Don’t Forget this ONE Thing about Mean Reversion

This is going to be a very small post. I tweeted this today.

“Many people (me included) are of the view that this mean reversion was long due. But ‘mean reversion’ doesn’t imply that things stop at mean only. Mathematically, it has to go lower to an extent it went up to let the mean remain ‘the mean’.”

Here’s the tweet:

I am afraid to say that the vast majority of market participants didn’t realize this. So for the readers of Stable, I thought I will elaborate a bit more.

Let’s say that some random thing that we are tracking has a mean (average) of 10. Now in the recent past, it went up to 15, i.e. 50% above the mean. So for the mean to remain the same mean (i.e. 10), as and when this thing falls, it will have to come down to 5. Why? Because only then, the mathematical mean of 15 & 5 will remain as 10 (which is the original mean).

This is ofcourse a very simple and crude way of looking at it. But it sends across the message.

And let’s just say that it doesn’t fall to 5 but only up to 8. Then the mean will shift. The new mean will be the average of 15 & 8 – which is 11.5.

So mathematically, it has to go lower to an extent it went up to let the mean remain ‘the mean’.

And this also means that ‘mean reversion’ doesn’t necessarily mean that once things begin to fall, they will just stop at the original mean only. They may go lower too. Why? Because for the mean to remain the mean, they have to go lower to an extent to which they went up.

This is one aspect of mean reversal which most people don’t give weightage to or more commonly, are unaware of. Don’t be part of that crowd. 🙂

Note – A few days back, I wrote this and this about current markets. You may want to read it.


  1. Not true… this implies that probabilities have memory, and ‘remember’ what the mean used to be.

Leave a Reply