Why So-Called Experts Make Lousy Decisions

Posted by: Rachael King on December 7, 2011

Data analytics, which is popular in business, is catching on in cities around the world. As I recently wrote, Memphis has reduced crimes such as robberies and rapes to their lowest levels in a quarter-century, thanks in part to IBM’s analytical software.

Why has this approach been so successful? Or put another way: Why did decision-makers, prior to data analytics, make such poor decisions? Part of that answer has to do with a kink in human reasoning, according Michael Lewis, author of “Moneyball,” which told the story of how the Oakland A’s turned to data analysis to find undervalued players passed over by baseball experts.

Lewis writes in the December issue of Vanity Fair that he did not know the answer to this question until he discovered the work of Nobel laureate Daniel Kahneman and his collaborator Amos Tversky. Both were psychologists and published work on judgment and decision-making.

“They had found that people, including experts, unwittingly use all sorts of irrelevant criteria in decision-making,” writes Lewis. This problem has to do with the way the brain accesses or remembers information. When making a decision, people tend to base it on examples that are easily remembered. But “reliable statistical evidence will outperform” that thinking, the article says.

Since this is a human problem, it affects not only baseball experts but also hedge fund managers, CEOs and even police officers.

In Memphis, police officers were able to get a better handle on crime patterns once they started using IBM’s data analytics software to figure out crime patterns and forecast where it was going to occur.

For example, Memphis police officers had long known that burglaries tended to spike in the month of March. But until the use of data analytics, nobody realized that the spike corresponded with Spring Break in Memphis-area schools. Once the police added more patrols during that week, the burglary rate dropped.

Reader Comments

AER

December 7, 2011 8:27 PM

It affects hedge fund managers, CEOs and...structured finance rating agency analysts! But finance generates profits for itself from the bad decisions of so-called experts, as we have just seen, so don't expect a rush to financial saber metrics.

rob beckman lapré

December 8, 2011 3:44 AM

I do agree with Rachael's statement; I am aware of referred article,and the software used to track(potential) criminal spots(was this software, as I recall without digging in my files, also used in other police depts.(Florida etc.etc).Read this on a blog dated 02/11/010 by "Bubbles"(short version)"This analytical software program( "streaming system,transmitting live video,audio, and data over bTCP/IP")is VERY robust and allows analysts to continue these networks to collaborate from all over the globe,can export to Analytical Notebook or Axis Pro Format.If you are looking for a true analytical capability I highly recommend it.It starts at $ 150 K if you are not part of somebody else architecture.(sorry,name of product and company to be withheld for obviuous reasons).

Peter Tarlan

December 8, 2011 5:20 AM

Interesting article Rachael. I saw a good Whitepaper on Twitter today actually on www.lovehatedata.com which highlighted how business intelligence has revolutionised the way data is analysed and used for decision-making nowadays.

But your right experts, do use all sorts of irrelevant criteria in decision-making, which obviously plays a big part in bad results.

Peter

James Young

December 8, 2011 3:49 PM

GET REAL! Are you a schell for IBM. What ever happen to asking the beat cop!It seems everyone has forgotten common sense.

RealExpert

December 8, 2011 9:39 PM

Some "experts" are not really experts at all. When the stock market goes up, a bull market "expert" get camera time and headlines. When the market goes down, a bear market "expert" get camera time and healines. Real experts understand, and can discuss, both directions.

rob beckman lapré

December 9, 2011 3:44 AM

Sorry folks, but my 12/08/011`reaction did NOT refer to the software used in "CRUSH"program in Memphis,Tn.This very interesting analytical program,which "pulls off one of the great computer science feats of the era;it combs through all available databases,identifying realted pieces of information,and puts everything together in one place"(comment of computer analist).Manufacturer's customers; US DOD,Army,Marines,Air Force, PD NY City,PD Los Angeles.With this type of analytical software "the bad guys"are targeted, and "what-if-scenario's"are nearing 100%(?) accuracy(my opinion).

Carlos

December 9, 2011 12:06 PM

That would be the whole point. Even the best human brain is not very good at spotting large-scale trends, mainly because we're extremely biased towards outlier events. Cops' opinions shouldn't be discarded, but they are definitely not immune to this. Pretty much every single piece of technology you own would not have been possible without statistical tools to understand the subtleties of nature, which is in many ways simpler than social or economic issues. It's about time we start using objective statistics in business and social contexts too.

Sabari

December 11, 2011 9:24 AM

This is one excellent usage of data analysis. But again i feel how we can determine the nature of humans. Moreover after establishing such patterns publicly, anti social elements become more alert and change their pattern, so the decision making tools need to be updated regularly.

Post a comment

 

About

Bloomberg Businessweek writers Peter Burrows, Cliff Edwards, Olga Kharif, Aaron Ricadela, and Douglas MacMillan, dig behind the headlines to analyze what’s really happening throughout the world of technology. Tech Beat covers everything from tech bellwethers like Apple, Google, and Intel and emerging new leaders such as Facebook to new technologies, trends, and controversies.

Categories

 

BW Mall - Sponsored Links

Buy a link now!