Chris Sawyer (Creator of RollerCoaster Tycoon) | Making of Documentary

In 1999, a wealth of critically-acclaimed PC games hit store shelves, including Unreal Tournament, Age of Empires II, Homeworld, Sid Meier's Alpha Centauri, and System Shock 2 - just to name a few.

Yet, among these heavy hitters, was a seemingly unremarkable theme park simulation game called RollerCoaster Tycoon that many in the industry believed would have very little appeal, only for it to defy all expectations by becoming one of the best-selling PC games of that year.

Its sole developer, known already by some for his previous work (Transport Tycoon), suddenly went from being relatively unknown to an overnight sensation! But just as quickly as his name appeared, it soon faded into obscurity, as several years later he decided to leave the video games industry behind to pursue other interests.

A game developer who is as renowned as he is elusive and who's legacy has left a profound impact on tens of thousands of players around the world, many of whom still enjoy playing his games to this day. A game developer called Chris Sawyer.

Chris Sawyer is a Scottish video game designer and programmer born on the 27th October 1967. Sawyer was introduced to video games as a child, when he played Pong at a neighbour's house - though he wasn't particularly excited by it.

Years later, as a young teenager, he'd write simple programs in BASIC using a Sinclair ZX81 demonstration unit they had on display in a WHSmith's store. There were other guys who did the same thing, and he stood in awe as they made big flashy objects move around the screen, while all he could do was print the word "Hello".

During his lunch break, he'd program in BASIC on his school's Commodore Pet 2001 and Research Machines 380Z computers just for fun, and then dabbled in a little machine code on the Pet's 6502 processor. These programming sessions often resulted in games of some description, and despite their simplicity they were fun to play and were great for showing off to his friends.

In the early 80s, he'd saved up enough money to buy his first home computer - a Camputers Lynx. It seemed like an odd choice at the time, given that his friends all had ZX Spectrums and VIC-20s, but the Lynx had a couple of standout features; namely its proper keyboard and its use of multi-colour bitmapped graphics instead of text-based pattern tables.

Popular Computing Weekly magazine was a regular read, and he'd excitedly study it from cover to cover, taking in all the technical details of the new hardware and software, whilst trying to understand the type-in program listings they published. Many happy days were spent engrossed playing games and programming the Lynx, but it wasn't long until a new home computer came to market that caught his interest; a Memotech MTX500 which he also purchased.

The MTX500 was notable because it had a built-in Z80 assembler which meant that, unlike on the Lynx where he'd hand-coded some small machine code programs byte-by-byte, he could now experiment more easily with Z80 assembly code. This became the start of his games programming career.

He continued writing simple games throughout the four years he spent at the University of Strathclyde in Glasgow, where he studied Computer Science & Microprocessor Systems. In the mid-80s, while still attending school, he sent a few cassette tapes of his MTX games to Memotech themselves and received a call from Jim Wills who invited him down to Witney to meet the company and discuss publishing his games through their own label Continental Software.

Wills eventually left the company to start his own publishing label, Megastar Games, and published several more of Sawyer's MTX games over the next year or so. His MTX game sales were never all that impressive, and despite having nearly a dozen of them published, the revenue was only enough to buy a couple of blank cassette tapes, his first printer and a floppy disk drive.

Eventually the Memotech MTX500 gave way to an Amstrad CPC 6128 and during his final year at university he bought an Amstrad PC1512 which became the start of his very long relationship with PC programming. Some of the last games he made on these early home computers featured an isometric perspective which were very challenging to program on such simple machines but were very satisfying once he'd gotten them working. This isometric viewpoint became a fascination for Sawyer and would become extremely influential later on.

During this time, he'd continued to have the odd game published for the Amstrad CPC, and had even designed and programmed a PC game as part of his university course's work experience requirement. The PC game was created for Ariolasoft UK, who never ended up publishing the game - presumably because it wasn't very good. They did however promise Sawyer a job once he'd graduated; working from home and programming games for them.

So as the end of his time at university approached, he didn't bother applying for any of the jobs his classmates were applying for; believing he'd have a well-paid job lined up in the games industry. However, after graduating in 1988, he discovered that Ariolasoft was struggling to stay afloat and was in the process of downsizing, meaning the job offer amounted to nothing, and he'd now missed the boat on applying for a job in the electronics industry as they'd all been taken by his fellow classmates.

It was only through recalling a conversation he'd had with another programmer a year earlier that he ended up cold-calling Jacqui Lyons at Marjacq Micro, who he'd heard was a business agent representing some of the most well-known UK games programmers at that time. Upon calling her, he was surprised to learn that she'd already heard of him. She took Sawyer on as a client and soon had him signed up with Logotron to do the first of many DOS conversions of Commodore Amiga games to PC - StarRay (also known as Revenge of Defender in North America).

He thrived on the work and loved the challenge of trying to get the PC to do anything like what the Amiga was capable of. Many more DOS conversions followed, with games like Virus, Xenomorph, Conqueror, Elite Plus, Birds Of Prey, and Campaign but, while he enjoyed the work, he would often be making his conversions in parallel with the original Amiga game's development and would find himself quickly catching up - only to then have no work while he waited for them to get ahead again. He solved this particular problem in 1993 with his final two conversions, Dino Dini's Goal and Frontier: Elite II, by doing them both at the same time.

All of these conversions were enjoyable and satisfying to work on, but by this point he was more interested in working on a new project he'd been tinkering with - an Interactive Transport Simulation game that would come to be known as Transport Tycoon. While finishing his remaining two conversions, Sawyer had also been developing a 3D isometric game world system, and having fun playing Sid Meier's Railroad Tycoon in his spare time. He started to wonder whether his isometric viewpoint would be better than Railroad Tycoon's overhead 2D map and if other modes of transport should be included. The two interests collided and he soon had some little trains running around his isometric world. From there, the concept for "Chris Sawyer's Transport Game" emerged. Other games such as A-Train, SimCity 2000 and Theme Park also likely influenced the way the game was developed.

By the time his remaining conversion contracts were complete, towards the end of 1993, he already had a fully-functioning version of the game up and running. The game had grown to include different modes of transport, a large world to play in and a simple financial mechanic for earning money. By March 1994 he had a roughly ninety-percent complete game and it was then that his business agent, Jacqui Lyons, began approaching publishers.

Just as they were about to sign with one publisher, MicroProse extended an invitation to Sawyer to come visit their offices and staff in Chipping Sodbury. He'd previously done work for MicroProse as a work-for-hire programmer when he did the conversion of Elite Plus, and after meeting with them he realised that they were the right choice. Having published Railroad Tycoon, MicroProse felt it appropriate to call Sawyer's game Transport Tycoon.

They also suggested a number of changes to the game, but few actually made it in. Sawyer had a very clear vision for what he wanted the game to be, and he also had a good handle on what was possible in the game, so although some of their ideas sounded good he knew that they weren't going to be feasible. He did put in their suggested references to other MicroProse games though, such as Subwar 2050 and UFO: Enemy Unknown (also known as X-COM: UFO Defense in North America) and they also helped him tie up a few loose ends, like what should happen when you beat the game.

After signing with MicroProse, Sawyer was assigned a producer and was tasked to create a project schedule to show what still needed to be done and how long each task would take. After adding up the estimates, it came to just over two years of work so he divided all of the estimates by four to keep the producer happy. In the end, he managed to beat all of those estimates anyway and they had the game ready to publish within four to five months.

Right from the start, Sawyer intended Transport Tycoon to span at least one hundred years, so he knew that the pace of the game would need to be carefully controlled to prevent the growth of towns and companies from getting out of hand. He also considered taking the game further into the 21st century, and having technologies such as matter transportation and other instant teleportation systems (like something you'd find in Star Trek) but he realised that this would make the endgame rather anticlimactic, as player's would suddenly find that all of their painstakingly constructed railways and roads were made redundant once the matter transport system was invented.

He also originally left out a lot of the complicated finance options present in Railroad Tycoon such as stocks and shares. He wanted Transport Tycoon to be as simple as possible to play, and believed that these mechanics would make the game unnecessarily complex. He did have ambitions to put in things like underhanded tactics, more detailed share dealing, and corporate takeovers but a lot of these features were dropped because, while he'd thought them through and could have added them, he felt like they just weren't going to work well enough in the game. Transport Tycoon also didn't allow players to alter the game speed because the complexity and detail of the simulation was already running flat-out, and there just wasn't any way to make it go faster.

Sawyer was only able to push the game to its limits because he wrote it in x86 assembly - a low-level programming language where you write the actual instructions that the CPU is doing and as such you can optimise it for efficiency in a way that higher-level languages like C can't because they need to be compiled into a lower-level language first. To put it another way, it's like the difference between speaking directly to another person vs. having to go through a translator first - the latter being much slower and far less efficient. At the time, home computers didn't have particularly fast processors so using assembly code was a necessity for Sawyer to be able to pull off what he wanted, as much of the simulation side of the game involved a lot of processor workload and needed to be highly optimised, otherwise the game would have been brought to a crawl - even on the fastest processors of that era.

By far the most difficult part of developing Transport Tycoon was designing the game's AI, as it involved looking in detail at how a human player builds railways, roads and so forth and how they make decisions for their transport company, and then trying to simulate the same decision processes for the computer opponents. The AI was far from satisfactory, with it often building very obscure railway lines or roads, but in order to make it more intelligent it would have required the PC to stop to think for several minutes every time a section of railway track or road was built, which simply wasn't practical for a real-time game.

Early test builds of Transport Tycoon were depicted in low-res display modes with Sawyer's own crude blocky graphics. These were later replaced with high-resolution bitmaps created by Simon Foster who Sawyer had the good fortune of crossing paths with. Foster happened to be doing the graphics for an Amiga game called Toado by Jim Wills, who had previously helped Sawyer publish his old MTX games. He recommended Foster to him, so Sawyer got in touch and paid Foster to come up to Scotland where he spent an afternoon looking at the game, and they agreed there and then that he'd start supplying the graphics for Transport Tycoon. Foster had a good understanding of what Sawyer was after, and he very quickly developed a style that turned out to be exactly what the game needed.

The two worked from home in offset hours, with Sawyer starting his day at seven in the morning and typically stopping in the middle of the afternoon, while Foster got up at noon and worked into the evening - so they only had a couple of hours in the middle of the day where they could actually talk to each other. They kept each other informed of what they were doing via telephone calls and by sending floppy disks in the post. Foster would work all day on the graphics, get a floppy disk in the mail by 5:30PM, and then Sawyer would receive it the following morning at around 8:30AM. This was a fairly common way to send data back then, as the Internet was still relatively immature, so everything ended up being sent by floppy disks. Despite sounding dysfunctional it actually worked quite well.

Many of the buildings in the game were influenced by real-world locations in Glasgow, as Sawyer had sent Foster several photographs of the area. These included; "Livingstone Tower" - a part of the University of Strathclyde that Sawyer attended, a historic building on 260 Clyde Street - the graphic in-game was modified but some of the features of the real building remained, houses situated near the Buchanan Street bus station along with the Clydesdale Bank building - the Transport Tycoon version gained an extra floor but was otherwise very similar, and lastly the shopping centre was very loosely based on the St. Enoch shopping complex in central Glasgow.

The game's soundtrack was composed by John Broomhall, who was the in-house musician at MicroProse. The composition began with the title theme. As he sat down to play the game, it didn't take long before his brain locked onto a tempo which just felt right. He thought a jazzy, funky vibe would be a good fit for the game as it conveyed a sense of time passing, travel and movement. Everybody in the office loved it. They then ran it by Sawyer and he liked it too! The positive reception to the game's title theme is what sent Broomhall down the path of jazz, but technological constraints really inhibited the creative process.

Writing the music seemed like the easy part, when compared to the job of trying to make it sound good on the midi soundcards of the day. In order to get the soundcards to do anything useful in DOS required the audio programming skills of Andrew Parton, with whom he was lucky enough to work with at MicroProse. While General Midi soundcards were supposed to have a universal instrument set so that musicians could guarantee their music would sound right on any given board, the reality was none of them actually sounded completely alike - so the 'universal standard' was something of a joke to the discerning ear.

When making each composition, he'd start out with the humble FM cards and once he'd gotten them sounding as good as possible, he then set about upgrading them for the Roland LAPC-1 and SCC-1. Because other GM soundcards of the day from Turtle Beach to Gravis Ultrasound were all slightly different, he tweaked and finessed a separate version of the midi files for each and every one of the main players, as he was determined to have the music sound good no matter the hardware being used.

Transport Tycoon was released in November 1994 and was an immediate and unqualified success for MicroProse. The game received great reviews and sold well in both the UK and Europe, though it did struggle in North America (likely due to the country still largely being console-focussed). A "World Editor" expansion pack followed shortly thereafter in mid-1995 which added a scenario editor and an alternative Martian graphics set. This came about after Sawyer and Foster brainstormed some ideas for what they could do and what would look good in the game. They wanted to add something extra to go with the World Editor and decided that the alien atmosphere that the Martian graphics presented would make the expansion just that bit more fun.

This was then followed by a Deluxe Edition of the game later that same year which featured some of the ideas and improvements that didn't make it into the original game, such as the ability to buy and sell shares, take-over rival companies, improved railway signals for more complex train systems, network play, and the world editor that had previously released separately, among other changes.

Following the release of Transport Tycoon Deluxe in 1995, Sawyer was commissioned by MicroProse to make a sequel. He worked on Transport Tycoon 2 on and off for several months, with the major focus being on re-writing the 3D isometric game engine and world data handling so that railway tracks and roads were no longer restricted to running at ground level. He reached a point where he had multi-level bridges and little block-shaped trains merrily running under and over them while depth-sorting correctly, but he soon started to lose confidence in the game as he didn't feel like it was enough of an evolution over the original to be worthwhile. This was further compounded by a lack of new features to add to the game as well as the trouble he was having adapting to Windows and DirectX development after years of working under DOS.

Needing a change of pace, he visited a couple of small theme parks with his friends in 1996 and started becoming interested in rollercoasters; specifically the engineering and aesthetics of them. He'd also been reading through a book called "White Knuckle Ride" by Mark Wyatt which featured photos of huge rollercoasters from around the world. Pages 44 and 45 were two that were engraved in his mind, as they showed the coasters "Twister II" at Elitch Gardens with its incredible wooden supports and "Vortex" at Paramount's Kings Island with its impressive steel loops.

So in September 1996, he put Transport Tycoon 2 on hold and tried adapting his isometric system to see if it could handle a simple rollercoaster. This marked the beginning of "Project White Knuckle" (the working title for RollerCoaster Tycoon) and within a month or two he had some basic rollercoasters up and running in the game. Unsurprisingly, the first two coaster types he created were the wooden coaster and a looping steel coaster that were modelled after Twister II and Vortex respectively. From there, the game gradually evolved from a simple construction set to a fully-fledged park simulation, with little guests riding your coasters and making your park a success or failure.

Sawyer bought himself out of the Transport Tycoon 2 contract in order to commit to working on RCT full-time. He didn't originally intend for RCT to be a commercial project, and was in fact told by those in the industry that if he tried to publish it, the game would have very limited appeal - perhaps only amongst rollercoaster enthusiasts. But Sawyer paid no notice, instead deciding to fund the game's development himself and keeping the project under wraps so that he wasn't distracted by people's negative opinions on the game's design. It was also much easier to sell a concept to a publisher when you could sit them down in front of a near-complete game and get them to play it.

With Transport Tycoon Sawyer had a clear vision of what the game was going to be and what it would end up like, but that wasn't the case with RCT. Once he'd proven that his isometric system could cope with a few simple rollercoasters, he moved onto getting guests working - as a rollercoaster is only as good as the enjoyment it brings to its occupants. Thus, the fundamental mechanics of the game were born; design and build good rides which guests enjoy, keep them happy and make money.

Having the guests be uncontrollable appealed to Sawyer a lot. He liked that they made up their own minds based on what they saw and it was satisfying to see the results of his primitive algorithm looking quite 'human'. Much like Transport Tycoon before it, RCT was written almost entirely in x86 assembly code; with a small amount of C to interface with Windows and DirectX. He rigged up two computers; a fast one for coding, and a slower one for testing. New features put a greater burden on the slower PC, so he needed to keep the game's scope in check.

By far the biggest headache of programming RCT was its pathfinding. It was easy to program a route-searching algorithm that worked perfectly, but it was of no use if it stalled the game for seconds or minutes at a time whenever it needed to make a decision. Sawyer chipped away at the algorithms, stranding many guests in the progress. While testers later criticised the guests' ability to get lost far too easily, if you built an oddly-shaped park, Sawyer recalled getting lost himself in exactly the same circumstances in real theme parks, so he thought it was probably right that the guests in RCT also struggled if the park layout was unintuitive.

The game's toolset grew organically from the things that interested him personally, as well as what was technically possible with PCs at that time. For example, the limited piece-by-piece rollercoaster construction method, with fixed angles and sizes, came about as a result of the limited power and memory available. These limitations were also responsible for one of the game's most iconic features; the ability to torment and kill guests. Sawyer never set out to deliberately allow such violence in his game but, with the drowning of guests, it was just simpler than the alternative of having guests swimming to shore and climbing out of the water. It would have involved more programming and graphics than a simple animation of a guest waving for help and then disappearing.

During the beginning of the game's development, Sawyer spent a lot of time researching rollercoasters. He bought every book and magazine he could find on the subject, joined both the "Roller Coaster Club of Great Britain" and the "European Coaster Club", and even visited places like Blackpool Pleasure Beach, Lightwater Valley and the now-closed Loudoun Castle just to take photos and ride the rollercoasters. While he would have liked to have visited more amusement parks, he simply didn't have the time and he'd collected enough technical information about the operation of the different types of attractions, from multiple park managers, to make the game - so he felt that he had sufficient research. Doing all of this research on his own also meant he could minimise the number of people who knew about RCT during its development.

As he continued working on the game, his passion for rollercoasters steadily grew. What had started out as a mere fascination with the engineering, design and aesthetics of rollercoasters eventually developed into an overall enjoyment and appreciation for them. When the game was already far enough along, Sawyer contacted John Wardley - who has been responsible for the design and construction of rollercoasters and other attractions throughout Europe. His experience and consultation proved invaluable to the project and only further deepened Sawyer's appreciation for rollercoasters.

While all of this research helped to make RCT more realistic, Sawyer always made sure to prioritise fun over realism when given the choice. If you take realism to its logical conclusion you get a simulation of real life, which is often boring, frustrating, time-consuming and not very rewarding. He understood that realism was important, in so far as giving the player familiar rules and recognisable feedback, but that the key to making a good game was having plenty of elements that were just plain good fun.

RCT underwent numerous changes throughout its two and a half years of development. Very early on rollercoaster track was made up entirely of straight pieces at various angles that could be plugged into each other to create hills and dips, however it didn't look very good and so this idea was abandoned in favour of smooth slope transitions. This was also around the time when the first batch of thousands of different angles for the rollercoaster cars were created. Depending on the type of track, each car could have up to 800 different frames of animation, plus a further 1600 frames for the passengers.

By March 1997 the state of the game looked remarkably cohesive given the relatively small amount of time Sawyer had worked on it. A variety of different rollercoaster types were implemented, along with some of the control schemes and scenery. Soon after, he began experimenting with different styles for the game's UI. RCT originally featured a wooden look-and-feel with the buttons at the top gently swinging back and forth while being selected or clicked. However, while it looked great, it had too many drawbacks and before long the game adopted a much cleaner and functional style which allowed for more information to be displayed clearly in more compact windows.

Another idea that was later dropped was the concept of rusty track that would appear as the rides aged, with the nice bright paint finish peeling off the steel track to reveal rusty and bare metal that would need to be re-painted. In the final game, the rides still lost value as they aged but there was no visual indicator to show this. RCT was also going to feature destroyed pieces of track whenever a coaster crashed however this idea didn't make it into the final game either.

A number of other ideas were dropped for one reason or another including: different styled entrance buildings and pavement, some strange tunnel graphics that looked more like drainage pipes, sculpted peacock trees, an odd panda object (the original purpose of which has been forgotten), a military tank for the Car Ride, a car model intended to be used for public roads and car parks, a large multi-lane slide ride that would have consisted of several sections pieced together, a robot entertainer costume, several scenery pieces and themed buildings, and a bus that was intended to bring guests to your park in some scenarios.

Simon Foster returned to do the game's graphics, with the music being composed by Allister Brimble. Sawyer felt the music Brimble created suited the game well, though he was adamant that the style of the fairground organ music for the Merry-Go-Round sounded a certain way, and wasn't happy with any of Brimble's attempts at synthesised versions. In the end, they obtained permission to use a 1970's recording of a real fairground organ which sounded great.

The look and feel of RCT was an important aspect of the game that Sawyer wanted to get right. The game's isometric perspective created a solid, accurate and consistent view that ensured objects always looked the same, with the same amount of detail, no matter where you were looking from. This consistency was also reflected within the game's UI. Instead of using a series of photographs or high-definition renders of rides when you select a new ride to build you're instead presented with little screenshots from the game. It was important to Sawyer that you could recognise the rollercoasters straight away, and that you could re-create all of them that actually existed. However he was also clear to Foster that the game should have a visual aspect closer to cartoons or comics than any other style, while making sure that the trains, structures and attractions were recognisable as real and feasible. He didn't think a game like RCT should be based on a photorealistic art style. Despite fully 3D games starting to become the norm, he knew that a projected 3D view wouldn't be able to cope with the level of detail he'd designed the game to handle. While it certainly would have been cool to see things from the guests' perspective as they rode you rollercoasters, it would have also been extremely difficult to pull off with the current-day technology.

Sawyer worked relentlessly on RCT, working up to 16 hours a day, seven days a week. It was a gruelling workload, though at the time he thoroughly enjoyed every minute of it. RollerCoaster Tycoon was released in March 1999 to critical and commercial acclaim. What Sawyer found most satisfying was hearing that the game appealed to such a wide audience; not just boys who liked strategy games, but also girls, younger children, adults and even grandparents too. Some were fond of the strategic challenges, some just liked building things, and others enjoyed interacting and looking after the guests. It was vindicating, after being repeatedly told by games industry people that there was no market for RCT, for it to become the overall No.1 best-selling PC game in the US in 1999, and by 2002 the game had sold four million copies worldwide in total.

However just a few days after the game had been released in the US, Sawyer awoke to discover that something had gone catastrophically wrong with the game. In the early morning of Sunday 4th April 1999, he was browsing various online forums when he noticed that players were complaining that the game had somehow lost their progress and sent them back to square one in the game's scenarios. Daylight savings had just taken effect in the US and it seemed too coincidental to be nothing. Within a couple of hours, he'd uncovered the connection. The adjusted time-stamps on the saved game files, which were there to safeguard against tampering or corruption, weren't matching up. The system call he used should have yielded the time in UTC, unchanged by daylight savings and thus creating no problem. Only for some reason, it did change. The issue appeared to be with Windows itself, but it was easier for Sawyer to patch RCT than argue with Microsoft, so he quickly coded a fix and later a utility to patch the broken saved game files.

Sawyer was also in the fortunate position to own the RollerCoaster Tycoon IP and trademark. Since Sawyer had created most of the game completely independently, it made it easier to hold onto the IP as there was no question that it was his game - rather than being funded or instigated by a publisher. He was also lucky to have his business agent Jacqui Lyons handle the negotiations, as she always strived to ensure her clients held onto as many rights as possible. Hasbro Interactive also likely didn't realise the potential long-term value of the IP at the time, as it seemed like a risky game with a niche audience.

With RCT's commercial success, an expansion pack was inevitable and so Added Attractions (also known as Corkscrew Follies in North America) was greenlit and released in November 1999. Sawyer hadn't designed RCT with an expansion pack in mind however once the pressure of completing the game was over, he found himself still inspired by the game and before long he'd written a long list of improvements that could be added without compromising the gameplay or compatibility. His philosophy behind Added Attractions was "more of the same" rather than changing the game in any major way. This meant more rollercoasters (including incarnations of real world B&M coasters), more rides, more scenery and more scenarios. Existing rides also saw improvements with new research options, such as the 'turnaround' section for the Log Flume ride, which allowed boats to be turned around mid-course and travel backwards, along with the 'vintage cars' which could be researched for the Car Ride.

There were also a handful of ideas that were cut from the original game that made their way into the expansion pack - namely the Mini Golf ride - a customisable miniature golf course, with guests going through the course taking shots. Mini Golf never made it into the original game because of the amount of work involved in getting it running. The construction, animation and operation of the ride all created their own problems, so it never went beyond some initial graphical tests in the original game. By the time Added Attractions was in development however, Sawyer had worked out a simpler way of making the ride work, and it seemed like a great addition to the other rides in the expansion pack.

A second expansion pack called Loopy Landscapes released a year later in September 2000 and was very much a snap decision by Sawyer due to problems he was having with his second attempt at a Transport Tycoon sequel. A month or two into development he was getting nowhere with the project and so he decided to take another look at RCT. Within 24 hours of making that decision, Hasbro had accepted his proposal for Loopy Landscapes and agreed terms, so he suspended work on Transport Tycoon 2 for around four months to work intensively on the new expansion pack. Loopy Landscapes included new rides, shops, scenery and scenarios as expected, with one notable change being the removal of the gate fee in the new parks. Sawyer did this so that the design and theming of the individual rides became more important and crucial to the success of the park. He'd realised that with an entrance fee, it didn't really matter if the guests didn't enjoy all of the rides in the park, and as such there was little incentive for players to really think about the design of every single ride. Without an entrance fee though, every ride suddenly becomes important because a poorly designed ride is no benefit to the park's finances. Although this change initially made it seem as though it would be harder to make money, it was actually possible to make more money than before because players could charge higher prices for the rides, especially when they were brand new. This change would also later make its way into the game's sequel.

While both expansion packs were criticised by reviewers for a lack of new gameplay features, they sold well - with Loopy Landscapes managing to put RCT back into the number one spot in the US PC games sales charts by December 2000. Following the release of Loopy Landscapes, Sawyer looked at the feasibility of doing a third expansion pack but realised that he'd already stretched the game to its limits and that it simply wasn't possible to add the kind of features he wanted while still keeping everything compatible with existing saved games. It made more sense to create a sequel, but he couldn't see himself immediately jumping into it. He did however have a clear idea of what RCT2 might look like, both in terms of what he wanted to do and also in terms of what players wanted, such as a landscape editor for creating custom scenarios. This wasn't possible with the original game because it had never been designed to manage an infinite number of scenarios, and several of the built-in scenarios had already reached the limits set by the game logic.

When interviewed at the time, Sawyer didn't think a sequel for RCT would happen for several years because he felt what was already out was the definitive version of the game, at least in terms of what was technically possible with PCs at that time. As new processors and improved hardware came to market however, it would offer new possibilities for a sequel. He didn't completely rule out the possibility of RCT2, but at this point he had twice abandoned work on a new instalment of Transport Tycoon and so he wanted to try his luck for a third time... But it didn't pan out, and so in June 2001 work began on RollerCoaster Tycoon 2.

RCT2 was based almost entirely on the code of its predecessor and featured a whole bunch of new rollercoasters, rides, shops, scenery and scenarios (including ones based on real Six Flags parks). One of the biggest new additions was the fan-requested Scenario Editor, which allowed players to change the map size and typography (with the maximum size of maps doubling to 256x256 tiles over the original game's 128x128 tiles), move the boundaries and entrance of the park, plant trees and bushes, specify the selection of attractions, and set the game goal and budget among other things. A wide range of new rollercoaster types were added to the game; both new and old were accounted for in order to give players untold freedom to build exactly what they wanted, with the ability to create, test and save custom coaster creations with the new Track Designer. Scenery placement was much improved with the ability to place pieces like walls and roofs individually, and new placement controls allowed for scenery and rides to be placed at varying heights with ease. A multitude of quality of life improvements were also added, such as the automatic removal of scenery during ride and path construction (a common complaint with the first game), the ability to construct rides and shops completely underground (in the original game only rollercoasters could tunnel underground), and scenarios no longer needing to be played sequentially in order to unlock them, with them all being made available to play right from the start. The latter being a point of contention amongst some fans.

There was a lot of pressure and expectation for RCT2 to be more free-form and less structured, which is why Sawyer configured it so that scenarios could be played in any order the player wanted. But what this added in terms of flexibility, it also took away in terms of challenge and sense of achievement that the original game had. This change also caused problems with the game's difficulty scaling because the scenarios were sorted alphabetically instead of by difficulty like in the first game. What's worse is that because of its alphabetical-sorting this meant that the scenario order changed depending on which language you were playing in, so some countries ended up having an easier or harder time with the game as a result. Sawyer also believed that the scenarios he created for the first game were better than what were available in RCT2.

A major criticism of RollerCoaster Tycoon 2, upon its release in October 2002, was that the game engine and user interface were almost identical to the original game, with only minor improvements to the game's graphics and animations. The most common critique being the lack of backwards compatibility with the original game and its expansion packs, with certain rides like the Steel Mini Roller Coaster being incompatible with its RCT2-lookalike the Junior Roller Coaster. To many reviewers, RCT2 felt more like a standalone expansion pack to the first game than a true sequel that pushed the series forward in any meaningful way, with comparisons made to upcoming games like SimCity 4 that featured more noticeable changes over its predecessors. RCT2 was even a runner-up in GameSpot's 2002 "Most Disappointing Game on PC" award, which went to Civilization III: Play the World.

Just like the first game, RCT2 had two expansion packs that added new rides, scenery and scenarios. Wacky Worlds and Time Twister both launched in 2003 but unlike the original game's expansion packs, Sawyer didn't develop them himself. Instead, they were created by Frontier Developments. RCT2 was the 10th best-selling PC game of 2002, with The Sims and its expansions taking up five of the top ten spots, and by April 2004 the RollerCoaster Tycoon franchise as a whole had collectively sold more than seven million copies worldwide. Despite its middling reviews, RCT2 was enjoyed by many players around the world and Sawyer himself considers the game to be his personal favourite entry in the series, as to him it was the "ultimate" version of the game.

After finishing work on RCT2 Sawyer felt as though he'd done everything he wanted to, and didn't have much vision for how to take the series forward. He knew that it would be impossible for him to carry on doing a new version of the game the way he'd been working; with him doing the programming and running the project, with only a single graphics artist and musician. Video games were getting bigger and more complex, with teams consisting of dozens of different people. He couldn't see himself working with a large team, and so he made the decision to license the rights for the next game to Atari (who had since acquired Hasbro Interactive and rebranded from Infogrames).

In March 2004, Atari officially announced RCT3 was in development and that it was being handled by Frontier Developments - the same company responsible for RCT2's expansion packs - with Sawyer providing design and technical consultation for the project. Despite being involved with RCT3 as a consultant, Sawyer gave Frontier complete creative control over the project. He decided that it was best to let those with the inspiration and expertise to design and lead the newer games themselves. Trying to work on somebody else's vision of how the game should be expanded, updated or modernised seemed pointless to him, and he didn't want to constantly interfere with their creative process. He'd also realised just how tough it was going to be to move the game forward without losing the playability of the original games. He knew the logical way to go was to add more detail using a fully 3D viewpoint and to remove some of the block-like construction restrictions but he couldn't get all that excited about taking the game in that direction for some reason, so the right thing was to hand it off to others who had the passion and drive to take the series forward.

Once work on RCT2 concluded in late 2002, and Frontier Developments took over development of the series, Sawyer once again gave Transport Tycoon 2 his full attention, using all of the ideas and experience he'd gained from his previous attempts, as well as from the original Transport Tycoon game. It would take a further two years to complete but unlike his previous attempts, this time he actually made it across the finish line. As a quick refresher of where we're at in the timeline; before creating RCT, Sawyer had made a graphical engine and world data structure purely for his Transport Tycoon sequel, but when that didn't pan out the engine was adapted for RCT instead. Then, after finishing work on the game's first expansion pack, Sawyer once again attempted a sequel for Transport Tycoon, only for that too to go astray prompting the decision to make a second expansion pack for RCT which became Loopy Landscapes. Between RCT and RCT2, Sawyer tried his luck for a third time, using an updated version of the RCT engine as a starting point, but re-written and fine-tuned for the new game. When the project was postponed yet again, much of the game engine was transferred to RCT2. But with RCT2 now finished, Sawyer was finally able to use the engine for what it was originally intended for - a transport, constructional world-based game.

With memory not being as much of an issue on PCs anymore, the scale and complexity of the game could be increased considerably, and thanks to more modern computing power he could simulate better competing companies than what was possible in the original game. He also had ambitions to design nearly all of the data the game used in such a way that additional vehicles, objects and even modes of transport could be later added without having to update the game code at all. Chris Sawyer's Locomotion, as the game ended up being called, was very much made with the gameplay of Transport Tycoon in mind, but in a larger, more detailed world that was simpler to play than its predecessor, as Sawyer had learnt over the years that adding complexity didn't necessarily mean a game was more fun to play. He designed the game in such a way that if you'd mastered the rollercoaster construction in RCT, then you'd be adept at building transport routes in Locomotion. Since he chose to forgo adding unnecessary complexity to the game, the depth of Locomotion came instead from the interaction with the landscape and the activities of competing companies. It was an uphill struggle to get the game reasonably well-balanced, as its free-form nature meant that you couldn't easily predict how the game would progress or what style of gameplay the player would end up using, so catering for all of the styles and scenarios proved difficult. It wasn't like with most games, where there's only one way to play and one way to win. There were so many different ways to play depending on what interested you or what you found fun (as he'd observed with RCT), and the challenge was to make sure that most players could achieve satisfaction and enjoy playing in their own particular way.

The increased scale and detail of Locomotion meant that designing the AI for the competing companies became vastly more complicated. Fortunately, the increased processing power of even low-end PCs at the time meant he could get the competing companies playing really well. In fact, they initially worked so well that he actually needed to slow them down considerably from early test builds, as they were monopolising the entire map within a few game-years. Like his other games, Locomotion was developed mostly in isolation without any input from Atari or other sources. Sawyer didn't know how long it would take to make the game, or even if he could achieve what he set out to make, so by developing the game independently he could take on the risk himself without having to commit to a particular design or completion date. It also meant that he was free to shift focus and experiment with other ideas along the way, whether that be in terms of completely new games or changing the game design depending on what did or didn't work well. One such example was with traffic lights, which were originally planned for Locomotion but it turned out that they only served to slow down the average delivery times and didn't help prevent gridlock. So instead Sawyer just made it so that vehicles automatically gave way at junctions.

Chris Sawyer's Locomotion was released in September 2004 to generally mixed reviews, with many outlets unfavourably comparing the game to its predecessors, citing minimal improvements to the game's presentation, mechanics and user interface. When interviewed about this, Sawyer expressed disappointment that Locomotion was received worse than Transport Tycoon, given that he believed it was a much better game in all sorts of ways. It was also the best piece of programming he'd ever done, with it being well designed, well planned and well written with some incredibly complex algorithms to handle the competing companies' AI, along with the plug-in data system, and getting the networked multiplayer working reliably. Locomotion also didn't sell very well, with Sawyer noting how the game wasn't as well marketed or supported by Atari the same way RCT was, but he also didn't solely blame that for the game's poor sales.

He realised that he'd missed the mark in terms of gameplay. While making Locomotion a bigger, more flexible and nicer-looking version of Transport Tycoon, the game wasn't able to capture the simplistic charm that the original game possessed, with its more restrictive grid-like straight track and roads.

He also recognised that Locomotion didn't have flashy 3D graphics that most modern games had at the time, and that it was easier to advertise and promote something good-looking rather than something which plays well. He believed that people would come to appreciate the game more with the passage of time, and he was right to some degree, as retrospective assessments of Locomotion were much more forgiving than initial reviews.

While Sawyer considered Locomotion to be nearly everything he wanted it to be, there was one area that he thought the game might lend itself well to; massively-multiplayer online gameplay. He believed that an online version with a large number of players playing within a single world would be fantastic, though he never got around to making it himself.

By the end of 2004 Sawyer had been living, breathing and working on nothing but games for nearly twenty years and he decided it was time to take a break and put more time and effort into his personal life and other interests, rather than spending 16 hours a day in front of his computer. He also felt as though he'd done what he wanted to achieve at the time and lacked the inspiration for taking his old games forward or creating new ones.

He was especially put off by the trend of games becoming more violent and destructive, as well as how the games industry was moving towards bigger and more complex games that required large teams of people to develop, which he had no enthusiasm for. His enthusiasm was soured further by a dispute with Atari over royalty sharing, which led to Sawyer suing the company in November 2005, in which he claimed he was owed $4.8 million in unpaid royalties from sales of the RollerCoaster Tycoon games.

According to the suit, Atari miscalculated payments to Sawyer between 1999 and 2003. He also alleged that Atari acted in breach of contract by refusing auditors full access to accounts between 1999 and 2001. Atari then counter-claimed Sawyer alleging that he had broken an agreement by letting Frontier Developments create a demo based on the RollerCoaster Tycoon franchise, and therefore induced Frontier to breach its contractual obligations to Atari.

In March 2007, the case was set to go to the UK's High Court, but Atari and Sawyer later settled outside of court in 2008. While the terms of the settlement were confidential, with the parties involved being unable to discuss details, the appointment of David Gardner as Atari's new CEO may have not been coincidental. In his new role Gardner was tasked with "rejuvenating the Atari brand" and, with the closure of this dispute, planned to revive the development community's faith in one of gaming's oldest brands.

In March 2010, Sawyer formed 31X Ltd to act as a holding company for the Transport Tycoon IP after Guy Herbert, a legal consultant, successfully negotiated the reversion of the MicroProse license. Sawyer's business agent, Jacqui Lyons, had persuaded him to bring the original game to modern touch-screen mobile devices but while he didn't initially have much vision for the project, he agreed with her that the game would be well suited for these platforms, thanks to their tactile inputs and high resolution displays.

Wanting to remain hands-off with the development, he enlisted the help of Origin8 Technologies to build a remake of the game for iPhones, iPads and Android devices, using Locomotion as a base. As time went on though, he was drawn into working closely with the team; overseeing the design and helping with debugging.

They initially had a few false-starts and setbacks, mainly due to the difficulties of converting his old x86 assembly code, whilst ensuring they kept the game's scale, depth and level of detail intact - as he didn't want to overly simplify or dumb down the experience. Fortunately, Steve Clark, a programmer working on the project, understood x86 assembly code and could also program in C, which was the preferred language for iOS and Android devices.

He had to figure out what each of Sawyer's routines did, as well as work out how to get them to work in C without being too slow. Sawyer worked with Clark to help him understand how bits of his code worked and to ensure the game ended up with the same functionality that the PC game had.

Getting the mobile device's display system to cope with the style of graphics the game needed was a huge hurdle, as the graphics chips were designed to handle polygons or sprites, and not pixel-mapped graphics. Another challenge stemmed from the game's heavy reliance on memory access to find data, such as the trains needing to know which track they're on and what part of the route they're taking.

While the memory in these devices was faster than back in the day, the access to each individual bit or byte of memory was slower because you needed to go through so many layers, rather than just telling the processor the specific bit of memory you wanted and what's inside it. Clark needed to do a fair bit of optimisation to get around these issues.

The game's UI and controls also received a complete overhaul to make them better suited for mobile platforms, and ensure the game was fully playable with only touches and gestures. The lack of a right-click, the ability to hover over buttons for tooltips and the precision of a mouse pointer were all things the team needed to think about and make concessions for.

Transport Tycoon Mobile was released in October 2013 as a premium title, instead of the more commonly used free-to-play model that most mobile apps had. Sawyer wasn't a fan of the way free-to-play games tried to make money through in-game purchases or advertising - particularly those that encouraged players to spend real-world money to take shortcuts - and didn't believe Transport Tycoon was a game that would be very well suited to that revenue model. He was also confident that the game wouldn't have gone over well with players if it had.

Marketing Transport Tycoon to mobile gamers, who were more accustomed to simpler freemium experiences, proved quite challenging. They wanted players to understand that this was in fact the actual classic game brought to life, with the same depth and strategy as the original. To that end, the team developed a brand new tutorial for touch-screen gameplay, with the goal of providing as much hands-on guidance as possible so that mobile players would understand how to use the mechanics and interface the same way as PC players did with the original.

New achievements were also added to introduce a bit of fun to the progression, such as "Morlock" which is awarded after building a station underground and "Swiss Cheese" which is awarded for building lots of tunnels. The team continued to provide regular updates and support for the game for several years after its release including adding new scenarios, maps, and achievements and even commissioning the game's original composer, John Broomhall, to completely re-arrange and re-record the original soundtrack.

The game was well-received by critics and players alike, however opinions have somewhat soured over the years due to the lack of updates and support for more modern devices, with the last official update for the game releasing on the 6th April 2017.

With the proven success of Transport Tycoon Mobile, Sawyer and the team at Origin8 set out to re-create the entirety of RollerCoaster Tycoon 2 for iOS, Android, PC and Mac. Despite being based on RCT2, the team decided to include all of the park scenarios from the original game to give players the challenge that was sorely missing from its sequel. These scenarios needed to be modified to work with the second game's toolkit to ensure that the experience remained well-balanced.

Much like with Transport Tycoon Mobile, development was far from smooth sailing, as converting the games' more than 500,000 lines of x86 assembly code once again proved difficult, but with the experience they'd gained from doing TTM they were confident they could tackle RCT2. TTM had proven that a mouse-driven PC game could be adapted well for a mobile touch-screen environment and that there was a market for older-style games on mobile platforms.

Newer versions of RCT had moved on in terms of their graphical style and the type of gameplay they offered, and Sawyer felt that this had left a gap in the market for bringing back the original games with their stylised but characterful graphics and their more simplistic gameplay. His hope was that the game would appeal to the players who remembered the original games, as well as new players who appreciated the solid gameplay and the amount of depth and content the game provided.

RollerCoaster Tycoon Classic was released for iOS and Android in December 2016, with PC and Mac versions arriving a year later in September 2017. The game received high praise from both critics and fans alike for its faithfulness to the original games and for being sold at a low price, with the only in-app purchases being the Wacky Worlds and Time Twister expansion packs, as well as a Toolkit featuring a Park Scenario Editor and Ride Designer for the creative types who wanted to build their own custom park scenarios and experiment with custom ride designs, with the ability to share their creations with friends.

Over the years, a number of fan projects have taken Sawyer's original games and converted them into open-source projects that aim to emulate the original games as closely as possible, whilst also adding new quality of life features and bug fixes. These open-source projects include OpenTTD, OpenRCT2 and OpenLoco which are reimplementations of Transport Tycoon Deluxe, RollerCoater Tycoon 2 and Chris Sawyer's Locomotion respectively - with a vibrant and active community helping to ensure these classic titles remain relevant for many years to come.

As for Sawyer himself, while he hasn't released any new games since his mobile conversions, he supports his local community by providing volunteer work, such as helping out the media team at a local primary school, as well as continuing to ride and enjoy different rollercoasters all over the world, with his personal coaster count exceeding 700 coasters. His favourite, Taron at Phantasialand in Germany, looks like something straight out of a Tycoon game.

If you enjoyed this video, and would like to support the work I do, then please consider checking out my Patreon or becoming a channel member - where you can get exclusive access to new documentaries before anyone else, see monthly updates on the progress of whatever documentary I'm currently working on, and even have your name featured at the end of each documentary you've funded. You can also send a one-time tip via Ko-Fi or a Super Thanks if that's more your style. Click the card, or the links in the description below.