Almost anyone is thrilled to read articles like ‘top 5 successful traits of business executives’, books like Jim Collins’ Good to Great, successful traits of certain individuals and how we can emulate them and not to mention the bizarre following and hero worshipping of certain individuals and the likes.

But the truth of the matter is you and I should treat them like fantasy novels.

Why, you retorted.

That’s because it’s a mirage of reality.

I wish I could explain it in a humourous and simple way but it’s difficult (I’m not skilled enough) so I’ll just lay down the points below:

Consider that a group of 100 people are given two random tests, one each on Day 1 and Day 2. At the end of each day, all these 100 guinea pigs are force ranked according to their scores.

- At the end of Day 1, how would you predict the top 10 scorers and the bottom 10 scorers for Day 2? The top 10 scorers will not do any better than the bottom 10 will fare badly.

The best explanation for the above is:

- Success = talent + luck.
- Therefore, all things being equal, the top 10 scorers are just luckier than average on Day 1. On Day 2, they are
*more likely*to be less luckier and thus their scores will regress to the mean (meaning get closer to the average). And vice versa for the bottom 10. - Before you start screaming blue murder, note that this is on average for the 100 people. It is still possible to have one or two top scorers who continue to do well on Day 2 (and vice versa).

Reasons dimpledbrain, reasons…

- Comparing these two extreme groups (top ten and bottom ten) will lead to
**asymmetrical sampling**. It then follows that any conclusion drawn from the distorted observation will be misleading. Think about successful traits of the rich. It’s a distorted sample. That is why paradoxically, we’ll observe that the very same attributes for success are also the perfect recipes for disasters. **Regression to the mean**simply means that following an extreme random event, the next random event is likely to be less extreme.**Correlation coefficient**you may recall as the relationship between two sets of data. 0 means no correlation; 1 means perfectly correlated.**Regression to the mean & correlation coefficient are actually two different perspectives on the same thing**, i.e. the lower the correlation coefficient between two things, the higher the likelihood to regress to the mean.- Regression to the mean doesn’t imply
**causal relationship**.

On the last point, it means that just because the top 10 did well, they are not doomed to fare poorly on Day 2. There is no causal explanation (i.e. why)…save for a mathematical inevitability. Think about it. If we look at Day 2 scores, we can also guess how scores on Day 1 will be distributed, i.e. those who did well on Day 2 probably didn’t do so well on Day 1.

Borrowing the words of Daniel Kahneman Nobel Laureate in Economics, he said,** “the fact that you observe regression when you predict an early event from a later event should help convince you that regression doesn’t have a causal explanation.”**

I find the above hard to swallow too, as much as when I learn that** intuition doesn’t work** except for under stable environment with immediate feedback loop. (That means treat it with caution). My heart was crushed. It’s a party pooper or a wet blanket I know, but reality is often cold and harsh. Like in the movies where we are conditioned to believe that good guys always triumph at the end, I am not so sure that it’s even a tad bit reflective of the universe we’re in.

**Tags:** asymmetrical sample · causal relationship · correlation coefficient · human behaviour · mirage · regression to the mean1 Comment

This reminds me of the book: Fooled by Randomness