When you take risks based on data

Today is Super Bowl XLIX. One scoring play we are most likely to see today is the point after kick. Teams seldom miss it. It is a sure one point. They do have the option of going for two instead of taking the sure one point. Teams go for the two only when they are forced to. That is when they re in fourth quarter and a single point still puts then at a disadvantage or keeps the door open for the other team.

Successful two point conversion is not given. It is risky. But the time to take risk is when you have time and resources to recover if it does not pan out as expected. Not when when that is the only option. Armed with data it is easier to take risks than with just gut.

Why don’t they go for it every time? Should they go for it? Let us look at some data.

Today’s Super Bowl features defending champion Seattle Seahawks and three time champions New England Patriots. Here is how the team statistics look like (source)



In the past seven years Seahawks attempted five fewer than Patriots and have less than half the success rate of Patriots.
So it is likely Seahawks won’t go for two unless they are forced into it and are most likely to miss it.

Patriots have a 50% average success rate over the past seven years. That is an expected value of 1, same as assured point after kick. So they should go for two every time. There is only upside for them.