The tails are going to be particularly important for those very high and

very low outcomes.

But the intent of the probability distribution is to give us that

approximation.

If we're too far off it's not mimicking the data, and it's going to put all of our

conclusions, it's going to call them into questions.

Right, so the next exercise,

what we're going be looking at is using that binomial distribution and using that

as a means of saying how much variation do we have around our expected outcomes?

We use that to characterize the variable that we don't

necessarily have perfect knowledge of.

And based on the extent of uncertainty that we're having, we can see how that's

going to have an impact on the decisions that we ultimately make.

All right, so we're going to take a look at this in terms of the airline industry

immediately dealing with the number of no shows for a flight and

should we sell more tickets?

If we take advantage of the fact that people don't show up,

we're going to be able to sell more tickets, generate revenue that way.

But there is a chance as we sell more tickets that the flight goes into an over

sold condition and it costs the airline.

So, what's the appropriate number of seats for us to be showing?

Well, it's going to depend on my best guess for how many no-shows we have.

So we're going to assume that we know the likelihood that people don't show up for

the flight.

We're going to use a binomial distribution to characterize that and use that as our

best guest to say, how many seats should we be selling to maximize profit?

In our next session, what we're going to look at is products breaking down,

product failures, and needing to be serviced.

Well, if we know that products are going to break down and

need to be serviced should consumers buy warranty coverage for

them and should retailers offer that coverage?

Well, how do we go about pricing that?

Depends on how likely the products are to break down,

how much it's going to cost to repair, what could consumers do on their own.

So that will be the next application that we take a look at.

But if we've got resource allocation decisions, maybe I don't know

how many people are going to show up at a restaurant on a given night.

How many people do I need to have working?

Am I going to have enough tables?

Do I need to order new equipment to meet capacity?

These are all case where there is uncertainty in our decision making.

There's some factor that we don't have perfect knowledge of so

that's where the probability distribution is ultimately going to get used.