the handover was complete, I resigned. If they were going to go forward under new management, they might as well do it from the outset.”
Ellis is now founder and managing director at Crash Lab, an independent mobile studio that just earned over one million downloads of The Snowman and the Snowdog, a tie-in with Channel 4’s festive animated film. It hit the top of the App Store charts and served as a debut victory for his new team.
He is indicative of the industry today, where veterans such as Cliff Bleszinski and Peter Molyneux leave their positions at triple-a studios to fulfil personal goals or realise pet projects on their own terms, away from the marketing and hype of the blockbuster circuit. The tools are there to turn dreams to reality, and depending on who you ask, we are now in a golden age of development.
Regardless, we at VG247 have seen many public campaigns to see TimeSplitters resurrected at Crytek UK, and at a time where Kickstarter makes fan-requested projects a reality, it seems as if TimeSplitters 4 is a no-brainer, but even Cevat Yerli feels wary that the game would under-perform. Maybe its time has come and gone, and that it’s best left to memory?
“I’ve had a lot of time to reflect since 2008”, Ellis mused, “and in many respects I think that the publishers were probably right. When you look at games that succeed commercially and games that don’t, aside from the perpetually successful games with annual updates – like CoD or FIFA – the marketing message is usually a critical factor. It’s not just about making a great game.
“We, and the publishers we spoke to, weren’t able to find a clear, compelling marketing message to communicate what TimeSplitters 4 is about. Without this, it probably would have sold mostly to the people who already liked TimeSplitters. We might have reached a million again, but not many more.
“However, due to the costs of developing a next-gen shooter, it would need more than the 2 million people who bought TimeSplitters 2 to buy it in order for it to recoup its costs. That wasn’t likely to happen, so they were probably right not to do it.”
Ellis concedes that the final decision on the fate of TimeSplitters rests with Crytek, which now owns the license, but stressed again that there simply aren’t enough fans out there to warrant a revival in this cash-critical age, where investment must be returned ten-fold to convince publishers that an IP is worth supporting. Ellis added that Kickstarter is out of the question too, as he explained that modern FPS titles now cost £10 million upward to develop. That’s a lot of charity.
“Looking back it’s hard to say what we should have done differently,” mused Ellis. “Again, I’ve had a lot of time to reflect since 2008 and haven’t really come up with any definite answers. Maybe we shouldn’t have taken on the second LucasArts project. We probably would have still lost the first one, but with a smaller overall financial impact that we could perhaps have managed.
“Maybe we shouldn’t have pursued TimeSplitters 4 for as long as we did – but it wasn’t a big team so we didn’t spend huge amounts of money on it and it couldn’t have made much difference. Maybe we shouldn’t have thrown away the old [C] engine and embarked on a rewrite – but maybe then we would have suffered all of the problems that we were trying to avoid with the rewrite.
“Maybe we shouldn’t have attempted to create the new AI system for Haze – but how would anything ever move forward if companies didn’t try to innovate? Maybe TimeSplitters 3 would have been more successful if we had stuck with Eidos rather than switching to EA, and perhaps then we could have had a better chance of signing TS4 – but then Eidos had their own problems through 2004, so they probably wouldn’t have been able to afford to do a better publishing job than EA did.
“Maybe we should have diversified sooner, to be less reliant on the risky and expensive business of developing shooters – but that’s easier said than done. It’s hard to convince a publisher to fund a project in a different genre than the one that you’re best known for.
“Who knows? It’s easy to come up with a hundred different things that we could have done differently but it’s hard to be sure that any of them might have made a difference.”
You can follow Crash Lab and Steve’s new projects over at the developer’s official site.