Now as you can see, we have
three prediction rules,
and the first question is
what are the three predictions
that they give for this week.
So, prediction 1, attendance last week,
well we can see that that
attendance last week was 40,
and the prediction is 40,
so all of these actually have the
correct answer for prediction 1.
The second one is 100 minus
attendance last week - 60.
100 - 40
And then 20 plus attendance last week,
20 + 40 is 60
so that shows us that answer A
is the correct one.
Now notice that none of these strategies
used this data of the attendance
two weeks ago.
They don't have to, even though
it is available to them,
they don't have to use it.
Now let's go to question 2.
This asks what did the three
prediction rules say last week.
So last week, the attendance we see was 40
but before we even knew that,
we wanted to know what the
prediction rules predict.
So prediction 1 says attendance
last week,well that would have been 70
because it was the week
before we're predicting.
So prediction 1 is 70
prediction 2 is 100 minus attendance
last week. 100 - 70 is 30.
And prediction 3 is 20 plus attendance
last week. 20 + 70 is 90.
So the correct answer is C.
So now we want to know which strategy
is the current best.
And that was defined as the one that was
closest to predicting correctly last week.
Last week the real answer was 40,
and the closest prediction to 40,
well, 30 is closer to 40 than 90 or 70,
so prediction 2 gives us
the best prediction,
and that's the one we're going
to use this week.
Prediction 2 for this week was 60.
Well that's just at the threshold
and we said the rule was
if you are at the threshold,
or below, then you go.
So our answer is to go.