When was the Bronze Age? Players put it 1,500 years too late

Of 4,301 guesses, more than half landed after 500 BC, in the world of the Greeks and Romans or later. For the typical object, 85 in 100 guesses landed closer to today than the real date. I didn't expect old dates to lean so hard one way.

Researchers have seen this pull on a far longer scale. In a 2009 study, Kefyn Catley and Laura Novick asked students to date events in Earth's history, and the events "were generally estimated to have occurred more recently than they actually did". They called it forward telescoping.

How the guessing worked

Each round shows the name of a thing with a small picture. You drop a pin on a timeline from 3000 BC to today, or type a year. No starting pin, nothing to compare with; some rounds offer a one-line hint like "Egypt was already building pyramids". Where the country is known, for about half the guesses, a third came from the US and one in eight from the UK, the rest from 71 other countries. I don't ask for age or gender.

The Bronze Age lands in Roman soil

The first palace at Knossos went up around 1900 BC. Players typically put it in 309 BC.

The Nebra sky disc, bronze with gold stars, came out at 120 BC. It was made around 1700 BC.

Not one typical guess landed earlier than the real date.

My hypothesis: what we don't know gets filed with the Romans

My hypothesis was that for most of us, "ancient" means Greece and Rome, and anything older without a clear picture gets filed there. Two things would follow. Objects from the classical world should be dated about right. And famous Bronze Age objects should slide far less than unfamiliar ones from the same centuries.

Both checks held. The nine classical objects were typically missed by about 300 years, the Bronze Age ones by about 1,500.

The Sphinx and Mohenjo-daro, a major Indus Valley city, were both built around 2500 BC. Players put the Sphinx at 1653 BC and Mohenjo-daro at AD 754.

The further back, the stronger the pull

So my hypothesis only partly held: it is a slope, not a single magnet. Almost a third of the Bronze Age guesses went past AD 500, into the Middle Ages or later. Across all 90 objects with at least 100 guesses, the older something is, the more it slides towards today. Things from after 1900 even go the other way, dated a few years too early.

Is it just the timeline?

The fair objection is the game itself. Its timeline gives the distant past little room, and guesses drift away from both ends, so the oldest objects get nudged inwards. That explains part of it. It cannot explain why the Sphinx and Mohenjo-daro, from the same century and the same spot on the line, land 2,400 years apart.

Two findings fit as analogies: people without schooling space numbers logarithmically (Dehaene and colleagues, 2008), and memory may keep the past on a compressed timeline (Howard, 2018).

What this can and can't tell us

These are players of a daily browser game, not a representative sample. The pull is not one country's, though: in the US, the UK and Germany, between 58 and 66 per cent of Bronze Age guesses landed after 500 BC. When 15 of 15 objects move the same way, the direction is hard to argue with, even if the game's timeline adds to it.

Next time something is called ancient

Ask which ancient. The Great Pyramid was older to Cleopatra than Cleopatra is to us, going by WorldAtlas's dates. For other one-way misses, read why the Falcon 9 nearly triples or why the Hindenburg shrinks to a fifth.