In his recent interview on the InvestED Podcast, Jake Taylor, co-host on our Value After Hours Podcast provided a great hack on how to know if you have skill or you’re just lucky in investing. Here’s an excerpt from that interview:
Taylor: One of the most difficult things in this whole investment game is untangling luck versus skill as we talked about in the first session it’s so random any one year right so here’s a little secret to kind of getting a better handle on that luck versus skill.
Now it actually comes from a 1950s weatherman, a statistician named Glenn Brier and what we’re doing here is we’re going to track our predictions and we’re going to assign probabilities to those predictions. So let’s say that… and how I break it down is on sort of the five drivers of where returns come from.
So there’s revenue, margins, gross margins, there’s earnings, so change in earnings, change in the market multiple assigned to those earnings, and then dividends. So I have these five things that are sort of the key drivers of of investment returns.
I’m writing these down so I’m going to interrupt you and ask you to repeat them. Revenue, margins, earnings and what were the other two?
Multiple, so what the market assigns a multiple to those earnings, and then dividends. Technically there’s another thing like with a change in the balance sheet. So if a company pays off their debt but we can ignore that for now.
So if you know these five things I can make a prediction about each of those five things in a probabilistic way. I can say I think there is a 75% chance that Apple will have 10% higher revenue in one year from now, and I tend to make
these on an annual basis. So step one is making a probabilistic prediction. Step two is actually scoring that prediction. Was I right or not? So if I have a very high probability estimate and it turns out I’m right that gives
you the highest score in this. It’s called a Brier Score and based on Glenn Brier who invented this.
If I give a very high probability and I’m wrong that gives me the worst score because that’s like you thought you knew what you’re doing and you don’t. Now in between those where you have low probabilities and it turns out right or wrong that’s not really worth a lot because you didn’t really have much confidence in what you were saying. So what the trick is and this is why this is such a good hack is that for every one data point of returns which is what I get in a given year right. Like what did Apple’s stock do in this case I get five data points on these five key drivers that tell me do I know what I’m doing or not.
So if I’m wrong on all five of my predictions and yet the stock went up I know that that was probably dumb luck and that I should probably not think that I know what I’m doing quite as much. This is if you don’t do this and you only looked at what the return was you’re setting yourself up for the big loss later because you think you know what
you’re doing and you get overconfident.
Host: You’re measuring yourself only based on the market as opposed to the more shall we say reality based scorecard that is this.
Exactly and we can get more data points in a faster time period to see do we have luck versus skill.
Host: Oh yeah true.
Conversely let’s say that you got five out of five of your predictions were right. You had high confidence and the stock went down. Well you should then realize oh man that was kind of just bad luck I know what I’m doing like I’m getting the right drivers and I’m making the right predictions I just didn’t get the result that I wanted.
But if I keep going I know that eventually luck tends to iron out and I think I should end up with a good result. I have some skill here and it’s not just purely luck. So I think that’s especially… early on as you’re learning and tracking how you’re developing I think it’s hugely important to do these type of exercises.
You can watch the entire interview here:
For all the latest news and podcasts, join our free newsletter here.
Don’t forget to check out our FREE Large Cap 1000 – Stock Screener, here at The Acquirer’s Multiple: