Native American hunter gatherers were using dice for gaming and gambling more than 6,000 years before the practice appeared anywhere else, a new study argues.

It says dice were being made and used on the western great plains of North America at the end of the last ice age, more than 12,000 years ago.

It had been thought that the earliest examples of dice were in the bronze age societies of Mesopotamia and the Indus valley.

Robert Madden, author of the study, said: “Historians have traditionally treated dice and probability as old world innovations. What the archaeological record shows is that ancient Native American groups were deliberately making objects designed to produce random outcomes and using those outcomes in structured games thousands of years earlier than previously recognised.”

This suggests that ancient Native Americans possessed “a basic working knowledge of chance, randomness and probability” and consequently “were early movers in humanity’s emerging understanding and practical application of these concepts”.

The study says games of chance and gambling allowed disparate groups with no relationship to “interact; exchange goods, information, and mates; and forge new social bonds”.

Madden, who is a PhD student in archaeology at Colorado State University, said: “These findings don’t claim that ice age hunter gatherers were doing formal probability theory. But they were intentionally creating, observing and relying on random outcomes in repeatable, rule-based ways that leveraged probabilistic regularities, such as the law of large numbers. That matters for how we understand the global history of probabilistic thinking.”

The earliest examples identified in the study come from late pleistocene archaeological sites in Wyoming, Colorado and New Mexico.

Up to 12,800 years old, the two-sided dice were crafted from wood or bone that would have been tossed in groups on to a playing surface, the study says.

Madden’s research is based on re-examining artefacts in collections that were often labelled as “gaming pieces” or were overlooked altogether. He argues that these wrongly classified or ignored objects were dice.

He tells CSU’s The Audit podcast that he spent a long time methodically going through online databases and libraries looking for examples of what he argues are dice.

“It was like a treasure hunt and I just kept looking for examples and keeping track of them,” he said. “Finally, after about three years, I compiled the dataset that makes up this paper and tracked this practice all the way from the well-documented historical era over the last 2,000 years as far back as I could.

“That’s where we see this incredible persistence of this practice going all the way back to the late pleistocene – about 12,000 years ago – which is more than 6,000 years older than the earliest dice known in the old world.”

The paper uses the word gambling, although it is not gambling as we might know it today.

“What we think of gambling is like at a casino where you bet against the house,” Madden said. “And as we all know, any time you play against the house, the odds are always slightly tilted in favor of the house.

“These games are one on one; there’s no house. This is me against you. It’s a fair game, everybody’s got an equal opportunity, equal conditions, and it was used as a form of exchange … particularly between groups of people who did not come into frequent contact with each other.”

The paper, Probability in the Pleistocene, is published in the journal American Antiquity.