This is fascinating.

That said, if it's possible to do better than random guessing, then does this reflect the fact that the five charts are presumably hand-selected to be "interesting"?

My naive guess, and I'd be very curious to learn if this were wrong, is that something very close to the efficient market hypothesis is true; that, if it were possible to beat the monkeys on randomly chosen stocks on random dates, then someone would have figured this out already and deployed bots to capture whatever profits are available.

> That said, if it's possible to do better than random guessing

Yes - by always picking Up.

Random stock on a random day has 53-55% chance of closing higher. Over 5 days, you will be right ~60%

1996–2016: 53.3% of days were positive. 2016–2021: 54.9% of days saw gains.

The baseline is the market average, not 0.

The interesting answer by the author, will_asouka, has been marked 'dead' for some inscrutable reason.

Thanks for flagging, summary is: Charts are from a seeded random draw. And the instinct is pretty spot on players state ~74% average confidence but hit ~54%. Accuracy is basically flat across confidence brackets. Up only strategy quietly beats the coin flipping monkeys (stocks generally go up and to the right) but players call down 43% of the time.

They're from a seeded random draw. Random S&P 500 stock, random 60-day window that's at least 12 months old, every player gets identical charts. Gemini writes the narrative vignette.

Your instinct is pretty spot on. 6,000+ calls in: players state ~74% average confidence but hit ~54%, and accuracy is basically flat across confidence brackets. Up only strategy quietly beats the coin flipping monkeys but players call down 43% of the time.