A common competition run by vendors of fishing equipment is a ‘guess the weight and win’ where an image of someone holding a fish is posted and it is up to you to guess it’s weight with the closest guess winning a prize.

The ‘law of large numbers’ implies that the average of the guesses of many is superior to the average of the guesses of a few, so the ‘best guess’ should be close to the average of all guesses…

Motivated by the possibility of winning some fishing tackle I set about messing about with R’s regular expressions to create a tool that would enable me to make an informed guess based on the guesses of many.

The function below reads in a text file containing each persons guess (provided via a comment), extracts and cleans the guesses, transforms the guesses into a common unit (kilograms) and provides summary statistics and a histogram that would suggest the best guess you could make. Of course this function could be adapted to suit a ‘how many jelly beans in the jar?’ competition also!

Here is the output of one such competition:

Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
4.00   12.50   17.00   17.35   19.90   85.00 

In this case, I would guess the weight of the fish to be around 17 kilograms!