title: “Confirmation Bias Consumes It’s Next Willing Victim, Mike Nesbitt #nipolitics #prediction” date: 2016-05-07 categories: - “data” - “data-and-statistics”

- “northern-ireland”

As human beings we tend to over estimate our own predictions and beliefs. When someone has a deep connection with the thing they are estimating then the bias tends to increase again.

So spare a thought for Mike Nesbitt who put his overconfidence, admirably, on the line with the exposure to public ridicule where requested. In all honesty he wasn’t that far out, a prediction of ¹⁸⁄₁₉ with an actual of 16. Oddly enough he’d revise his estimate down 10% then he’d be in the right area (see weather men and women of the world over estimate the chance of rain by 10% and under estimate the chance sun by 10%).

It’s also very hard to predict anything where the last event was, in reality, five years out. If anything Mr. Nesbitt only reminded us that prediction is hard when there’s a lack of data.

Revealed - how close was the UUP leader’s prediction of the number of Assembly seats his party would win:https://t.co/m6i8kuWC8e

— BBC Newsline (@bbcnewsline) May 7, 2016

//platform.twitter.com/widgets.js

So Mr. Nesbitt, I know you’re not reading this and I’m neither a supporter or affiliate of the party, but at least you stuck your neck on the line and letting the BBC make a meal out of it. Next time though, revise down 10%