“We want to believe, and we often get stuck and intertwine our belief with our identity…”

Feedback in:

Hi Michael,

Your podcasts, intertwined with my reading, keep bringing up ideas. One of the recent ideas is the topic of Type 1 vs Type 2 errors.

I was reading in (I think) “Thinking in Bets,” that we all have a ‘bias to believe.’ We want to believe, and we often get stuck and intertwine our belief with our identity.

The author describes our evolutionary heritage as a possible reason for our desire to believe. For example, if we hear some rustling in the bushes, it is in our interest to move or run, as it may be a Lion. If we don’t move or run, and it is a lion, we die, a type 1 error. Yet, if we don’t run, even if the odds are 1000:1, we are dead the last time, a type 2 error.

So it is good to be ‘safe,’ in this world.

In a recent podcast, discussing the Turtles, and investing: the concept of making type 1 errors was also ‘ok,’ but it is not good to make a type 2 error.

While the consequences are not a dire, a type 1 error, the downside can be addressed by: 1. limiting an investment amount/Kelly criterion, or having 2. stop losses. What we don’t want to miss is the Type 2 error, of missing the (potentially) unlimited upside–which we can’t predict, and don’t know when they will happen.

The insight that you described, along these lines, is that the APPEARANCE of safety, in the type1 scenario, leads to mediocrity or being average. WHILE the actual value, the actual living life, the actual big gains to be had, are NOT getting involved, thus making a type2 error. The error of Opportunity Cost, etc.

I really didn’t think that I would use these statistical definitions outside of a classroom, yet here it is… and this idea of type1 and type2 errors, and how to act once they are understood, now have real life.

I expect these ideas will now be part of my broader conversations, and actions.

Thanks for the podcasts,

(and I am one of your premium members of Trend following [Flagship]… still studying and learning)

Best,

Michael [name]

Thanks!

The Author