Game 003 - Nimrod: A Mighty Computer Before the Lord

Release Date: 4th of May, 1951

Platform: Standalone Device

Genre: Strategy

Developer(s): John Makepeace Bennett, Raymond Stuart-Williams

Publisher(s): Ferranti

You think I was joking when I said last article that video games had become an international sensation by 1950? Well, Nimrod here sees us jumping across the Atlantic Ocean to Great Britain to a game designed by an Australian (you beauty!), built in England, and shown off in two separate countries! How's that for international?

Nimrod, also happens to be the first of many, many, many computerised versions of the game of Nim we'll be seeing over the formative decades of video game history. Yes, I did just say decades - that wasn't a typo. As a simple and wholly predictable mathematical logic game, Nim was extremely popular with programmers and computer scientists.

Aussie, Aussie, Aussie!

It's a personal point of national pride for me, seeing an Australian-designed game appear so early on in the history of video games. Sure, we have top-tier pedigree nowadays (Hollow Knight, anyone?), but seeing that Aussie involvement in gaming can be traced back to the very formative years of the medium does bring me a little joy.

John Makepeace Bennett (OA) was born in Warwick, Queensland on the 31st of July, 1921. Little was recorded concerning his formative years in this regional town - we know, at the very least, that he wasn't educated there. Rather, his education took place at a Gold Coast boys-only boarding school known as The Southport School. Bennett chose to pursue higher education at the University of Queensland, graduating with a Bachelor of Civil Engineering - a choice of convenience inspired by his older brother, Ron, who previously completed the course. John had free access to all of his brother's study notes - no doubt making the course rather easy for him.

John M. Bennett (OA, 1921 - 2010)

Bennett's tertiary study years also happened to coincide with the outbreak of World War Two. Once Bennett had that shiny piece of paper in his hand, it was off to war with him. Fortunately, his smarts kept him off the battlefield, and he was instead stationed with a RAAF (Royal Australian Air Force) radar unit on the Wessel Islands - a group of uninhabited islands at the very tip-top of the Northern Territory, almost a stone's throw across the ocean from Papua New Guinea. Once the war was over, Bennett returned to education, studying electronics and mechanical engineering, and mathematics.

1947 would be the year to define the course of Bennett's life work. It would see him leave his homeland for the computer labs of England, where he would work on some of the most significant early computers, which would, in course, lead him to design Nimrod. While working in Brisbane, Bennett would hear a radio talk concerning one of those new-fangled computers being designed over in the United Kingdom. This computer was the Pilot ACE - the scaled-back version of the ACE computer designed by Alan Turing in 1945. Inspired to bring the technology back Down Under, Bennett left for Cambridge University. As a side, Turing and Bennett did indeed cross paths - in 1950, when both were working for Ferranti.

At Cambridge, Bennett would not only complete his PhD, but also assist in the development of several important early computers - we'll talk about those more fully with the next few upcoming games. As a research assistant under Sir Maurice Vincent Wilkes, Bennett would be part of the team that built the Electronic Delay Storage Automatic Calculator - EDSAC for short. EDSAC was one of the very first stored-program computers, and is significant for us with regards to video games, as we have not only evidence of several games programmed for it, but one of the contenders for the title of "first video game," A.S. Douglas' OXO, is still playable today (though only in simulated form.)

EDSAC. Looks more like a library than a computer.

While completing his PhD, Bennett was approached by the electronic engineering firm, Ferranti. Ferranti might be a familiar name to those of you who read the Bertie the Brain article, as the company that provided Josef Kates' University of Toronto with the FERUT computer. Ferranti wanted Bennett's specialty knowledge for a project being undertaken for the University of Manchester. Ferranti had, at this stage, already completed their Ferranti Mark 1 computer - the successor to the earlier Manchester Mark 1 - but further upgrades were required. Bennett's new-found familiarity with computers made him a valuable asset for Ferranti, and he accepted a role on the project in 1950. He would develop the instruction set for the upgraded Ferranti Mark 1, commonly known as the Ferranti Mark 1 Star (usually written as Mark 1). 

Ferranti Mark 1* in operation.

This is where he would work alongside both Alan Turing, and the parents of Tim Berners-Lee, future inventor of the World Wide Web. The web of connections grows exponentially the further I dig in! Bennett would also meet his wife, Mary Elkington, while working on the Ferranti Mark 1 in 1952. Bennett had already done much in his brief career, but he was yet to lead a project. That opportunity came in 1950.

Faster Than Thought

In the late 1940s, plans were commenced for creating a celebration of British ingenuity. This was the Festival of Britain, and it was due to commence in the summer of 1951. A year prior, the event organisers approached Ferranti - who were at the time one of the leading computer manufacturers in the world - asking if they would be able to create a full-scale computer display for the Festival. Ferranti agreed, but as the deadline approached, it became obvious to all at the company that the promised device would not be completed in time. Ferranti made the unfortunate news known to Festival authorities on the 30th of November, 1950.

Enter John Bennett with a solution.

Bennett proposed that something could still be done for the Festival. Something of a smaller scale, but could still display the basics of computer science and programming. He suggested that they create a computer that could play the game of Nim. He worked through December of 1950 to develop the design for this gaming computer, and it was presented to both Ferranti and Festival authorities. They both approved it, and work commenced in January, 1951. Progress was rapid, with construction undertaken by a team of nine, spearheaded by Ferranti engineer Raymond Stuart-Williams, and the computer, named after the Biblical character Nimrod, was completed in April of the same year. Just in time for the Festival!

A poster for the Festival of Britain.

Nimrod was transported from Manchester to London for the Festival over April, and the four ton, nine-foot tall beast was fully constructed and working on the 1st of May. The Festival commenced on the 4th of May, with Nimrod front-and-centre in the science section of the South Kensington location. Advertised as an "electronic brain" and "faster than thought," the Nimrod display captivated its audiences for the entire five-month run of its exhibition at the Festival. It's reported that Bennett's colleague, Alan Turing, tried his hand at beating Nimrod. It did not succeed in having a flawless performance, with a few minor malfunctions occurring during the months, which were easily resolved. Although, Bennett recounts, its reception was not quite as intended,

"The machine was a great success but not quite in the way intended, as I discovered during my time as spruiker on the Festival stand. Most of the public were quite happy to gawk at the flashing lights and be impressed. A few took an interest in the algorithm and even persisted to the point of beating the machine at the game. Only occasionally did we receive any evidence that our real message about the basics of programming had been understood."

(The source for this quote can be found here. This website is chock full of Nimrod facts and trivia, and is well worth your perusing.)

Nimrod, in the flesh.

Curiously, this reception mirrors the reception of Bertie the Brain in Canada a year prior. There's quite a few parallels between Nimrod and Bertie, as a matter of fact:

• Both games were designed by one man, with countries of birth foreign to the country their games were built in (Josef Kates was Austrian; John Bennett was Australian - don't get the two mixed up!)
• Both games were designed specifically for a national exhibition.
• Both games had a primary purpose other than entertainment.
• The reception of both games ran contrary to their intended purpose.

Unlike Bertie, however, Nimrod was not immediately dismantled after the conclusion of the exhibition. That was the intention, but Nimrod was offered a second life. In the last week of the London exhibition, a suggestion was made to the Nimrod team by the Board of Trade to take Nimrod for a Continental excursion to visit the old enemy, Germany. The Berlin Industrial Fair was due to begin on the 6th of October, and there was a dedicated United Kingdom exhibit planned for it. It took a bit of convincing, but the Nimrod team eventually agreed, upon realising that it would be quite easy to accomplish the task of dismantling and flying the machine overseas in the matter of a fortnight - all they had to do was work 24 hours a day, and work twice as quickly! How hard could that be?

An artist's rendition of Nimrod.

I won't recount the full tale of Nimrod's trek to Berlin, but needless to say it was quite the arduous journey! There's an entire article available to read online, written by Raymond Stuart-Williams himself, where he recounts the full adventure with typical dry, British wit. The liner notes version would be that the whole saga was a comedy of errors, featuring a plane too small to store all of Nimrod, another dodgy plane with a flat tire and doors that wouldn't close, and having to convince Dutch customs officials that Nimrod was not actually an atom bomb! Maybe the instinct to refuse initially had merit, after all...

Needless to say, Nimrod did somehow manage to arrive at its destination unscathed. Stuart-Williams recounts that the team, plus a few German engineers, had Nimrod up and running within 28 hours of arrival. The German audiences were just as - if not more - fascinated by Nimrod, and several dignitaries tried their hand at beating Nimrod - and failing! Word of the machine even made it over the Berlin Wall, with one university attempting to build a replica.

Nimrod in Germany.

Once the German exhibition concluded, it was time for Nimrod to return home for a well-earned retirement. There's some misinformation floating around the internet suggesting that Nimrod also made an appearance in Canada, but this is flatly debunked by Stuart-Williams' article. He states that, once Nimrod was returned to Britain in November, 1951, having completed its task, it was promptly dismantled.

A Computer of Renown

I'm actually quite pleased to say that Nimrod possess a legacy and influence, unlike the first two games of our chronology. Although, it's only tangential influence, rather than direct. But it's better than nothing!

As formerly noted, John Bennett worked on development of EDSAC prior to designing Nimrod. Very soon, people are going to start making games for EDSAC - most notably Sandy Douglas' OXO in 1952, which might be the first "true" video game, under some definitions. The other contender for that title, Christopher Strachey's Draughts, also has a potential, yet more speculative connection due to Bennett's involvement with Ferranti during the Ferranti Mark 1 and Mark 1* development - the former of those being the platform Strachey would get his Draughts program fully functional on. There's a strong possibility that Bennett and Strachey crossed paths in 1952 while both were working on their separate projects at the University of Manchester.

This gives us a substantial update to the Video Game Family Tree, as we'll see below. Aside from the Air Defense Simulation connections, we actually have some real connections to make!

The upgraded family tree as it stands at 1951-52.

As for John Bennett post-Nimrod, he would make his way back Down Under in early 1956 to become an Aussie computer pioneer. He took up a role with the University of Sydney's Department of Physics, where he would be in charge of overseeing the university's brand new supercomputer - the SILLIAC, named after the American ILLIAC. This was regarded as one of the most powerful computers in the world in the mid-1950s, and Bennett would - quite literally - write the book on it. He also developed Australia's first computer programming course during this time. Bennett would remain involved in computer science for his entire professional life, chairing several state-based computer societies, and he eventually became the inaugural president of the Australian Computer Society in 1966.

In 1983, Bennett was made an Officer of the Order of Australia for his contributions to computer science and technology, and retired in 1987. He was also awarded the Centenary Medal during Australia's centennial celebrations in 2001 for his services. After enjoying a long retirement, whilst still being regularly tuned into the latest developments in computer science, John Makepeace Bennett passed away on the 9th of December, 2010, at the ripe old age of 89.

John Makepeace Bennett.

Personally, I've quite enjoyed learning about this great Australian, whom I was totally unaware of prior to writing this article. Part of the joy of doing this blog is discovering the stories of great people, who achieved great things that still have an impact on us today, who would otherwise be forgotten about.

Gameplay

Let's finally talk about how Nimrod actually plays. Nimrod plays your standard, garden-variety game of Nim. What Nim is warrants some explanation here, as it's technically the first time we see Nim appear in the chronology.

Before diving in, it's worth mentioning that you can play Nimrod today! There's an online simulator available on a Nimrod fan site here. This is a "stripped back" version of the author's full simulator, which he boasts is for the "most modern operating system", HAIKU/BeOS. I have my own thoughts on this decision, which I'll keep to myself to avoid offense. I'm thankful that we've been given a publicly available, browser-based version of this simulator, at least.

The Nimrod simulator.

Nim is an ancient game centred around object removal. There are multiple variations of the game, but the core gameplay element that does not change is that two players take turns removing objects from one or multiple piles of objects,. with the objective being to either be the player to take the last object, or to force your opponent into taking the last object.

Nimrod was designed with a fairly standardised ruleset for Nim, in which there are four piles of objects, represented by the computer as four rows of seven lights - seven being the maximum number of objects allowed in each pile. Nimrod allows the player to choose how many objects are in each pile, though the standard, 4-pile setup is 7, 5, 3 and 1. The player would setup the game by pushing a lever on the computer to put the game into "setup" mode. The number of objects they'd like in each pile could be selected by pressing a button above one of the lights in each row, corresponding to the number of objects in each pile, ascending from left to right - to have zero objects in a pile, you'd push the leftmost button, to have a full seven, you'd press the rightmost button.

Look at the pretty colours!

The player could also use other levers to adjust various game settings. There's not a whole lot of variation with Nim, but Nimrod only includes the bare essential settings - whether the winner is the one who takes the object (called "simple" mode by the game), or the one who forces their opponent to take the last object (called "reverse".) This setting was also changed by a lever, as was whether the player or computer played the first move. Quick hint: if you want to win Nim, always go second! Nimrod plays a perfect game, meaning the player can never win going first.

Once the player's preferred settings were selected, they'd then push the "setup" lever back down into the "play" position to begin the game. For the player to remove objects, simply press one of the other buttons above a light, and all the objects to the right of that button will be taken. We're so far back in computing history here that the computer can't automatically make its moves - the player has to press a button to get the computer to make its move! (If you think this is complicated, wait until you see Strachey's Draughts.)

My playing experience of the simulator was confusing initially, fumbling about with the controls until I figured out how it all worked. This, thankfully, didn't take too long, and I was able to beat the computer on a couple of occasions once I remembered how the logic behind Nim worked. I show you how to beat Nimrod in the video at the top of the article, and, to save me the explanation, there's an article here explaining the binary logic behind Nim. There's not really a whole lot to say about the game once the novelty of it being a computer game from 1951 wears off. It's just computer Nim, like any other computer Nim I've played before.

This was a winning setup for me.

On that note, I think it's a good time to move onto the scores.

Scores

Time Played: 25 minutes
I had to re-record my Let's Play, as OBS ate half of my voice on the original recording due to my filter settings.

Difficulty: 5 (Moderate)
Difficulty is tricky to judge with Nimrod, and by extension all other early Nim variants. The computer plays perfectly, meaning that it's either impossible to beat on certain settings. On the flipside, one can also take advantage of the settings and computer (provided you know how Nim's logic works) to also win infinitely.

Gameplay: 2
Nim is inherently a simple and not overly interesting logic game to me. These early computerised versions tend to be frustrating in that they're either impossible to beat due to the computer being programmed to play perfectly, or are infinitely beatable with the right settings, once you know the logic behind the game. The lack of variation in the computer's competency makes the game far less compelling than it could be.

At base, I give Nim in all its variants 1 point for being an actual game. Nimrod gets another point for the variety of game settings available, which is impressive for such an early computer game.

Controls: 5
Nimrod actually has a fairly simple control scheme, despite the appearance of complexity. Game settings are changed with the use of levers. Each light in the four rows, representing an object, has a button attached to it. You setup the game by pressing the button lining up with the number of objects you'd like in that pile, with the rightmost button being the most at 7. If you want no objects in a pile, you'd press the leftmost button. Actually... the more I try to explain it, the more complicated it actually seems...

Initially, playing the simulator, I did get a little confused on taking away objects at first. The game only takes away the objects to the right of whatever button you press, but not the object below the pressed button. The computer does not automatically make its moves, either - the player has to press another button to have the computer make a move.

Sound: N/A

Visual: 2
A score for visuals is quite odd to determine for Nimrod, as it doesn't exactly have a "traditional" display. The visual component is represented by individual lightbulbs - whether this counts as a true video display or not is up for debate. It's at least more interesting to look at than the average text-based game, so I'll give it a 2.

Story: N/A

Functionality: 4
This score is purely based on the first-hand accounts from Bennett and Stuart-Williams concerning Nimrod's actual performance. A few malfunctions occurred, but they were only minor, and resolved quickly.

Accessibility: 3
I mean... if you lived in London in 1951, and were able to attend the exhibition, then sure - it's accessible.

As a game, it's relatively easy to understand, as the Nim concept is simple to grasp. Although, the game's display appears overwhelming at first with all its lights, buttons and switches, which is where it loses a couple of points for me.

Fun Factor: 1
Not particularly exciting once the novelty wears off; it's just like every other game of Nim I've ever played, just with more involved controls.

Overall: 17

Proto-video games aren't going to score very well by their nature. 17 (weighted to 21.25) is a pretty terrible score, however. Not terrible enough to land Nimrod in the F-tier - but only just. One also has to keep in mind that games like Nimrod were not designed with a modern game dev's mindset - nobody knew what video game design principles were back in 1951. Nimrod was a showcase for a company's technical prowess, and its design reflects as much. I find it fascinating how much Nimrod parallels with Bertie the Brain, despite the two games being created on opposite sides of the world, and with no contact.

Next Up: Mate-in-Two (Gaiden Article)