00:01
All right, so we're going to use the regression analysis tool in excel to estimate a regression equation that predicts winning percentage.
00:10
And we have the data down here.
00:12
This is our data.
00:15
Make it bigger for us.
00:19
So the way we do this in excel is we have our y range, which is a winning percentage.
00:26
That's our y.
00:27
And then everything from runs to teamera.
00:32
You select the whole table here essentially is the x.
00:38
Include the labels and now we'll pop all of this summary right here.
00:43
And if we're looking at it, we're told to note the coefficients and the significance at the 10 % level.
00:53
So 10 % significance levels.
00:57
Can we get rid of any of these variables? because it says, are there any independent variables that appear to be unnecessary? so if we're looking at them, the only ones that appear necessary are runs, because it's got the p value 0 .05 and team era because it has a p value of 0 .05 but these other ones they have a higher p value so you can it means that this significance level we don't consider them to be as significant here so that's why runs and team era are significant because they're within that 10 % threshold and the others are not now we're going to rework the model based on what we just found.
01:42
We're going to only look at runs and team era.
01:48
And that's what this is right here.
01:54
Same thing.
01:56
Winning percentages are y, but then runs and team era are the x variables.
02:01
Here's our output.
02:04
And we can see the p value for both the runs and team era is still really low here, which is still within that same 10 % level of significance.
02:19
So we go.
02:22
And let's compare these two.
02:24
And what we're going to do is look at the adjusted r squared values and the standard error.
02:34
So here in the original model with all the variables, the adjusted r squared is 0 .915, the standard error of 0 .0214, 0 .2267.
02:46
Whereas here, the adjusted r squared is just a little higher, but the standard error is smaller.
02:55
So for that, we could say we might want to go with this one because there's less error, slightly less error...