Experiment of the week #1: The randomness of Swedish krona coin toss


A coin toss is generally regarded as giving a 50 % probability of either outcome – heads or tails. But to what degree does it really work? The sides of the Swedish 1 krona coin differ from eachother, and I would think the head side (featuring a portrait of the king) is heavier. Would this cause this side to land on the ground more often (similar to a buttered toast more often landing with the buttered side)? And to what degree would such an effect be detectable? Would this seriously affect the practice of tossing a coin in Sweden?


I use a regular Swedish 1 krona coin, flipping it 200 times in as similar a fashion as possible, noting down the results on a notepad. I estimate a statistical margin of 10 %, so results within 40 – 60 % would be expected if the coin toss really is a good predicter of pure chance. Any results within 100 – 60 % or 0 – 40 % would warrant some further investigation with a larger number of coin tosses.


My results were 102 coin tosses showing heads and 98 showing tails, or in other words exactly 51 % heads and 49 % tails. This is well within the statistical margin (40 – 60 %) and so my results confirm the common idea that a coin toss is a good determiner of 50 / 50 % probability.