Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
The Forecasting Fallacy (alexmurrell.co.uk)
12 points by anarbadalov on Sept 19, 2020 | hide | past | favorite | 9 comments


This seems to fundamentally miss the reason why we forecast. To quote Eisenhower:

>Peace-time plans are of no particular value, but peace-time planning is indispensable.

The end result of these forecasts isn't particularly valuable, but the thinking used to develop the forecast can help us preempt, react, and respond. This is like criticizing climate scientists because they didn't get the exact amount of warming or sea level rise correct in their predictions. I'm not saying every (or any) consultant forecast is worth reading, but to dismiss the idea of forecasting because we are wrong is a fundamental misunderstanding of the value of forecasts.


> I'm not saying every (or any) consultant forecast is worth reading, but to dismiss the idea of forecasting because we are wrong is a fundamental misunderstanding of the value of forecasts.

I agree with your original statement but in the portion above you talk about reading someone else's forecast as if that was equivalent to the forecast 'thinking process' you had mentioned.

Reading the end-result of another individual's analysis won't necessarily give you the benefits you just mentioned.


The only way in my opinion to really separate snake oil sellers from serious forecasters is to require that the companies or pundits register their predictions in some database, so one can then assess how often they are wrong.


Even more than this: they should personally invest.


Hat tip to the fantastic Browser newsletter (https://thebrowser.com/) for this one. As its editor & head curator writes:

"How can it be that forecasts of global trends and outcomes are almost always wrong, yet nobody seems to mind, no stigma attaches, and the appetite for them seems only to grow?"

From the article: "To understand the extent of our forecasting fascination, I analysed the websites of three management consultancies looking for predictions with time frames ranging from 2025 to 2050. Whilst one prediction may be published multiple times, the size of the numbers still shocked me. Deloitte’s site makes 6904 predictions. McKinsey & Company make 4296. And Boston Consulting Group, 3679. ...

I believe the vast majority of these to be not forecasts but fantasies. Snake oil dressed up as science. Fiction masquerading as fact."


The article is misguided in thinking anyone cares about forecast accuracy in the first place. Accurate forecasting would usurp political authority, since it would provide a mechanism by which you can accurately hold policy creators accountable.

“It’s hard to get someone to understand something when their job depends on not understanding it.”

Robin Hanson wrote very well about this years ago:

https://www.cato-unbound.org/2011/07/13/robin-hanson/who-car...


We can predict GDP. Everyone knows ballpark where it will be without calamity.

The GDP is an overal measure of something not based on 'events'.

What we cannot predict are 'events' - like COVID, or a rash of market information coming to light like 2008 causing a domino effect.

Predicting marketing spend is like predicting what colour someone will chose for their favorite colour: it's within some range, but it's a choice.


I want to point out that Tetlock went into checking if there were ways to get better forecast and better forecaster.

It was the Good Judgement Project, and the results are slightly more positive


To what degree does the predictions alter the future?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: