Brian A. answered • 10/01/15

Tutor

5
(7)
Former Math professor and CFA Charterholder for Math and Finance Tutor

I bet Maurizio;s answer is the one your teacher wants but is is surely not a very good answer. In fact, it's one of those really disappointing questions because it misses an opportunity to talk about something beautiful. The number of fatal car crashes in a day is a Poisson r.v.. It has to be if you believe only some very small assumptions about how fatal car crashes happen. That means you estimate the Poisson mean with the data above (0*144 + 1*172+...)/489 and then calculate Poisson probabilities using the formula you can get from Wiki. That would give about 0.673 compared to Maurizio's answer of 0.645 which isn't much different here.

The Poisson estimate is much more efficient (i.e., uses the data better). Among other things Maurizio's estimate treats a day with 4 crashes in the same way it treats a day with 5 crashes so important information is lost.

It's awful to work hard at convincing people that statistics is boring...

Brian A.

Yep, and if you caught him at a weak moment he would say that he would cut off his hand to be able to write Ph.D. after his name.

Report

10/01/15

Maurizio T.

10/01/15