The cone of uncertainty

Yes, its ‘cone of uncertainty’ time again for hurricane watchers as Dorian ends up over 350 miles from where it was projected to be 5 days ago. Given that the 5 day margin of error on these forecasts appears to only around plus-or-minus 125 miles, this must temper President Trump’s relief that the current forecasts project Dorian veering north to miss Mar-a-Lago. Perhaps these weather forecasts are predicated on a low confidence coefficient, to exclude the ‘outliers’. As in equity markets however, do this and one might encounter a few too many of Goldman Sachs CFO, David Viniar’s 25 Sigma events .

Like any equity forecaster, the National Hurricane Center (NHC) has a tough task in trying to predict future events in an apparently chaotic world, because one only has a grasp of a few of the variables. While responsible weather forecasters recognise this margin of error, with their ‘cones of uncertainty’, even if occasionally massively under-estimated, this is rarely mentioned when you’re being sold an equity story. Here the closest you might get is a blanket ‘risk’ adjustment to an equity premium that is based on a probably unrelated and irrelevant historic share price beta. But in marketing, admitting to uncertainty when predicting the future is rarely a winning sales strategy to trigger a commission generating event, even if it proves to be the better performing investment approach. Appreciating that forecasts on all events can carry an ‘natural’ margin error can save your life when it comes to a hurricane, or some expensive losses on your equity portfolio. That is why quantifying volatility on prospective outcomes ought to be a central requirement on any equity valuation modelling, as it is on the GrowthRater. If you know, you can take a measured risk, if not, you’re just gambling.

Back to Dorian, meanwhile. 5 days ago it was supposed to have been half way across Florida by now and just north-east of Tampa, having already rolled ashore near Trump’s Mar-a-Lago.

5 days ago, Dorian was expected to have made swift progress into Florida by now

Instead, Dorian has decided to loiter around northern Bahamas, which is approximately 350 miles away from where it was forecast to be just 5 days ago. Against a 5 day margin of error on the ‘cone of uncertainty’ of around +/- 125 miles, that doesn’t appear to have been materially altered, this ought to alert us of two points. Firstly and most obviously, that hurricane forecasting is an unreliable art and of limited reliability. The second, is that notwithstanding an outcome that has fallen substantially outside its own predicted margin of error only 5 days ago, this hasn’t permeated through to any adjustment in the error range going forwards. But then, given the exponential impact in expanding the size of the cone, that might seem to devalue to results and who knows, possibly also the funding. Perhaps weather and equity forecasting isn’t so different after all!

It doesn’t seem to matter how wrong the forecast is, the projected margin of error doesn’t change!