By: Greg Firestone, Analytics VP, Allstate
Through a series of blog posts I want to discuss two topics that are near dear to my cold calculating heart… results oriented thinking and why we should think probabilistically.
Let’s play a game, called “Easy Money”. Here is how it works: You get to pick either 1 or 2-10. I will roll a ten sided dice (chosen to make math nice) if your number comes up you win $1,000. The game cost $100 to play. Should you play? What would you choose?
The answer is yes and 2-10.
I now want to imagine that you run an investment firm and you have two employees working for you, Spike and Timmy. They each have $1,000 to invest and you have set the goal of making at least 10% return on the money. They both decide to play Easy Money. Timmy chooses 1 and plays the game ten times and wins every time. He has made $10,000 easily surpassing his goal. Spike picked 2-10 and lost every time and has made no money. Who should you fire?
If we base our decision on achieving the goal we should fire Spike. He made no money and Timmy made tons of money. Yet, Timmy made some really bad choices. The probability of Timmy making $10,000 was only 0.00000001%. Spike made some good choices he was expected to make $9800. So, do we decide based on achieving the goal or based on who made the better decision? We will discuss this next time…..