I’ve spoken and written many times in the past about the folly of prediction. It’s something that we are seemingly hard-wired to look for as a species (if someone can tell us the future, we mitigate away a whole series of risks); the people who predict the most extreme versions of the future are the most newsworthy and so both get the most coverage and yet are also generally the most wrong. These modern day seers, however, have little or no accountability for their work (see, in particular, the credit rating agencies on that), so continue to produce their gobbledegook. For more on this check out either the half hour Freakonomics podcast, or Dan Gardner’s book-length “Future Babble“.
The thing is, though, making predictions about the future is really tempting. In the future we shall all wear silver clothes. See! There I go again.
In a conversation with Tony Cocks on Twitter this week I realised what is needed is a consistent relative probability scale so that predictions are put into a context that people who don’t get statistics (ie most of us) can understand. The Watford Probability Index is therefore born.
First of all, nothing is completely certain and nothing is completely uncertain. However, some things are incredibly likely, and others remarkably unlikely. Those two ends will be marked by “the world will still be spinning tomorrow morning” (incredibly likely) and “the world will stop spinning tomorrow morning” (remarkably unlikely).
Now we need some markers for things that are really quite unlikely, but might or might not happen. For that we must refer to that well known tax on people who don’t understand statistics, the lotto. Our benchmark for really quite unlikely is me winning the lotto in my lifetime. The fact that I don’t actually play the lotto doesn’t make much difference to the overall probability of that outcome, and if you don’t agree then can I suggest that you give me your hard-earned pound each week?
Now we come to a marker of something that is pretty darn unlikely, but could possibly happen. For that we draw on this scale’s title: Watford to won the Premier League in the next ten years. Much as I would love for this to become a reality, in reality it probably won’t.
Finally we need a 50/50 option, so we resort to the chances of me flicking a coin and it landing on “heads”. This scale is a bit logarithmic.
So, let’s apply to some topical subjects…
The chances of global warming being man made? Much, much more likely than that of me flicking a heads.
The chances of the global warming deniers being right? Only slightly better than of Watford winning there league.
The chances of The Singularity happening in my lifetime? Somewhere between Watford and a Lottery win. The chances of the American government never opening up again? Between a lottery win and the world stopping spinning.
And as for hover boards, the riches so cruelly predicted in Back to the Future? Within my lifetime? Hmm. Between a coin toss and a Watford win.
This might need some refining… but as a serious point, using reference points that people get, rather than abstract percentages or decimals, would go a long way to help more people understand probability and risk.