Monday 22 November 2021

SimEarth: The Ultimate God Game?

Fig. 1: Does old necessarily have to mean primitive?

God-games are a genre of video games that on one hand feel obscure, but, on the other hand, it also feels like everyone has played at least one game of this type during their childhood. What actually counts as a god-game is difficult to define, as simply putting the player into a controlling position over people, cities, wars and societies would mean that most strategy games like Civilization or Command & Conquer would have to be counted as god-games as well, but nobody does that. Instead, the genre is commonly imagined as putting you into powers outside of human influence and there have generally been two approaches by game designers in how to do that. One is the Peter Molyneux one of games like Populous or Black & White, which let you roleplay as a deity as imagined by ancient mythologies. The other is a more simulation-based approach, which lets you play with or mimic natural progressions, which in reality happened over geological timespans, such as the evolution of life and the formation of natural landscapes. Of the latter kind the most recent and still widely remembered big budget example is 2008’s Spore by Will Wright and Maxis, which put you in charge of guiding an alien species from planktonic beginnings all the way to interstellar exploration. Spore promised a lot and I remember it being the first game in my childhood I actually felt hype for after reading about it in a newspaper. When it released it fell short of expectations however and, while not at all bad, is remembered by many people as a disappointment (personally, I had a lot of fun with it when I was 10 and still play it from time to time). While the game offered unprecedented editing tools in the creation of your own lifeforms, vehicles, architectures and so on, which I have yet to see replicated to the same degree in a modern game, many of the other gameplay aspects were very basic and not fleshed out. The game’s stages of progression are basically simplified versions of multiple other genres, which just makes you wish you were playing deeper examples of those. More egregiously perhaps, the game does not actually simulate anything outside of the locomotion and movements of your created creatures and vehicles. Evolution does not progress in the game naturally, things only change when you make them, which has prompted some reviewers like LazyGameReviews to note that a game about evolution may be unintentionally endorsing the pseudoscience of intelligent design (Maxis or at least the game’s marketing team seem to have been aware of this contradiction and jokingly referred to its philosophy as “creatiolutionism”). The pathway that the game gives you is also relatively linear. While you are given the choice to be herbivorous, carnivorous, omnivorous, pacifistic, religious, war-like or commerce-oriented through the consequences of your actions, all your created species must, for example, progress from water to land in order to advance (some interviews with Will Wright indicate that, the option to stay in the oceans throughout the creature stage and all the way into the civilization one seems to have been a planned option at one point).

Fig. 2: Oh how much I miss these times.

While Spore was not a commercial failure, seeing how it was successful enough to warrant multiple spin-offs, a sequel has never been in the works and this, combined with its shortcomings, has made people long for a true successor which addresses all the flaws and in turn creates the ultimate simulation game, SimEverything if you will (Spore’s original working title). There have been some interesting attempts, which scratch at least a few of the itches left behind by Spore, such as S.P.E.C.I.E.S.: Artificial Life, Real Evolution, Ecosystem, Universe Sandbox, Adapt, The Sapling, Niche or Tyto Ecology, but many of these do not have the same polish or reach the scope which EA-backed Spore tried to tackle, due to being made by significantly smaller studios, sometimes just one person. Many are also still very early in their development with progress being very slow. The God-Game-Heiland, the ultimate simulation game that addresses all of Spore’s failures and simulates an entire world and its history, has yet to be made.

Or has it?

Background: Will Wright and James Lovelock

Fig. 3: Will Wright, pictured here during a Spore presentation with an arm possibly broken by Electronic Arts executives after not producing enough The Sims DLCs. After the release of Spore, Wright left Maxis and the gaming industry at large, though in 2018 he announced his comeback with a still mysterious mobile game named Proxi.

The year is 1989, the Berlin Wall has fallen, Stephen Jay Gould’s Wonderful Life has climbed bestseller lists and a man named Will Wright has just released a game called SimCity. SimCity was born out of the first game Wright worked on, Raid on Bungeling Bay, in which you need to shoot down military bases and cities in a helicopter. During development, influenced by books like Urban Dynamics or A Pattern Language by authors such as Jay W. Forrester, Christopher Alexander and Stanislaw Lem, Wright found it far more fun to build the cities than to destroy them and decided to make a game out of that. SimCity was a milestone in software history, not because of its technological achievements, but because of its concept. It was a “game” which did not have a definitive win or lose state, even though this is one of the defining characteristics of games, be they digital or otherwise. If you did well with your city, you just kept on going until a natural disaster hit or you got bored. Instead of referring to it as a game, Wright coined the new term “software toy” to describe it and many other games of the Sim-genre. At the time the idea of a software toy was perplexing for publishers but at the same time very refreshing for audiences and the game became an immediate success. A sequel was a no-brainer, but it needed to be larger and newer in scale. What is next after simulating entire cities? The whole planet, Wright and Fred Haslam apparently said to themselves, but who could help them simulate that? By coincidence, one of Wright’s neighbours at the time was Stewart Brand, an editor of the magazine Co-Evolution Quarterly, and Brand was by coincidence also a friend of one James Lovelock.

Fig. 4: James Lovelock, seen here clearly having a good time.

Lovelock is an important though controversial figure in the history of earth system sciences (what I am currently studying as my minor) and other geosciences. In the 60s, while working as an advisor for NASA, he was the first to propose in a significant way that the different components which make up the Earth and its inhabitants, such as the atmosphere, hydrosphere, geosphere, biosphere and anthroposphere, are not separate entities, but instead influence each other, sometimes massively, to form one giant interconnected earth-system. Nowadays, this seems pretty obvious and not controversial (unless maybe you are a climate-change denier) and is pretty much the basis for all earth sciences. Lovelock went one step further, however, and argued that this earth-system is able to regulate itself through feedback-loops between its sub-systems in order to continuously maintain habitability and stability. For example, when the sun’s energy output was up to 25% lower than today, the biosphere put out more greenhouse gases in the form of methane, so that the atmosphere could heat up the rest of the sphere, but as the solar input increased, the biosphere evolved lifeforms in the form of plants, which could absorb the excess carbon from the atmosphere and through fossilisation store it safely in the geosphere in the form of coal, maintaining a stable global temperature throughout the whole process (a more crass example I heard a colleague make, and which I do not endorse, is that the biosphere evolved Covid-19 like an antibody to impede the Earth’s cancer that is humanity). In this sense, the Earth functions like a living organism, with the components functioning like organs. Since its inception, Lovelock’s idea was strongly criticized, accused of being teleological, sometimes even esoteric (Lenton 2016). While Lovelock never claimed that this self-organising system was conscious, instead evolving entirely through natural processes, it certainly did not help that he embraced William Golding’s coining of this idea as the Gaia-Hypothesis, named after a Greek titan that has become prevalent in many new age religious movements. There are also various examples of the Earth not being able to stabilize itself, during mass extinction events for example (though one could argue this is just the earth-organism being sick or injured for some time before beginning the healing-process).

Even if a hypothesis might be problematic, it can still be useful, as Will Wright found out. As Lovelock notes in an interview with the gaming magazine The One, most computer models at the time, which tried to individually simulate atmospheric or geospheric processes on their own, quickly broke down into instability. Only when the spheres were linked together, did the simulations reach long-lived, stable conditions. The Gaia-Hypothesis thus became the base for Will Wright’s SimCity-sequel, and Lovelock one of the game’s advisors. He would also end up writing the foreword for the manual, as well as for the official strategy guide, written by Johnny L. Wilson (If you ever get the chance, I recommend getting your hands on it, because it is a very engaging read). In the span of just one year, SimEarth: The Living Planet was released in 1990, over 30 years ago.

The basics of the game

Fig. 5: The game’s editing window. Taking my own screenshots of the game was a bit difficult, so please excuse the low resolution.

Like SimCity, SimEarth is a tile-based two-dimensional game viewed from a top-down perspective, but things are significantly more zoomed out, as you are in charge of a whole planet. You can view it from three windows: The zoomed in Editing Window, in which you can interact with the world, the zoomed out Map, on which you can apply various filters to see, for example, wind patterns or tectonic movements, and the rotating Globe. SimEarth is about as complete of a planetary simulation as you can get. The geo-, atmo-, bio- and anthroposphere are all distinct, but interconnected entities, which are simulated in detail. And you can change every aspect of them. In the geosphere window, you can for example edit the pace at which continents move, volcanoes erupt, meteors fall down or heat escapes the planet. In the atmosphere window you can change albedo, rainfall, the greenhouse effect and more. The biosphere model lets you edit the rates of mutation, speciation, “advancement” and so on, while the civilization window lets you adjust what sorts of technologies and energy sources cultures on your planet value, which will in turn have an influence on wars, pollution and the greenhouse effect. Every change on your planet you can analyse in more detail by pulling up graphs and statistics menus.

Fig. 6 – 8: Three of the different control panels in the game.

Apart from these options you can also terraform the planet in a more hands-on way by placing buildings and tiles SimCity-style, such as CO2-generators, biome-tiles and lifeforms. As befits a god-game, you can also unleash natural disasters, such as comets, meteors, plagues and nuclear bombs. These actions cost energy-points however and how many of those points are available to you is determined by the difficulty in which you play a new game. In experimental mode your budget is thankfully unlimited. Unlike Spore, and this is an important difference, you do not have to interact with the world at all to see interesting results, as everything from technological and biological evolution to plate tectonics and ocean currents are fully simulated by the game. If you wish you can just let it develop on its own from a starting point and observe what will happen, like a gigantic digital ant farm.

Fig. 9: The many tools and lifeforms the game provides you with.

The game’s timescale spans billions of years, but its pacing changes depending on which phase your planet is in. In the early stages, before life has developed, it runs at a geologic timescale where millions of years pass in seconds. The next phase, where you have evolved animal life but not yet sapience, it runs at the slightly slower evolutionary timescale, and afterwards the civilized and technological timescales. You can nonetheless speed up the game’s time yourself. The closest to a winning condition the game has, is once a civilization has reached past the nanotech-age and decides to leave the planet in a mass exodus (mind you that this game came out one year before Sid Meier’s first Civilization game with its Alpha Centauri victory), but the game does not end then, instead the planet just reverts back to an evolutionary stage and allows sapience to evolve again from scratch. The official end of the game, the closest to a losing condition, is once 10 billion years have run their course, because at that point the sun enters its red giant stage and swallows the Earth. Do not worry about that, if you do not mess up catastrophically, most of the interesting stuff will have happened on your planet long before you reach that point.

Fig. 9: World map seen through the altitude filter.

When beginning a new game, SimEarth offers you 8 different scenarios to choose from:

  • Random World: You are free to choose at which developmental stage your randomly generated planet starts in and do not have any hard-set goals. Just mess around or watch things play out on their own.
  • Aquarium World: The whole planet is covered in water and stuck at the evolutionary stage. You need to form landmasses in order for life to advance.
  • Stagnation:  Humanity is stuck in the stone age and you need to help them progress.
  • Cambrian Era: You get to start at the very beginning of animal life and have to help it progress towards sapience.
  • Modern Day: You are in charge of Earth as it was in 1990 A.D. and have to avoid humanity destroying itself through environmental pollution, nuclear war or pandemics.
  • Mars: You need to terraform the red planet to make it habitable in 200 years. You are not allowed to change the geosphere or atmospheric panels and instead need to use tools and your limited budget.
  • Venus: Same as above, just even more difficult.

And lastly there is the Daisy World, which was likely put in at the behest of James Lovelock. In order to show that the phenomena described by the Gaia-Hypothesis arise through natural, not supernatural, processes, he designed the Daisy World model in 1983 to show how lifeforms can unconsciously regulate their own environment to maintain habitability for a long time. Imagine a planet that is only inhabited by black and white daisies. The petals of the black ones are better at absorbing light, the ones of the white ones are better at reflecting it. At the beginning of the model, the planet is irradiated by a sun whose energy output is still very faint. At first the black daisies dominate, as they are better able to absorb the sun’s heat. As their populations spring up and cover their whole planet, their absorption also heats up the planet, to a degree which would not have been reached if there were no daisies at all. Now the planet becomes warm enough that the white daisies can get a foothold. As the luminosity of the sun increases, some areas on the planet become too hot for black daisies, allowing the white ones to expand their range. In turn, their white petals reflect more sunlight, which keeps the temperature of the planet lower than it would be with no life at this stage.

Fig. 10: A classic Daisy World graph. Compare the temperature curve of the world with and without daisies.

This continues until equilibrium is reached between the daisy populations and in turn the planet’s global temperature. As the sun’s luminosity further increases, white daisies become dominant and so increases the planet’s albedo, helping it cool down. Eventually white daisies cover the whole surface, until the sun’s luminosity becomes so great that even they cannot regulate the temperature anymore. As we watch the Daisy World’s temperature on a graph, we will see that the planet’s temperature curve is, except for the beginning and the end, a nearly horizontal line, while on a Non-Daisy World, it would just move gradually upwards. The daisies, therefore, without knowing, regulated their planet to keep it habitable for as long as they could. SimEarth lets you simulate your own Daisy World scenario, where you can plant your flowers and watch how they influence the planet’s conditions. You can even go further than Lovelock’s original model, by looking what happens, for example, when you introduce animals who feed on the daisies into the mix. Will they help stabilize or destabilize conditions?

Possible Worlds

A basic summary like this does not really do the game justice. We need to go a bit more in-depth. Let us start by looking at the types of lifeforms, simply called taxa by the game, which can evolve on your planet. The game has 15 taxa, which can all diversify into a primitive-to-advanced spectrum of 16 species, certain subsets of that spectrum being able to evolve into new taxa. In the biosphere panel, the rate at which a taxon diversifies among itself is controlled through the advancement toggle, the rate at which new taxa evolve out of old ones through the mutation toggle.

  • Prokaryotes are the first to evolve in your world’s oceans and will from that point on already begin to significantly alter your planet by removing methane from the atmosphere.
  • Eukaryotes arise shortly after and will be your planets first major oxygen-producers. They can arise out of the eight last prokaryote species to evolve.
  • Radiates are the first group of multicellular animals to evolve on a naturally generated world, ushering in the evolutionary timescale. They evolve out of the four most advanced eukaryote species. The class Radiata has not been in use for a long time and the game seems to treat this simply as a wastebasket taxon for all radially symmetrical animals, as the in-game icons represent it with both jellyfish and starfish, which in real life are radically different beasts (while the radially symmetrical state was likely ancestral to cnidarians, starfish and other echinoderms evolved from a bilaterally symmetrical animal not unlike our own ancestor and secondarily reverted to this more “primitive” bauplan). Unlike the two unicellular taxa, radiates do not thrive equally as well in all of the three oceanic tiles, instead preferring the shallows and open ocean over the deep sea. Unlike their real life counterparts, the radiates of SimEarth can also live on land, in jungle and swamp tiles! You will soon read why.
  • Arthropods evolve from the first eight species of the radiates. Like them, they thrive in continental shelf and open ocean tiles and also do well in swamps, but only barely survive in jungles. As seen by the in-game icons, this class is more meant to represent marine arthropods such as crustaceans and trilobites. The reason is:
  • Insects are treated as their own taxon apart from arthropods, likely to better represent their wide diversity and range of habitat given the gameplay limitations. Otherwise, the arthropods would have likely become too overpowered in every playthrough of the game. Insects evolve out of the fifth to twelfth species of the arthropods and they can survive in almost all terrestrial biomes (except arctic), more than almost any other taxa in the game, and even thrive in continental shelf.
  • Carniferns are one of only two taxa which the player cannot plant themselves through editing tools, they only appear under certain conditions. They are carnivorous plants, which in in-game terms “evolve” out of the first four species of insects. What this really means is they arise through co-evolution as a natural response to high insect-populations. Carniferns can live in boreal, desert, grassland, forest, jungle and swamp tiles, but do not thrive well in any of them unless the insects are thriving as well. Carniferns do not give rise to any future taxon (though read on).
  • Mollusks evolve out of the first four species of the arthropods (perhaps meant to reflect the superficial resemblance between chiton snails and trilobites). Like the radiates, they all do well in continental shelf, open ocean, swamp and jungle tiles.
  • Fish evolve out of the fourth to seventh species of mollusks. Wilson’s walkthrough muses that this might be a nod to the earliest vertebrates being heavily armored animals not unlike the nautiloids. Fish thrive in all three ocean biomes and, what the manual does not tell you, can also survive in the swamp biome, which like with the radiates has further implications later on.
  • Trichordates are a weird inclusion, in part because there is no taxon in real life with that name. The game’s manual states that “Trichordates were a class of animal with three-chord spines. They lived and died out long ago on Earth. We felt sorry for them, and are giving them a chance for survival in SimEarth.” Their appearance strongly indicates that they are based off the real life Trilobozoa, a phylum of animals (?) possibly related to cnidarians, which appeared and quickly died out in the Ediacaran over 550 million years ago even before the Cambrian radiation (though there is evidence that the Paleozoic conulariids may have descended from them). Why exactly Maxis renamed them (and interpreted their tentacles as backbones) I am not sure. Interestingly, trichordates can have two origins in the game. They can evolve either directly out of the ninth to twelfth species of radiates (implying that they descend from starfish that lost a few arms) or the ninth to twelfth species of fish (implying that they are vertebrates which somehow grew two extra backbones). Trichordates can survive in swamps, deserts and grasslands, but truly thrive in boreal forests and jungle biomes. They can therefore surprisingly be the first major terrestrial taxon to evolve on your planets. Trichordates are not the ancestors to any later group, technically making them a dead end (though read on).
  • Amphibians can evolve from the first eight species of fish. They can live in boreal, grassland and forest tiles, but do best in continental shelf areas, jungles and swamps.
  • Reptiles evolve out of the first eight species of amphibians. They can live in boreal and swamp tiles, but do best in grasslands, jungles and deserts. They in fact stand out for being the only taxon which can truly thrive in desert biomes.
  • Dinosaurs are treated as their own taxon because, let us be honest, they deserve it. They evolve out of the first eight reptile species and can live in roughly the same biomes, but can not deal with deserts and grasslands as well. They actually do best in ordinary forest biomes, perhaps reflecting their long co-evolution with conifers and flowering plants.
  • Avians, as the game calls them, or birds as they are better known to the layman, evolve out of the first four dinosaur species (which is actually interesting, as, while on the winning-lane, the hypothesis that birds descend from dinosaurs was still being debated in the late 80s and the game could have represented the classic thecodont hypothesis as a second origin for them by having them directly descend from the reptiles). Birds barely survive in desert and swamp tiles, do okay in grasslands and thrive in forests, jungles and boreal forest, making them come in conflict with their predecessors.
  • Mammals can have two (technically even three) different origins. They can evolve either out of the ninth to twelfth species of the reptiles or the fifth to eight species of the dinosaurs. Why the latter path is even an option, I do not know. Maybe it is a nod to the fact that prominent mammal-ancestors like Dimetrodon are often erroneously referred to as dinosaurs. Interestingly, according to Wilson’s guide, their ancestry actually has an influence on their environmental preference. If mammals descend from reptiles they will do better in grasslands, if they descend from dinosaurs they will do better in the jungles (I never actually tested this). Apart from that they will also do well in swamps, forests and boreal forests, but barely survive in deserts.
  • Cetaceans, like with the insects and the arthropods, are again treated as separate from the mammals to better reflect environmental separation. Note that the in-game icons indicate that seals and manatees are included in this taxon as well. They evolve out of the fifth to twelfth species of mammals, do well in all ocean tiles and can also live in swamps. Fascinatingly, should mammals somehow die out on land, the last four species of cetaceans can evolve into land mammals again to restart their evolution (the third mammal-origin I mentioned).
  • Machine life is the final taxon, which at the time of the game’s release was a complete easter egg. It can appear in your world if a city on the nanotech technological level is destroyed by a nuclear bomb. Like all the other taxa they start out simple, as robotic home appliances that survived the blast, but will gradually adapt to the wilderness and diversify and advance through self-improvement. They can survive and thrive on all tiles of the planet and, if advanced enough, can outcompete all the biological taxa. If you let things go really sour, they will wipe out all life and eventually cover every tile on the planet, meaning the game lets you simulate a form of Kim Drexler’s famous Grey Goo scenario.

Now here is the big surprise, the great kicker, the smackdown: ALL OF THE MULTICELLULAR TAXA CAN EVOLVE INTO A CIVILIZATION. Yes! All of them! From radiate jellyfish to nanobots! From trichordates to dinosaurs! Even the mainly marine ones, such as fish and cetaceans, have a chance at sapience if they happen to live on swamp tiles and the carniferns too get to build log-cabins out of their non-sapient relatives if you wish so. Some taxa are more likely to develop sapience than others. As the manual states: Mammals, dinosaurs, birds and reptiles are the likeliest, these are then followed by cetaceans, insects, amphibians and carniferns, which are then followed by fish, arthropods, mollusks and radiates (Curiously the manual does not state where the trichordates lie, and I heard conflicting information on these. Some online sources state they are the least likely to develop sapience, while Wilson’s Bible says they are very likely to if given the chance). Though one may be sceptical on how accurate the manual actually is, as Wilson’s guide notes that one of the most common complaints people had with SimEarth at the time is that mollusks had a disproportionate tendency to develop sapience, to the point where the guide actually has to give tips on how to prevent the Tentacle Acres Ending. Some of this may be due to differences between versions of the game. However, unless you are playing on a limited budget or setting yourself up for a challenge, none of these likelihoods have to matter to you, as the game allows you to choose who gets to eat from the tree of knowledge. One of the terraforming tools provided to you is a monolith, which when placed on a population of your choosing has a 1/3 chance of elevating a taxon into the stone age, a straight up reference to 2001: A Space Odyssey by Arthur C. Clarke and Stanley Kubrick (of note is that, unlike the movie/novel, the monolith in some versions of the game has a circle-shaped hole in the top half. The exact same type of monolith would later appear again in Spore as an uplifting-tool during the Space Stage).

Fig. 11: When you just want to give natural history the middle finger.

This open-endedness of who will get to dominate your planet is, I feel, highly significant, not just from a gameplay factor but also a socio-cultural one. For context: Before the 80s a widely held, though often subconscious opinion held in paleontology and evolutionary biology was a teleological one, meaning that the emergence of human intelligence or at least something very similar to it, was something inevitably programmed into the makings of the universe and the history of our planet. Developments during the 70s and 80s deeply disturbed such a worldview. The Dinosaur Renaissance showed that the dinosaurs were not destined to be replaced by the mammals, but would have just kept on ruling the planet if it had not been for the asteroid. The redescription of the Cambrian Burgess Shale fauna by Harry Whittington and Simon Conway Morris led to the realization that animal life itself started with a far grander availability of body types, but through apparently random decimation at the end of the period became restricted to the few ones we have today. The aforementioned book Wonderful Life by Stephen Jay Gould highlighted these facts and made the, at the time bold, claim that none of the major events of our evolutionary history were predictable or a given from the start. Go back into the Cambrian, accidentally step on a trilobite, come back to modern day again and giant echinoderms might now roam the savannahs of Africa instead of elephants. I do not know if Will Wright ever read Wonderful Life (he was well-read on scientific matters and early concept art of Spore does indicate a strong fascination with the Burgess fauna, so there at least is a strong likelihood), but SimEarth feels almost like a direct response to it. It released just one year after and is the earliest piece of entertainment I can think of which presents such an open-endedness to evolution. Also, the game lets you literally play out Gould’s thought experiment of rewinding the tape of life (do modern readers even know anymore what is referred to here by “tape”?) back to the Cambrian and watching Earth’s history play out again. That is hard to accept as a coincidence.

Fig. 12: The cover of Wonderful Life’s first edition.

From a gameplay perspective all of this means that SimEarth provides something which Spore could not: True historical contingency. Whereas every playthrough of Spore is linear, with the only variables, apart from cosmetics, being your diet and policies, no two runs of a SimEarth simulation will feel alike. Yes, ultimately you only have a limited amount of taxa which can evolve sapience, but how each will get there and how they will run their course will always be a unique story. On my very first playthrough, my world stagnated very early in its evolution, with radiates and arthropods being the height of complexity. As if wanting to take matters into their own hand, the radiates randomly evolved sapience at one point (much to my dismay as I obviously wanted sapient dinosaurs) but got permanently stuck at a pre-industrial level of technology. The reason? Apparently the game accurately portrays the fact that, if not enough diversity and biomass has accrued over the eons, civilizations will later on not have enough fossil fuels to power their industry. On another playthrough I had a healthy biosphere, but a freak increase in volcanic activity heated up the planet, making jungles, swamps and especially insects thrive, which in turn led to a high population of carniferns. I decided to mess around and dropped a monolith on the little audreys (not knowing at the time this would actually work). They did well, until bombing each other with nuclear weapons in the atomic age. On a recent run, a regular, albeit slightly cold planet developed a human civilization. Again they did well, until they reached the information age and started a massive greenhouse effect. The poles melted away and the interior of the landmasses became nearly uninhabitable deserts, with most settlements clustered around the poles where temperatures were still somewhat tolerable. From this sorry state they advanced all the way past the nanotech stage and began their mass exodus from the ruined planet. Once they were gone and the world reverted to an evolutionary timescale, the only taxon which could thrive in this Mad Max ecology were reptiles, who evolved sapience almost immediately after the exodus (perhaps salvaging what was left in desert ruins). Another experience had by a gaming journalist in the 90s was when due to environmental degradation all the continental landmasses suddenly became uninhabitable, but because the civilization he had on the planet happened to descend from amphibians, they were able to cope and simply moved all their cities into the ocean, Gungan-style (Wilson 1991, p. 103). I can confirm that amphibians will indeed build cities on ocean tiles. Another fun run with dinosaur-hating bird people was recently recorded by LGR.

What stories might you be able to tell with SimEarth? Will trichordates make it past the stone age? Will dinosaurs develop space travel? Will carniferns defeat the nanobot invasion? You are god, you decide.

Quick note for worldbuilders

It should be noted that, apart from being a very fun earth history simulator, SimEarth also makes for an excellent worldbuilding tool. Plate-tectonics are realistically simulated and can be observed over geologic timescales and what biome a tile becomes is determined by the temperature and humidity of the area and the planet. If you need to create a natural-feeling fantasy or sci-fi world from scratch, for example when you need to design a tabletop roleplaying campaign, you can simply start up a random planet and watch as it evolves its landmasses and climate zones. Once you have grown something you like, screenshot it or sketch it. You can create some quite interesting geographies through fairly simple adjustments of the game’s sliders. Crank up plate tectonics to the maximum and your landmasses will simply be a series of island archipelagos, shut it down completely and you might end up with something similar to Mars. Messing with the axial tilt will also allow you to create interesting climate zones. Put the tilt at minimum (0 degrees) and you will create an ice age climate, put it at maximum (54 degrees) and all regions on the planet will receive the same amount of sunlight throughout the year, creating tropical climates throughout most of the globe. Using the game’s raising and lowering tools and placeable tiles it is also entirely possible (though probably tedious) to build existing maps, like Westeros or Middle Earth, yourself and then see how natural forces would act upon them.

Fig. 13 – 16: A SimEarth map viewed through the game’s altitude, plate tectonics, sea current and biome filters.

Fear not, for I am with you: How to play SimEarth nowadays

SimEarth released for a wide variety of platforms at the time: MS-DOS, Macintosh, Amiga, TurboGrafx-16, Sega CD and the Super Nintendo Entertainment System. Under the Maxis Classics label the game was re-released in 1996 to be compatible with Windows 95. Gameplaywise they are all the same, but the graphics differ widely, with the DOS-version being pretty crude to look at, while the console versions had nicer sprites (though the game is handled better with a mouse and keyboard). The most recent way you would have been naturally able to play the game was with the Virtual Console for the original Wii, which is how I experienced the game for the first time, where you could download the TurboGrafx-version. But alas the VC was discontinued in 2019 and it is unlikely that the game will reappear on the Nintendo Switch Online service (which I do not recommend anyway; have you seen those prices and the online quality?). So, if you want to experience the game in an authentic way, it really looks like you have to cram out the old Super Nintendo or Amiga and buy a used cartridge or floppy disk online.

But, since SimEarth can pretty much be considered abandonware, there are thankfully other ways in which to play it for free. For example, there are various emulating sites with which you can simply play the game online through a browser:

While easily accessible, this is nonetheless an awkward solution, especially since you cannot save your games. The next obvious option is then of course actual emulation, but I must admit that I am a total dunce when it comes to that. At the very least I managed to emulate the MS-DOS version on my laptop thanks to online guides:

  • Step 1: Download DosBox.
  • Step 2: Download the game’s files here. You can also download the original 228 page manual there.
  • Step 3: Unzip the file and move it to C:\\DosGame\
  • Step 4: Run DosBox
  • Step 5: Type in (without the quotation marks): “mount c c:\DosGame\”
  • Step 6: Type “c:\”
  • Step 7: Type “simearth.exe”
  • Step 8: The game should start now. If it opens in a small window, you can expand to fullscreen by simply tapping Alt+Enter.

Fig. 17: SimEarth as viewed through the prettier TurboGrafx graphics (say that three times fast).

Of course, if you are better at emulating than me, you are free to play any other version of the game. Like I said, the SNES and Turbografx version are recommendable in the sense that their graphics make it a lot easier to discern what is actually happening on your planet, but the game simply was not made to be handled with a controller, which why the MS-DOS version is a lot less frustrating. Judging by the screenshots, the version for the Commodore Amiga would then be a good middle ground.

The fate of SimEarth and software toys

SimEarth, I believe, is to videogames what Harry Whittington’s redescription of Opabinia is to paleontology. It really is quite remarkable how advanced it is for its time and how alone it stands among games today. As mentioned before, according to Lovelock, most professional computer models used by earth scientists at the time did not interlink the earth spheres. Neither did atmospheric simulations take into account the effects of clouds and albedo on the climate (Wilson 1991, p. xii). And yet here we have a videogame which does both. “Strangely these atmosphere models are like SimEarth but less complex.”, to quote the man himself. For fairness’ sake to these climate models, they were still a lot more detailed in how they handled the structure of the atmosphere and in how much data they computed, while SimEarth obviously had to simplify many of its variables and is not an accurate model of Earth. This still does not take away from the fact that SimEarth was possibly the first software that tried to simulate an interlinked planetary system and did so successfully enough to be remembered fondly by scientists years later and perhaps even inspire actual models. To quote astrobiologist David Grinspoon from his book on Venus:

Have you ever played, or at least heard of, the computer game ‘SimEarth’? This clever game lets you ‘play God’ by altering various environmental factors and watching the effects on the evolving Earth system. Like any computer model, the game is only as good as the equations that go into it. It is far too simple to be an accurate simulation of Earth’s evolution, but it does give a feel for the way various components of our environment interact. We have started to construct evolutionary climate models of Venus that include simulations of the surface-atmosphere reactions and the greenhouse effect. In a sense we are trying to create a “SimVenus” that really works.” (Grinspoon 1997).

If something as complex as SimEarth could run on early IBM PCs with MS-DOS, imagine how much more intricate, complex and graphically more appealing a game such as this could be on later or modern technology. But alas, here lies one of the most frustrating and intriguing mysteries. There never has been anything like SimEarth. No sequel, no spiritual successor, no rip-off. Ever. The games most closely comparable, such as SimLife from 1992, only tackled certain aspects of the whole which SimEarth tried to portray. Why is that? Why could, for example, Spore not have been SimEarth 2 in all but name? Multiple factors might have led to this dearth:

  • Even though SimEarth got very high critical scores by reviewers at the time, it was not nearly as popular with gamers or as well remembered as SimCity and other Maxis titles. Most people are probably more likely to remember SimAnt than SimEarth. The probable reason is, ironically, the game’s high complexity (Reminder: The manual is over 220 pages long). This, coupled with its very simplified graphics and somewhat convoluted computer-window-presentation likely made the game seem rather inaccessible for casual audiences, who are not as interested in the subject matters the game tackles. I fear my own enthusiasm I have with the game might be clouded by my own background in earth sciences. I know what albedo is, but did 9-year-old Jimmy in 1990 without widespread access to the internet?
  • The technology behind videogames becoming more complex does not necessarily mean that games themselves become more complex. Games like SimEarth at the time were able to be complex because everything apart from the core had to remain simple out of necessity. Instead of focussing on graphics and sound, most of the work could be put into gameplay and programming, which was also easier to manipulate given the primitiveness of the technology. As such, SimEarth was able to be programmed by just eight people. Technological progress and market forces since the early 90s have, counterintuitively, put the whole gaming industry into many constraints. If a modern remake by a big studio were attempted, its publisher would likely demand the game to be three-dimensional and of high graphical fidelity in order for it to be deemed commercially viable, reallocating significant amounts of workforce that might be needed in other departments. Many of the more complex or demanding features might also be cut in order to not alienate casual audiences. The fact that videogames have become a multimillion dollar industry means that mainstream studios have to direct massive resources into projects in order to be competitive, but this in turn means they cannot allow high risks, as they could in turn face massive losses. This of course stifles experimentation. It is therefore no wonder that, apart from a few holdouts like the Sims, the era of big budget software toys is over, as is that of mainstream god games. The aforementioned forces likely led to the decline of Lionhead Studios (though Peter Molyneux’s own “marketing” strategies may also be to blame, but that is a different story) and to Spore’s troubled production. According to Will Wright, his original concept of Spore, when it was still titled SimEverything, was to simply be a toybox galaxy, where the player flew around in his spaceship, exploring other planets and manipulating their properties and biospheres (If you want to check out a modern game with a similar premise there is Heliopedia). This truly could have become a successor to SimEarth, but the lack of hard goals in the game may have forced him to instead hop to the Stage-model the game has now. The space stage, if you have enough money and tools, does technically allow you to terraform a planet and colonize it with creatures of your own design, but in the end the Spore planets are nowhere near as intricate or fun to play around with as in SimEarth. The game originally also went for a darker, more gritty and generally more biologically grounded design, with the different eye-types your creature has, for example, messing with the resolution of your screen. But eventually the dev team split up into a science-group and a cute-group, the latter of which wanted to direct the game into a more cartoony and videogamey direction, likely to appeal to more audiences. While Wright said the final game reached a good middle ground, the cartoony look, especially when compared to the 2005 demo, was criticized and was likely a major factor in the disappointment people had.
  • If major companies do not want to make a SimEarth-type game, there is still the indie market to look towards, right? Unfortunately, most indie devs in the god game genre want to mimic either the Molyneux style games or directly remake Spore in their image. Some truly fun and promising life simulations have appeared out of the latter attempts, such as S.P.E.C.I.E.S. or Ecosystem mentioned at the start, but since they are often developed by just one person and three-dimensional, they severely limit the scope at which they can simulate and therefore just create limited biosphere simulations more similar to SimLife. To create something like SimEarth again also requires an interest and know-how in the earth sciences. Wright and Haslam had luck in getting James Lovelock as an advisor, but a modern indie dev who is not already as interested on these matters might have to do some extensive research before being able to make a good simulation. Conversely, while they do have to work a lot with programs, I do not know many earth scientists who are interested in making videogames, even though such projects would be great for public outreach and education.

The ultimate god-game?

I admit, I titled this post provocatively on purpose. Is SimEarth the ultimate god game? Probably not. There likely can never be a perfect god-game (or probably game in general), due to how vaguely defined this genre actually is and how much tastes vary. For other people, Black & White 3 might instead be the ultimate god-game and that is a totally valid perspective. There is also some fair criticism for SimEarth, apart from just the somewhat confusing presentation. While it offers so many variables to play around with, it ultimately still restricts itself to Earth and its components. Creating truly alien worlds is not possible and the game’s evolutionary progression is pretty straight and goal-oriented. Its simulation of civilizations is very simple as well. All of the taxa in the game are from Earth’s biosphere (though I imagine the trichordates were included to give you at least some raw material for slightly more alien worlds) and you cannot directly influence their biology. This is where SimLife from 1992 and of course Spore are better. SimLife especially seems like a final missing component of SimEarth as the game allowed you to make your own critters construction-kit-style, letting them loose in the ecosystem and then watching how they would survive and adapt. A perfect mix of the two, as well as Spore, a game where you could theoretically make, for example, something akin to Wayne Barlowe’s Darwin IV, might not be the perfect god-game, but it might scratch the itch of a generation.

The ultimate god-game - A modest proposal:

I do not know much about programming, and I have never made a game, but that does not really stop me from dreaming and writing a concept based on what we have learned so far. My hope is that someone more capable than me might read this someday and hopefully get inspired.

 Fig. 18: My good friend Bob Guan was nice enough to illustrate for me what I am roughly imagining (Thanks again for that). You may know him better from our "little" side-project.

For the groundwork, I imagine the game to be two-dimensional and tile/grid-based like SimEarth (though of course still with prettier sprite graphics), as I imagine this would make things a lot easier to program. Unlike that game, I imagine the tiles to be hexagonal instead of square. At the start of a game, all land-tiles initially start out as either wasteland or desert, but will change into new biomes depending on what vegetation you introduce, as well as the humidity and temperature. Like in SimEarth, the multiple systems of the planet work interconnectedly, such as the atmosphere, lithosphere, hydrosphere, plate tectonics, biosphere and so on. However, in the main mode, the “campaign” if you will, you are the god of life and only in charge of the biosphere, trying to hold it up against the ever-changing forces of nature. You start with a planet that has high volcanism, no oxygen, a lot of CO2 and only a single lifeform: an extremophile prokaryote that lives in deep sea hydrothermal vents. As the population of these bacteria grows, it starts accumulating mutations. These mutations are counted as mutation-points, which are the main currency of the game. The larger the populations are and the faster they reproduce and make copying errors, the more mutation-points are produced. Once you have enough mutation-points you can open up a sort of “tech-tree” (similar to ones from historical strategy games like Civilization), but instead of technology you unlock biological adaptations. Let’s say your first bacteria-population generates 50 MP. These you then use to unlock photosynthesis in the prokaryote tech-tree. You then open up a two-dimensional creature-editor and edit your existing bacteria by adding a new plasmid to create a new lifeform (you can choose whether your new species replaces the old one or lives alongside it) that lives through photosynthesis. This lifeform you then place in a new environment, the open ocean. If your new photosynthetic lifeform thrives, it will generate oxygen and terraform your planet’s atmosphere. If you are not careful though it uses up all the CO2 and your planet will freeze over, so you soon thereafter have to produce aerobic organisms that feed on the bacterium and so on. This is just an example of how early gameplay would work. After you accumulate more MP you will unlock more taxa, like eukaryotes, fungi, plants, animals or something completely alien like vendobionts, which all have their own creature-editor, parts and tech-trees (viruses could also be in the game, but they would be generated by the computer and live off your creations, so you have a sort of adversary). You will have to colonize and design all available ecosystems, such as the mountains, deep sea, shallow sea, open sea, desert, rainforest, tundra, etc. with your creations. Depending on your difficulty setting your planet will be hit with disasters throughout the eons, like meteorite impacts, supervolcanic eruptions or even alien invasions, which disrupt the ecosystems. You either have to adapt your creations for these crises or create new ones out of the survivors. 

The lifeform editor I imagine to be similar to the one in Spore, but again two-dimensional (such a thing did exist once). Instead of giving stats like strength and defence, organs would instead add or deduct points to environmental preferences, niches and fitness (as well as environmental impact in case of the prokaryotes). Changing the skin-toggle from scaly to furry might for example deduct two points from desert biomes but allow the new creature to settle in arctic environments. Fitness is a stat that would define the creature’s reproductive success, efficiency and general competitiveness and would be determined by body parts such as mating displays, armour, tooth-size etc. A creature’s niche would be determined by its size and diet/mouth-parts.

Unlike in SimEarth or SimLife, each biome-tile could be occupied by multiple lifeforms, as long as they do not have the same niche. In order to not lose visual track of specific creatures, the game should allow you to apply specified filters to the map so you can look at the ranges of individual species. Creatures of lower trophic levels on the tile will feed those above them. In real life only 10% of a trophic level's energy is converted into the biomass of the trophic level above it, so in simplified in-game terms, if you have 10'000 plants on one tile this would be enough to feed 1000 herbivores and 100 carnivores. If two species of the same niche compete for the same tile, their fitness stat will determine who outcompetes the other from the space.

Of course, there should also be a pure sandbox mode, where you can take full control of each sphere of the planet. In this mode, natural selection should also be enabled, where the computer itself randomly unlocks nodes of the biotech-tree and mutates creatures, so you can just sit back and watch the game play itself as a software toy.

After you or the computer have created a stable ecosystem in all your planet’s habitats and unlocked the tool-using brain in the animal “tech-tree” (although I would not be against the idea of intelligent plants or fungi in the game), what is next? The game could simulate the development of civilization itself and you influence it through indirect means. More fun perhaps might also be if the game turned into a turn-based or real-time strategy game at that point, where like the god of the Old Testament you guide a single tribe through many perils, from prehistory into the information age and beyond, sort of like Empire Earth from 2001. The long evolutionary history you have formed on your planet would then have direct influence on your progress. If, for example, you never managed to evolve fruit-bearing plant species, your culture might have a problem with finding crops suitable for domestication and be permanently stuck in the paleolithic or resort to nomadic animal husbandry. Or if you happen to have evolved sufficiently large megafauna shortly before entering prehistory, your later iron age cultures might have access to howdah-bearing beasts of war.

 Fig. 19: Mammals never had the makings of a varsity megafauna.

I hope this whole read has been engaging to you and you and that people might walk away with a renewed interest in this old but still "clever game".

If you liked this and other articles, please consider supporting me on Patreon. I am thankful for any amount, even if it is just 1$, as it will help me at dedicating more time to this blog and related projects. Patrons also gain early access to the draft-versions of these posts.

Related Posts:

Literary Sources:

  • Gould, Stephen Jay: Wonderful Life. The Burgess Shale and the Nature of History, New York 1989.
  • Grinspoon, David Harry: Venus Revealed. A New Look below the Clouds of our Mysterious Twin Planet, Cambridge 1997.
  • Lenton, Tim: Earth System Science. A very short introduction, Oxford 2016.
  • Wilson, Johnny: The SimEarth Bible, Berkeley 1991.

Online Sources/Further Reading :

Image Sources:


  1. This is a post with really interesting ideas. My favourite god game has always been SimEarth and I'm also disappointed that there's no spiritual successors to this game. Out of all the games made in the last 30 years, we'd think there would at least be one, alas no. As I always wanted to do this for years, I started a SimEarth-inspired game a few months ago (named Stella Dei), the time to amp up my programming skill.
    And you're right, I really get why it hasn't been done much before, the amount of physics and geological processes I had to learn is so high I'm close to saturation hehe. But I expect to continue as I've never felt that motivated to work on a game.
    On my "sequel" graphics, I chose to do a 3D planet simply because I felt SimEarth in 3D would be interesting. I'm still thinking about how to do life and how to visualize sentient trichordates, but one of your ideas I really liked is the "tech-tree" which could be very interesting if done right. Right now, the game is an early prototype, but as I implemented rainfall in a few days, I can tell I'm going at a quite fast pace. I can't wait for some people to feel the same awe when playing Stella Dei as I did playing SimEarth.

    1. I am really glad this post was able to inspire your project a bit. If you ever need input, help with research or maybe even a playtester, you know who to contact.

    2. Zen1th, I would like to contribute to your project, if that is something you're willing to consider. I have been wanting to get people together on a SimEarth-esk remake for many years.
      I'm intermediate with Python, and my background is GIS/Geography/Geology. Earth/Planetary sciences have always been a strong interest of mine and I think I can bring a host of knowledge to your effort, if you'll have it. Let me know your thoughts on that:
      roadtripguy at g mail dot com

  2. Great post :) it's good to see classic games like this getting appreciation, especially from people with specialized areas of knowledge today. I've always been interested in city planners, systems managers, various STEM type jobs etc, discussing simulation programs that model their field of study.

    It would be neat to read your thoughts on other games and software. although sadly, it seems like a large gap in the market.

  3. Wonderful. SimEarth is a gem, isn't it? You'd expect to see more iterations and advancements on it, what with computers being literally millions of times more powerful, and yet... A host of small-scale independent project, at least. Even I'm trying my hand at one, though facing the teeny tiny difficulty that I actually know next to nothing about programming. I'd like to see one devoting more detail to the evolution of species, perhaps replacing the hardcoded list of organism groups (my least favorite aspect of the game) with the "tech tree" of organs and adaptations that you mention, but that each species would have to research independently (at least from their splitting point). Each tile would contain a number of ecological niches for which the species would have to compete, using each other as resources, based on their evolved attributes. Perhaps in several years I will be able to turn it into a reality.

  4. I play SimEarth at least once a year, usually on an emulated SNES port (quite a bit easier than the PC version). I set it up on the TV and just let it run for a few hours. The possibilites of what can happen still fascinate me to this day. No other god simulation ever quite captured it like SimEarth did. I really hope it gets a proper remake someday.