Forza 5: How McLaren Automotive's data served Turn 10

Forza 5: How McLaren Automotive's data served Turn 10
Will Freeman

By Will Freeman

November 27th 2013 at 12:45PM

A look at what a Formula 1 team with a sold out £1 million sports car has to offer Turn 10’s developers

With Forza Motorsport 5 off the grid, Develop caught up with developer Turn 10’s creative director Dan Greenawalt and McLaren Automotive’s P1 programme director Paul MacKenzie, to find out what a partnership between a first-party studio and a luxury sports car brand really means beyond marketing.

How did Turn 10 and McLaren Automotive come to work together on this project?
MacKenzie: Our relationship with the games dates back further than Forza Motorsport 5. But when Forza 5’s team were looking for the vehicle that could help launch their game, everything fell into place. We were developing our P1 super car and its technology just as Turn 10 were looking for a vehicle to launch their game.

It all just came together at the right time. That allowed for a real coming together of a next generation game and a next generation vehicle. It was really fortuitous, and there was a great parity of these technologies advancing.

Greenawalt: And recently things have really accelerated as well, in the advancement of real car physics, through things like their geometry and the way aerodynamics are progressing. The last five years have seen a lot of changes there, so Turn 10 needed partners that could help us recreate those advances from the cutting edge in simulation.

Of course, having the P1 helps us make a big splash, and McLaren’s car is perfect for that, but on the engineering side we need to push our simulation, and we need companies that are on the cutting edge there.

Yes, we did work with McLaren in the past, which helped. Turn 10, as part of Microsoft, were able to kind of get in the door with the partnership McLaren had to do the electronics for Formula 1 cars.

That allowed us very early access to McLaren – which is a very multifaceted organisation –  on the OEM side of things, so original equipment manufacture and the electronics side of things. That allowed us to swap notes with their engineers and such. This was many, many years ago, back during the SLR timeframe [around 2006/7]. So, for a long time, and culturally, there’s been a level of alignment between the two companies.

Together we’ve really tried to be on the cutting edge of design, and with the P1, that became really important. The P1 is a once-in-a-decade car that came when we had a need for technology and a great partner. McLaren could provide that.



How exactly did McLaren help with that need for technology? How did you help each other there?
Greenawalt: We need raw data, and that’s what it comes down to. We’re not going to advise McLaren on the best way to make a supercar. They are the best at what they do, and we are the best at what we do.

But we can very much use their data. There’s really very complex geometry and movement to a car, especially at high speed, and with things like the way the air moves through the different facets of a car as it heads forward. Similarly, the way the suspension behaves is incredibly complex. That means we’re very keen to get measurements; the rawest of raw data. Things like the weight of components and the angle they might sit become something we can programme into the game

Does that come from the likes of wind tunnel tests and McLaren’s own testing? How much effort is needed to make that data suitable for a game?
Greenawalt: Our job is really about recreating, is what it comes down to. So when McLaren might have a complex fluid dynamics model for simulation use, they are handling test software, but rarely does that run in realtime, as it’s typically on the server. We, meanwhile, do things in realtime, so that definition changes what is being computed.

But what’s misunderstood is this. People assume we would take test data such as 0-60 MPH data or top speed and program that in. The truth is that we don’t programme any of that, because it is physics that is recreating it.

What we need to do is look at things like the weight of the car as components, the weight of the doors, the drive line and so on. So that recreates the weight of the car. That weight of the car, combined with friction, horse speed and everything else is what gives our in-game P1 the 0-60 MPH speed it has.

We have to break everything down to its smallest unit, and that is as true of our simulation as it is their simulation. Now, the Xbox One is incredibly powerful, and much more powerful than the Xbox 360. That means that Turn 10 is now able to do things in real time that could not have been done in realtime last generation.

But, still, when trying to build such a detailed model of a car building together every element, building it offline makes much more sense. So there’s different types of simulation for different types of application, and our job is moving data between those, and ultimately building a realtime simulation in game.

It sounds like your targeting a level of detail and realism almost no player would be able to consciously recognise. What justifies that effort of making sure a car door’s weight is correct in-game?
Greenawalt: Having been with Forza for 10 years, I can tell you we don’t want to compromise on quality. And we’ve also always believed Forza should fit a very diverse range of players, and fit them all like a glove. In order to do that we want to let each player play the game the way they want to play it.

I admit I’ve never been totally sure just how far we can take that, but I’ve always been sure that through elegant design we can deliver a game that is both a simulator with no compromise, and something playable by a five-year-old.

Many games do that by trying to create something positioned between the extremes of their target audience. That’s not our goal. We build a simulation and add layers of assists, both real world and more fictitious, that can be stripped off and added, depending on what the player wants from a game.

The simulation to me needs to be cutting edge enough for that small percentage of players that are real professional racing drivers to use the game for practice. Stefan Sarazen, for example, used to race with Peugeot, and had never raced Atlanta. So he came to our studio for two days, practiced on our Forza simulator and got pole position when it came to the real race. And yet I have five-year-old twin boys who really can play the game and have a good time.

To allow us to do that, we need that level of detail, and access to real data, so we can cover the full spectrum.



And what does McLaren gain from this relationship, beyond the branding and marketing opportunity?
MacKenzie: There’s a lot of synergy between the two companies in the way that we’re developing innovative and advanced systems. And actually, there’s even been discussions between the two companies that have made us all realise how similar the processes are of developing a game and developing a car. The way we test and simulate at McLaren is actually very similar, as is the way we do the styling. So that has certainly been interesting and informative.

But for McLaren, what else do we want to get out of it? McLaren is known as a Formula 1 racing team; that’s what the world knows us for. We did the F1 car back in the 1990s, and projects like the SLR. But really we are a relatively new automotive brand in terms of cars like the P1, MP4-12C and the 12C Spider. For us, getting that automotive brand out there is really important.

McLaren obviously makes expensive cars, so we are out of the reach of a very large part of the population. We understand that, but it’s so important for us to get access to all those people. For people to be able to drive a P1 or 12C is amazing, and the game can do that in an incredibly accurate way. We give over the raw data, so Turn 10 can put every detail like the throttle map into the game. It’s a way for people to try our cars, and as such it opens up McLaren to a far wider audience. That helps us build the brand, and a game like Forza 5 can help us let people know that we are not just a Formula 1 team.

Also, the average age of a gamer today is similar to the ages of people buying our cars, so we’ve lots to gain from the partnership.

Has that emerging convergence taught Turn 10 much as a studio?
Greenawalt: It’s amazing how similar our industries’ approaches are. For example, when McLaren kicked off the design process, I was expecting something very different from what we do. I don’t know why, but I assumed the creative process would be very different, but it isn’t.

Both are about collaboration, dreaming big and setting long-term visions and long-term goals. It’s a very similar process with both cars and games, and I really enjoyed seeing like that. So as Turn 10 began working with McLaren it became apparent is was less about learning and more a bout a meeting of the minds. There’s so much alike, even down to the way you hire and attract talent, and how a business plan effects what you do and don’t do.

There’s also that culture of perfectionism that’s prevalent at Turn 10 as a first-party studio, and at McLaren as a company making very special cars.

And in delivering a simulator that respects the relatively restrictive conventions of the driving genre, you’ve still managed to innovate?
Yes. And we always have. In setting out to create these games, I was never aiming for certain results from a business plan. I wanted to make a difference in – and connection between – gamer culture and car culture, and that has meant embracing new trends in game and car design.

Our very first Forza had learning AI, called the Drivatar system, developed through Microsoft Research in Cambridge. That’s the sort of technology you don’t normally get as a game company, and it lets us bring in innovation to the driving genre. We also innovated in the series with our livery editors, which worked more like Photoshop than most other in-game paint tools of the time. If imitation is the sincerest form of flattery, we feel very flattered over the years, because our livery editor has made its way into not just racing games, but games of all sorts. And that was Forza 1.

In Forza 2 we started to introduce our work with tyre manufacturers to better implement tyre simulation, and we’ve improved on that with every iteration. Forza 3 introduced more, innovative assists that made the game available to a far wider audience, and we saw our audience change, or rather diversify, as a result.

Sure, it’s hard to innovate the pure ‘car on track’ form of the game, but if our innovations are bringing in more new players and getting adopted elsewhere, we must be doing something right. For example, the dynamic green-line in driving games came from Forza. There’s even things that have been adopted by other games already that I wanted to fix in Forza before they got adopted.

And in Forza 5, Drivatar’s cloud-based, big data-led approach is just a totally new way of looking at and using AI. Having real players teach the AI down to the smallest driving technique is really special; it can learn human decisions that we humans don’t even know we are making. That’s an innovation we feel really impacts the game players can experience.