From Daniel Kahneman “Thinking, fast and slow”
People forecast and predict all kind of things. Financial analysts forecast earnings, Economists forecast inflation, contractors predict the time it will take to finish their project, people forecast how much growth of
People often forecast using their intuition which leads to a forecast with Bias. “intuitive predictions tend to be overconfident and overly extreme.” (p. 192)
“Intuitive predictions need to be corrected because they are not regressive and therefore are biased.” (p. 190)
How to Get a More Accurate Forecast With Less Bias.
- 1. Get the Base Rate (also knows as the Outside View). In order to get the Base Rate, you must ignore detailed information and get the average rate of it’s class.
- 2. Now gather more information and evidence that’s unique for your case. This is called the Inside View.
- 3. Now analyze the information carefully. How would you rank the quality or validity of this information? Is the information strong or weak?
- 4. Once you decide how strong and accurate the new information is, write down an estimate of the Correlation Coefficient.
- 5. Write down your intuitive prediction.
- 6. Your forecast should be a number between the Base Rate and your intuitive prediction. Use the Correlation Coefficient to adjust between these two numbers. If the quality of your information is weak, you should stick closer to the Base Rate.
Example of a Forecast:
(Please note the calculations are done on my own, not from Kahneman’s book)
Julie is currently a senior in a state university.
She read fluently when she was four years old.
What is her grade point average (GPA)? (p. 186)
You probably guessed 3.7 or 3.8. Your intuitive prediction is overconfident and extreme. Let’s look how Kahneman would have forecast it.
Get the Outside View or Base Rate. Ignore any extra information about Julie. If you had no information about Julie, you must assume she’s an average student and her GPA would be a 3.0.
Now look at the quality and validity of Julie’s information. How strong is the correlation of reading fluently at age 4 with GPA in a University?
If you think Julie is in the 90 percentile for reading (at age 4), it doesn’t automatically mean she will be in the 90 percentile in GPA. There are many other factors that contribute to her GPA. We don’t have other data on Julie, so the quality of the information is weak. However you can give Julie some points for being able to read at an early age.
Give an estimate of Correlation Coefficient. Kahneman made a guess that reading at age 4 has a .30 Correlation Coefficient with GPA. It has some correlation but not very strong.
Now give your intuitive guess of Julie’s GPA. Most people said 3.7.
So the final formula is
Base rate= 3.0
Intuitive forecast= 3.7
Correlation Coefficient= .30
3.7-3.0= .70
.30 x .70 = .21
So you add .21 to the base rate (3) = 3.21
Your final forecast should be 3.21!