Artificially generatedThe Longest Game
Forty years of economic simulation, and why the recreations in particular have survived.
There was once a genre that wrote its manuals like textbooks. Anyone who wanted to play a shipping company in 1992 got a historical overview of imperialism and a lexicon of nautical terminology with it. These games are almost all dead. Their free recreations live on, on average for 24 years. Why this is no coincidence, how deep the economic models really go today, and where exactly the boundary between business simulation and game is currently disappearing.
A genre that came from the German-speaking world and made no apologies
In 1986, Rolf-Dieter Klein and Martin Ulrich built Ports of Call, a shipping company simulation for the Amiga. You buy ships, look for cargo, calculate fuel against demurrage, and finally maneuver into port by hand.
In 1987, Ralf Glau published Vermeer with Ariolasoft. A dying Berlin art collector, lost paintings, and to buy them back you run plantations for coffee, tobacco, tea and cocoa. Together with his Hanse, this is now considered a milestone in German game development.
In 1992 came 1869 from the Austrian studio Max Design, named after the opening year of the Suez Canal. And Der Patrizier from Gütersloh, from a studio called Ascon, which later renamed itself Ascaron.
What connects these titles is not the graphics. It is the attitude of their manuals.
Max Design called its series Erlebte Geschichte and included with 1869 a historical overview of imperialism from 1854 to 1880, plus a lexicon of nautical terminology. This was not a package insert, this was teaching material.
These games did not pretend that economics was simple. They assumed that someone who voluntarily wants to run a shipping company also wants to know what a bill of lading is. This is an assumption about audience that almost no one makes today.
And it worked. German-language economic simulation became its own genre with its own reputation in the nineties.
The three longest bars belong to no one
We looked at sixteen titles to see over what period work was actually done on them or their direct continuation. From first release to last documented edition.
The result is clearer than we expected.
Yellow are the three open-source recreations, gray the commercial titles. A bar extends from first release to last documented edition, sequels included. Very short bars are titles with exactly one edition.
| Original | maintained | Free recreation | maintained |
|---|---|---|---|
| Transport Tycoon 1994, Deluxe 1995 | 1 year | OpenTTD, since 2004, version 15.3 from April 2026 | 22 years |
| Civilization I, 1991 | – | Freeciv, since January 1996, version 3.2.5 from July 2026 | 30 years |
| Colonization, 1994 | – | FreeCol, since 2003, version 1.0 after twenty years, most recently 1.2.0 in July 2024 | 21 years |
A good ruleset finds people to maintain it
OpenTTD is the free recreation of Transport Tycoon Deluxe. The original was maintained for one year, then the company moved on. OpenTTD has been released since 2004, the current edition is from April 2026.
Freeciv is the free recreation of Civilization. First edition in January 1996, exactly thirty years ago, current edition in July 2026.
FreeCol recreates Colonization and took twenty years to reach version 1.0. This is not a criticism, this is the point: There was no deadline.
The mechanism behind this is the same one we have already described in another context: a ruleset spreads through the cost of its first correct implementation, and it survives through the number of people allowed to implement it.
A commercial game may be maintained by exactly one company, and it does so as long as it pays. An openly recreated ruleset may be maintained by anyone, and it pays for no one, which is why no one stops.
That these three rulesets were recreated and not others is the real distinction. No one recreates a mediocre system for twenty years. The list of free recreations is a best-of list that no one compiled.
No price is fixed anymore, each follows from something else
The question of which game has the most complete economic system is often answered with titles. More useful is the question where a price comes from. This reveals the depth of a model, and the answers today are astonishing.
In Workers & Resources: Soviet Republic there are no fixed prices. The price for steel results from the price for iron, the price for coal, the price of labor, and proportionally from the construction costs of the steel mill. Anyone who buys or sells in large quantities shifts the market price themselves.
The titles are not comparable because they make different things deep. That is precisely the interesting observation.
| Title | Where the depth lies | The trick |
|---|---|---|
| Victoria 3 (2022) | Hundreds of goods, from grain to automobiles. At the center are not buildings, but Pops, meaning population groups with income and standard of living. | Prices emerge from buy and sell orders. All orders are fulfilled, the difference creates or destroys in the simulation. |
| Workers & Resources (2019/2024) | Complete production chains from crude oil to consumer goods, plus citizens with education, work, party loyalty. | No fixed prices. Each price is calculated from its precursors, own trade volumes shift the market. |
| Anno 1800 (2019, with expansions) | Production chains from hundreds of components across three game worlds. | Depth through chaining rather than price formation. The feat is the logistics, not the market. |
| EVE Online (2003) | An economy entirely sustained by players, continuously for over twenty years. | The operator employs economists. For the successor EVE Frontier, a former economist from the Icelandic central bank was hired as head of economy. |
There is exactly one game that calculates economy and nature in the same model
The question of whether economic simulations are connected to an ecosystem anywhere has a surprisingly clear answer: yes, specifically in Eco by Strange Loop Games.
There, multiple players jointly build a civilization in a world whose ecosystem consists of thousands of plant and animal species and continues around the clock. All resources come from this ecosystem. Every action affects it.
What is remarkable is not the ecology alone, but the entanglement: the incentives of the economy drive groups with actually identical goals into financial conflicts whose consequences can destroy the world. The collapse is not a script, it is an outcome.
And the economic side continues to be expanded: rental market, mortgages, stock markets, and building on that, forms of government up to a central bank. Leading a functioning national economy is intended to become one of the main tasks.
Business buys the same thing, just without fun
While the genre became smaller in entertainment, it has become a market in companies that hardly anyone talks about.
Serious Games, meaning simulations for training and decision-making purposes, are at 14 to 21 billion dollars in 2026 depending on how you count, and are expected to roughly double to triple by the early thirties. The sub-market of pure business simulations was at 2.5 billion dollars in 2024.
Alongside this, and counted separately, the market for digital twins with generative AI is growing from 9.8 billion dollars in 2026 to an expected 32.5 billion in 2031.
Both markets want the same thing: a model on which you can try out decisions before you make them. They just come from different directions, and they don't meet.
The twin has the data and no loop. The game has the loop and no data.
A digital twin knows everything about a company and displays it. It is a very accurate dashboard of what has already happened. You cannot play it.
An economic simulation has the loop: decision, consequence, next decision, and fast enough that you get a feel for the system. It is only about an invented company.
In between lies a gap that remained open for forty years for a single reason: building a playable model of a real company was more expensive than the benefit.
Precisely this effort has just collapsed. Generating a model from existing data, attaching it to an interface and making it playable in a loop was a project year and is now a work order.
This means the question "what if" is no longer a presentation, but a game.
Four sentences to get the picture right
First: our timeline is a selection, not a complete survey. Sixteen titles that appear in this report. Anyone who chooses others will get different averages. However, the distance between 24.3 and 9.9 years is so large that a different selection would hardly reverse it.
Second: the market figures are estimates by market researchers and differ from each other by a factor of one and a half depending on the firm. We therefore state the range and not a single number.
Third: free recreations are not better software. They are longer-maintained software. That is something different and sometimes less valuable to the player, but precisely the point for the question of the durability of rules.
Fourth: we are partisan here. McGrinsey is building at exactly the boundary that section 07 describes. This does not make the market figures wrong, but it explains why we notice this gap. Anyone who wants to hear the same finding from someone who has nothing to gain from it should keep looking.
If you want to get to know the genre
Start with OpenTTD. Free, runs everywhere, and the ruleset is so good that people have been maintaining it for twenty-two years. After that Freeciv for the long line and FreeCol for the trading side.
If you're looking for depth
Workers & Resources for pricing from precursors, Victoria 3 for markets with population groups, Eco for the entanglement with an ecosystem.
If you're working on models
The most useful question for any economic model is: where does a price come from? If it's set, the model is a table. If it follows from something else, it's a simulation. There's nothing in between.
If you're making decisions in a company
For every dashboard, ask whether you can try something out with it or just look. A tool without a loop is a report, no matter how much data is in it.
Everything this report is based on
Nine sources and one proprietary analysis. The three free projects are directly linked because you can download them and check for yourself. For a report about durability, that's the most honest form of citation.
- OpenTTD, the project itselfFree remake of Transport Tycoon Deluxe, since 2004, current version 15.3 from April 4, 2026, GNU GPL. The evidence for the 22 years is the release list itself.openttd.org
- Freeciv, the project itselfFree remake of Civilization, first release January 5, 1996, making it thirty years old, stable version 3.2.4 in March 2026, bugfix 3.2.5 in July 2026.freeciv.org
- FreeCol, the project itselfFree remake of Colonization, since January 2, 2003. Version 1.0 was reached in 2023 after twenty years of development, most recently 1.2.0 in July 2024.freecol.org
- Ports of Call, manufacturer website and Steam entry1986 by Rolf-Dieter Klein and Martin Ulrich for the Amiga, published by Aegis. Still being sold forty years later, as Classic on Steam as well as in XXL and browser versions.portsofcall.de
- Wikipedia on the Vermeer series and on 1869Vermeer by Ralf Glau, Ariolasoft 1987, reissue 1997, successor 2004. 1869 by Max Design 1992, named after the opening of the Suez Canal, part of the Erlebte Geschichte series with historical outline and specialist encyclopedia in the manual.de.wikipedia.org/wiki/Vermeer_(Computerspielreihe)
- Workers & Resources: Soviet Republic, description of the economic systemThe source for section 04 and Fig. 02: "There are no fixed prices." The steel price follows from iron, coal, labor and proportionally the construction costs of the plant, large own trading volumes shift the market price.store.steampowered.com/app/784150/Workers__Resources_Soviet_Republic/
- Paradox: Deep Dive, Modeling the global economy in Victoria 3The developer report on the economic model: hundreds of goods, population groups at the center, pricing from buy and sell orders, where all orders are fulfilled and the simulation creates or destroys the difference.gamedeveloper.com/design/deep-dive-modeling-the-global-economy-in-victoria-3
- Strange Loop Games: Eco, Economy as GameplayThe basis of section 05. An ecosystem of thousands of plant and animal species that continues around the clock, intertwined with a player economy. Planned are rental market, mortgages, stock markets and based on that a central bank.developland.strangeloopgames.com/economy-as-gameplay/
- Game Developer: CCP hires a central bank economistThe EVE operator has hired Stefán Þórarinsson, formerly economist at the Icelandic central bank, as head of economy to treat the player economy of EVE Frontier with the same care as a real national economy.gamedeveloper.com/business/ccp-hires-new-economy-head-to-legitimize-eve-frontier-s-in-game-economy
- Market figures for Serious Games, business simulations and digital twinsThe figures from section 06, deliberately as a range. Serious Games 2026 depending on source 14.6 to 20.6 billion dollars, business simulations 2.5 billion (2024) with expected 6.7 billion by 2033, AI digital twins 9.8 billion (2026) with expected 32.5 billion by 2031.mordorintelligence.com/industry-reports/serious-games-market mordorintelligence.com/industry-reports/generative-ai-in-digital-twin-simulation-and-scenario-modeling-market
Proprietary analysis McGrinsey: The timeline in Fig. 01 and the averages of 24.3 and 9.9 years are calculated from the release and publication dates of the sixteen titles mentioned. This is a selection, not a complete survey, and for series the span counts to the last release, which biases in favor of the commercial side. Fig. 02 shows a mechanism, not game values. Market figures are estimates from various market research houses and are given as a range where they differ from each other.
This report has told the story of a genre and in doing so brought back a finding we hadn't been looking for: that rule systems outlive their owners when you release them. This is the same mechanism that also decides about protocols.
Games 02 goes to the boundary from section 07 and gets concrete: What belongs in a playable model of a real company, what must you leave out, and how do you recognize that the model is lying?
Related: Games are eating the world on the larger thesis, and Energy 02 as an example of what a model looks like whose numbers we measured ourselves.
