For decades, a small, worn rock, etched with a pattern of rectangles and lines, sat quietly in a Dutch museum, dismissed by experts as a broken piece of architectural salvage. Now, a study in the journal Antiquity proposes that this unassuming artifact was actually an ancient Roman board game.
By employing artificial intelligence to analyze wear patterns on the 8-by-6-inch chunk of limestone and running thousands of simulations, researchers believe that they have brought a forgotten pastime back to life, rediscovering how it was played and solving an archaeological mystery in the process.
Even outside the Colosseum, ancient Romans enjoyed a good distraction and they regularly played tabletop games. Tracing that history is challenging, however, because many games were played on sketches in the sand, or with perishable sticks.
Yet historical writings and archaeological remnants suggest a sophisticated culture of play. Games included rota (resembling tic-tac-toe), ludus calculorum (similar to the Japanese board game Gomoku), and ludus latrunculi (a precursor to chess).
Boards of ludus duodecim, akin to backgammon, have been discovered inscribed with bits of Roman trash talk such as “Idiota recede,” Latin for “You idiot, get lost!”
“Romans made a lot of objects for play and we happen to have a fair number of texts,” said Walter Crist, an archaeologist at Leiden University in the Netherlands and lead author of the new paper.
In 2020, while visiting the Het Romeins Museum in the Dutch city of Heerlen, formerly the Roman town of Coriovallum, Crist, a specialist in ancient games, was drawn to the oval-shaped slab. The object, discovered in the late 19th or early 20th century, featured a peculiar pattern.
Intrigued by a geometry he did not recognize, Crist examined the surface and noticed distinct signs of wear — a clue that the artifact may once have been a well-used board game.
“Players had, over and over again, pushed game pieces across the rock, indicating it was far more than mere decorative graffiti,” Crist said.
To break the code, researchers used high-resolution 3D scans that mapped microscopic imperfections on the limestone board, revealing telltale grooves. “Those line indicated the most heavily played routes,” Crist said.
The data was then fed into Ludii, an AI platform that had been trained on roughly 100 ancient Roman board games. By generating dozens of potential rule sets and pitting bots against each other in mock matches, the AI pinpointed a few specific, engaging variants that conformed to the physical wear pattern.
Programmers ran 130 rule combinations, testing thousands of scenarios against the grooves on the board. Crist said the simulation successfully matched the rub marks to nine plausible sets of rules, indicating that the stone provided a playing surface for a blocking game that required trapping an opponent in the corners of the board. Gamers can now test these strategies themselves by playing this once-forgotten battle of wits against a bot on Ludii.
Now called ludus coriovalli or the Coriovallum game, the surprisingly modern, asymmetrical recreation challenges previous ideas about the history of gaming.
Games that require complex blocking strategies were thought to have emerged much later, in the Middle Ages. Crist said the reconstructed game likely involved two players sliding small pieces of glass, bone, or earthenware along the lines, with one party trying to immobilize the other.
Veronique Dasen, an archaeologist at the University of Fribourg in Switzerland who was not involved in Crist’s research, said the study tackled an important problem in the archaeology of games: how to infer rules from material remains. She noted technology’s potential to uncover hidden history, suggesting that many other unidentified items found at Roman sites could also be game boards waiting for their rules to be rediscovered.
Not everyone is convinced by the study. Ulrich Schadler, an archaeologist at the University of Fribourg and director of the Swiss Museum of Games, is skeptical of the notion that the curiously etched stone is a genuine Roman relic. Unlike, say, typical, crudely scratched ludus latrunculi game boards, he said, the Heerlen artifact features straight-edged, noncoherent lines, making it a total outlier.
Does ludus coriovalli represent a breakthrough or just a bored stonemason’s doodle? Schadler has bet on the latter, arguing that our deep knowledge of Roman culture makes this oddment a product of modern projection rather than historical proof. “Draw a geometrical pattern and I will invent a game to be played on it,” he said.
Crist countered that the artifact’s uniqueness isn’t a flaw, but the very reason for its study. Unlike the ornate playthings of the elite that have been discovered in what were once the Roman Empire’s urban centers, this heavy stone board, he said, represented a lost local tradition on the empire’s fringes that had escaped the notice of Roman historians.
Just because a game wasn’t written down, doesn’t mean it wasn’t played, Crist said. “We therefore cannot rely solely on this method to say a game may have existed,” he added. “This research provides a potential way to do so when those sources are silent.”
This article originally appeared in the New York Times.
