Can you play your cards right?

Back in the 1980s there was a popular TV program in the UK called “Play your cards right”. The game proved so popular that it made a comeback in 2002 and is due back again in 2019. You can find old episodes on YouTube (of course). One part of the game involved a line of playing cards that are placed, face down. If the contestants could make the right call, of simply “higher” or “lower” than the preceding card, regardless of the card’s suit, then they won a prize. It was an entertaining test of the contestants’ ability to forecast turning points - the sequence of cards was, after all, almost entirely made of them.

Forecasting ability can be a mix of extrapolation, luck and confirmation bias (looking to prove that you are right rather than looking for evidence to the contrary too). Once something begins to fall (rise), people tend to think it is going to continue to fall (rise) and it often does, particularly in slow-moving asset classes like real estate. Picking a turning point from a rise to a fall or vice versa is not so easy as it requires forecasting a change in direction. If you are going to measure forecasting ability then you first need to look at turning points. Turning points are valuable if you can predict them and take action so that you can capture them. Turning point errors can prove costly, if you act on them and you are wrong. There is more to assessing forecasting ability, but assessing how good you are at forecasting turning points is a good start.

To do this, Henri Theil suggested a simple “prediction-realisation” diagram. This shows quadrants which are comprised of forecast change (positive or negative) on one axis and actual change (positive or negative) on the other. Then, simply plot the dots where each prediction and accompanying realisation lies on the chart. If you forecast a positive change and there was one, good news, similarly for a negative change turning out to be a negative change. But, if you forecast a positive change and it was a negative change (or vice versa) you have a turning point error. (This is the basis for the quip: “Economists have forecast eight of the last two recessions”.) It is then just a case of plotting the dots and seeing how good you are; how many turning point errors did you make?

Then you need a dose of honesty.

Plot some dots on another prediction-realisation diagram using a naive competitor such as “same change as last time”. So, if last year (or whatever period you are claiming forecasting ability over) there was a positive (negative) change your “forecast” for next year is the same positive (negative) change. Clearly this is a forecast somebody with no knowledge or skill can make. And, it will be right in every period there is not a turning point. Now, using your own forecasting method, did you make more turning point errors than the naive competitor or not?

The one thing that the contestants in the ‘Play your cards right” game had on their side was that they knew exactly where they were - they knew they had a 2, an 8 or an Ace. They knew that they had just uncovered a low, a central or high card and could guess the next card accordingly, with the odds on their side. Knowing where you are is the first exercise in understanding where you might be going.