The Model has been struggling for confidence in recent weeks: beaten by Ronnie d’Arch Smith, only matching Andrew Brown, and being overtaken by Paul Merson.
Therefore, the Model wanted a new forecasting challenge. Taking it on this week, we have two 3-year-old boys and one 5-year-old girl.
Instead of taking on adults, we will evaluate whether the Model can beat these children either individually or all together (i.e. taking their best forecasts for each match).
The Model’s forecast are here.
The forecasts of the children are below – but first a bit on our “scientific method”, in case anybody ever tries to replicate this experiment.
Method and Experiment design
- The children were randomly sampled (from among our own children).
- Each child was given 25 LEGO or DUPLO bricks of assorted colours and sizes.
- The Premier League Fixtures were written on scraps of paper, which were placed on a (kitchen) table.
- The children were told they were helping with daddy’s “work”.
- They were told to place as many bricks next to each team name according to how many goals they thought that team would score in the match.
- They were only prompted with the following statements “You don’t have to place a brick on every team; you can place more than one brick on a team.”
- No forecast revisions were allowed once all the bricks had been placed.
- The children are called D, E & R.
- D made his forecasts independently. R & E made forecasts sequentially, in the same location, but blind of each other, with R making her picks first.
- D placed his bricks from the side of the table closest to the Home teams (Home bias).
- E & R placed their bricks from the side of the table closest to the Away teams (Away bias).
- Evidence on the method and data collection process is presented in the featured image above and is available on request.
The forecasts of the children are below – they appear to favour teams scoring at least one goal, and their is clear evidence that the side of the table picks are made from matters.
|Child D||Child E||Child R|