It is 1865. Britain runs on coal: its mills, its engines, its ships. Every year it burns more than the year before. The dotted line is what happens if that goes on.
All the coal Britain has is in the ground, drawn here as black squares. Burn it fast and it makes a big fire that runs out. Burn it slow and it makes a small fire that lasts. You decide the pace, one generation at a time, for ten generations; a ledger keeps what each one got, and the score is how well the typical generation did.
Choose for Britain: a brief greatness, or a long mediocrity.
Choose for Britain: a brief greatness, or a long mediocrity.
It is graph paper, ruled like Jevons's own. The years run along the black line in the middle, from 1850 to 2070.
Above the line is Britain's fire: how much coal it burns in a year. The higher the curve, the bigger the fire. The black under the curve is coal that has been burnt, and you can count it in squares. If the fire grows taller than the paper, more paper is added at the top.
Below the line is the coal still in the ground: 167 black squares, in eight layers, down to 4,000 feet. One square below is the same amount of coal as one square above. When a square is burnt above, a square goes from below. The top layer is dug first. The deeper the digging, the dearer the coal.
The dashed line is America's coal. The hatched bar is people who went over the sea because there was no work for them. Orange is whatever is burning now.
The ledger at the bottom says what each generation got: its income per head (the bar and the big figure) and Britain's whole income (the small figure), both against 1865. Income comes from power, and power comes from coal and from how much work a ton of it does. Heads are all the people Britain would have had, those who went over the sea counted in. The last box, after, is what the coal still in the ground in 2065 could give ten more generations, burnt evenly.
You set ten turns of twenty years, from 1865 to 2045. Slide along the ruler under the sheet to say how much coal Britain should burn twenty years from now: cut back (half as much), a stand (the same), slower (half as much again), as now (twice as much: Jevons's own pace in 1865) or faster. Under each mark the ruler says how long the coal would last at that pace. The dotted line on the sheet is your plan, and the crossed squares are the coal it would take. Then press Burn. The solid line is what actually burnt, and it may leave your dotted line.
In 1885 and 1925 a new engine appears that does the same work on less coal, but costs more to build. It pays where coal is dear, and it gets cheaper every year. Britain's works fit it as soon as it pays for them; America fits it when it has become cheap enough. Nobody chooses. Watch what the fire does when an engine comes in.
The price of coal is told as the days of a year's earnings that fuel takes: about 18 days in 1865. Deeper pits make it more; a better engine makes it less. Works go where fuel takes fewer days.
If coal has grown so dear that the fire could not grow at any pace, the evening has come: the fire falls whatever you set, and the rest of the sheet draws itself.
Your score asks: what even share, given to every generation alike from 1865 to the ones after 2065, would be worth as much as your sheet? A rich generation cannot make up for a poor one: we count so that ×4 for one generation and ×1 for the next is worth no more than ×2 for both. Your even share against the best path's, out of 100, is your score.
The last page shows the best path we found, dotted on your sheet and in hollow bars in your ledger, and says why it goes as it does. Then the curve, continued, and why it stopped.
W. Stanley Jevons, The Coal Question: An Inquiry Concerning the Progress of the Nation, and the Probable Exhaustion of Our Coal-Mines, second edition, revised (London: Macmillan, 1866; first edition 1865). The game is built from the preface, chapter I (Introduction and Outline), chapter VII (Of the Economy of Fuel), chapter XVIII (Concluding Reflections) and the frontispiece. The full text can be read at the Online Library of Liberty (econlib.org). Game built from this text; the text is the authority where the two disagree.
Jevons argued that Britain's use of coal was multiplying at a steady rate (three and a half in the hundred a year), that the coal within reach was fixed, and that exhaustion would come not as an empty cellar but as deeper pits and dearer coal, until manufactures went "to where fuel is cheapest". Better engines would not spare the coal: "the very contrary is the truth". A growing nation could not be brought to a stand without "a great exodus of our people". So, he wrote, "we have to make the momentous choice between brief greatness and longer continued mediocrity." The game puts that choice in your hands, one generation at a time, on graph paper like that of his frontispiece.
The trade-off. The coal in the ground is fixed: 167 squares on the sheet. Every square of fire above the line is a square gone from below. So the shape of the fire is the only thing Britain chooses: a fire that is high is a fire that is short, and a fire that lasts is a fire that is low. The ruler shows both sides of the bargain at every mark: how big the fire gets (above) and how long the coal would last at that pace (below). The verdict reads your sheet in the same two measures, how high and how long, and ranks neither, because Jevons does not.
A toy. The game runs a small model built to illustrate Jevons's argument, set to the numbers his chapters give. Underneath it is one production engine: power is the coal burnt times the work a ton does (one in 1865, two with the air-engine, three when it is perfected); income, what is now called GDP, is the people times power per head to the power 0.7; fuel's share of income, told as the days of a year's earnings that fuel takes, is the price of coal times the coal burnt over the income, anchored at a twentieth in 1865; works go where power is cheaper; an engine is fitted when the fuel it saves is worth its cost; and the people's increase follows income per head. Everything the game says about days, income and who got what is read off this engine.
It demonstrates that the argument can work as he describes it. With a fixed store of coal, a higher fire means a shorter one. An engine that does more work per ton raises the coal burnt. Holding the fire level sends people over the sea. A rival with cheap coal passes you whatever you do. The end comes as dear coal, not as no coal.
It assumes everything the toy had to choose: the production engine's form and its two anchors (0.7, a twentieth), how fast price rises with depth, how strongly works follow cheap power and how fast they move, how big the rebound from a better engine is, what an engine costs and how fast it cheapens, how the people's increase follows income and how many of those without work emigrate, America's growth, and the coal lying evenly to 4,000 feet. Under this game's assumptions the shapes on your sheet and the figures in the ledger follow; under others the sizes would differ, and the far generations' income per head most of all, because it rests on a toy population two centuries long. The directions were checked to hold when each of these choices is changed.
The text establishes the argument and its numbers: the rate of 3½ per cent and the frontispiece's projection (2,607 million tons by 1961), "rather more than a century" at that rate to reach 4,000 feet, Hull's estimate of the coal within that depth (83,544 million tons, which Jevons adopts; the figure is in chapter II), the engine figures of chapter VII (4 lb of coal per horse-power an hour in a good engine, about half in an air-engine, under a third when perfected), and the reasoning. It does not establish any one future, and Jevons says so.
Turns of twenty years, ten of them, from 1865 to 2065, so that a fire that runs out is seen to run out and every generation scored is the player's; after the evening, the rest is drawn at the last pace set. The engines: the air-engine exists from 1885 and the perfected one from 1925; each pays where coal is three times the 1865 price when it appears, and gets 4 per cent a year cheaper to build, so that it pays at the 1865 price about 27 years later, which is when America, whose coal stays cheap, fits it (the preface: "a wasteful engine pays better where coals are cheap than a more perfect but costly engine"). Britain fits it as soon as it pays there, which is earlier only if Britain's coal is dearer. The price of coal: flat at first ("no recent rise of importance"), twice the 1865 price when about three-fifths of the coal within 4,000 feet is gone, and without limit near the bottom ("we cannot get to the bottom of them"). The pull of cheap power: the trade Britain keeps falls with the price of British power against American, to the power 2½. The rebound: twice the work per ton brings about 1.4 times the coal within the generation. Chapter VII's own magnitudes are larger (Scottish iron: a third of the coal per ton, ten times the coal), but they mix in ordinary growth. The people: 30 million in 1865, increasing 1.2 per cent a year while income per head holds, more slowly as it falls (the preface's "prevalence of marriage" following trade), and about one in seven of those the trade cannot employ go over the sea while the rest "sink down into poverty". America: a fifth of Britain's coal in 1865, growing about 6 per cent a year, always cheap, and fitting every new engine at once. The words noon and evening are the preface's ("we are now in the full morning of our national prosperity, and are approaching noon"); our noon is the top of your curve, and our evening is the point where the trade could not grow at any pace. The shapes are named in Jevons's words when the noon is at least five times the fire of 1865 (greatness) or at most twice it (mediocrity).
Left out: the household grate, a third of the coal and the one saving Jevons says is real ("laid up for the use of posterity"); substitutes, imports and iron (chapters not given); the two parties of chapter I by name; the preface's duties (schooling, the end of child labour, the national debt), which are read on the last page rather than played.
Income per head in the ledger is per head of all the people Britain would have had at the census rate, those who went over the sea counted in. Per head of the people at home it would be higher on every path that sends people away, and the best path would then be to send them away first; counting them in removes that. The last box of the ledger, after, values the coal still in the ground in 2065 as if it were burnt evenly over ten more generations, with the engines of 2065, for the people of 2065: what the generations after the sheet could have from it.
The score is our choice, and the one place the game ranks what Jevons would not. It is an even share: the income per head which, given alike to every generation from 1865 to the ones after 2065, would be worth as much as your sheet. A rich generation cannot make up for a poor one: ×4 for one generation and ×1 for the next is worth no more than ×2 for both. In an economist's words this is the geometric mean of the eleven boxes of the ledger, the equally-distributed equivalent of a planner with log utility who counts every generation the same and discounts nothing. Your score is your even share against the best path's, out of 100. The best path is searched over the ruler's five marks (optimum.js), so it can be told in the ruler's words and followed: faster, then slower, then a stand three times, then slower twice, then faster twice, then as now; it gives an even ×1.5 and leaves 21 squares. A steady "slower" scores about 90, "as now" about 70, a stand about 56, "faster" about 52. A planner who weighed the present more, or the far generations more, would rank them differently.
The best path is a plan for the engines that come. Without the two new engines the best even share falls from ×1.5 to ×0.9; with them a generation earlier it rises, and the best path's noon moves to 1913; a generation later, its pause and its second push move later too. The shape (burn hard early, hold, burn again for the many later people) survives without the engines: it comes from the cheap shallow coal at the start and the many people at the end.
The last screen ends with a button, Why it stopped. The coda draws all of Britain's energy from 1965 to 2024, by source, counted in the tons of coal it would take (Energy Institute data published by Our World in Data, non-fossil sources by the substitution method, converted at the same factor as the coal curve); before 1965 the data has no breakdown, and coal is drawn alone. It also searches the best path again with a substitute for coal, our choice, that gives every generation from a given year (and the box after) at least twice the income per head of 1865: from 1965 (about when nuclear power and North Sea gas came), the best path burns faster in nearly every generation and leaves 2 squares; from 2025 (about when wind and sun became the cheapest power), it leaves 7. This is the economist's backstop result (Nordhaus, 1973): the best rate of using a store depends on when its substitute comes. The dates: Calder Hall, 17 October 1956, the first nuclear power station to feed a national grid; North Sea gas ashore at Easington from the West Sole field, March 1967; Ratcliffe-on-Soar, Britain's last coal power station, closed 30 September 2024. "The cheapest power there has ever been" is the International Energy Agency's World Energy Outlook 2020, of solar power where it is well placed and cheaply financed.
The ledger's boxes from 1985 on say what the toy gives the generations after the player's turns, and they carry a line: supposing, as Jevons did, that no substitute for coal is found. He supposed it in chapter I ("no probability that when our coal is used up any more powerful substitute will be forthcoming"), and Tyndall's letter in the preface supposes it harder. The real curve on the last screen fell for other reasons: oil, gas, nuclear power, renewables, imports, and the end of heavy industry in Britain, with about two-thirds of Hull's coal still in the ground. The sheet knows none of them.
It shows the coal raised in the United Kingdom each year from 1865 to 2024 on the same sheet, with the ground emptied by it at the sheet's own rate, and the date on which you play. The figures before 1900 are the standard series (Mitchell, British Historical Statistics). From 1900 they are the Energy Institute series published by Our World in Data, converted to tons at the factor that gives 225.2 million for 1900 (1913: 287 million, the peak). The game says nothing about them.
Bodoni Moda (Owen Earl, SIL Open Font License). Built with Claude from the text under the paper-to-game skill; the mechanism, its calibration check and the design memo are in the repository (steffenhuck/paper-to-game, folder jevons).