1
00:00:00,909 --> 00:00:02,250
- [Instructor] Adrianna gathered data

2
00:00:02,250 --> 00:00:04,249
on different schools' winning percentages

3
00:00:04,249 --> 00:00:07,968
and the average yearly
salary of their head coaches

4
00:00:07,968 --> 00:00:12,225
in millions of dollars in
the years 2000 to 2011.

5
00:00:12,225 --> 00:00:15,619
She then created the following
scatter plot and trend line.

6
00:00:15,619 --> 00:00:17,604
So this is salary in millions of dollars

7
00:00:17,604 --> 00:00:19,936
and the winning percentage.

8
00:00:19,936 --> 00:00:24,056
And so here, we have a coach
who made over $4 million,

9
00:00:24,056 --> 00:00:28,413
and looks like they won
over 80% of their games.

10
00:00:28,413 --> 00:00:29,831
But you have this coach over here

11
00:00:29,831 --> 00:00:31,313
who has a salary of a little over

12
00:00:31,313 --> 00:00:32,823
a million and a half dollars,

13
00:00:32,823 --> 00:00:35,451
and they are winning over 85%,

14
00:00:35,451 --> 00:00:38,119
and so each of one of these data points

15
00:00:38,119 --> 00:00:39,576
is a coach,

16
00:00:39,576 --> 00:00:41,701
and is plotting their salary

17
00:00:41,701 --> 00:00:44,770
or their winning percentage
against their salary.

18
00:00:44,770 --> 00:00:48,326
Assuming the line correctly
shows the trend in the data,

19
00:00:48,326 --> 00:00:49,625
and it's a bit of an assumption,

20
00:00:49,625 --> 00:00:51,070
there are some outliers here

21
00:00:51,070 --> 00:00:53,508
that are well away from the model,

22
00:00:53,508 --> 00:00:54,679
and this isn't a,

23
00:00:54,679 --> 00:00:55,911
it looks like there's a linear,

24
00:00:55,911 --> 00:00:58,050
a positive linear correlation here,

25
00:00:58,050 --> 00:00:59,582
but it's not super tight

26
00:00:59,582 --> 00:01:01,121
and there's a bunch of
coaches right over here,

27
00:01:01,121 --> 00:01:03,781
in the lower salary area,

28
00:01:03,781 --> 00:01:05,875
going all the way from
20 something percent

29
00:01:05,875 --> 00:01:07,483
to over 60 percent.

30
00:01:07,483 --> 00:01:10,209
Assuming the line correctly
shows the trend in the data,

31
00:01:10,209 --> 00:01:15,169
what does it mean that the
line's y intercept is 39?

32
00:01:15,169 --> 00:01:16,786
Well if you believe the model,

33
00:01:16,786 --> 00:01:19,784
then the y intercept of being 39

34
00:01:19,784 --> 00:01:22,357
would be the model is saying

35
00:01:22,357 --> 00:01:24,437
that if someone makes no money,

36
00:01:24,437 --> 00:01:26,583
that they could, zero dollars,

37
00:01:26,583 --> 00:01:28,054
that they could win,

38
00:01:28,054 --> 00:01:31,666
that the model would expect
them to win 39% of their games,

39
00:01:31,666 --> 00:01:33,230
which seems a little unrealistic,

40
00:01:33,230 --> 00:01:36,072
because you would expect most
coaches to get paid something.

41
00:01:36,072 --> 00:01:37,760
But anyway, let's see
which of this choices

42
00:01:37,760 --> 00:01:40,323
actually describe that.

43
00:01:40,323 --> 00:01:42,984
So let me look at the choices.

44
00:01:42,984 --> 00:01:46,124
The average salary was
39 million dollars, nope.

45
00:01:46,124 --> 00:01:48,108
No one on our chart made 39 million.

46
00:01:48,108 --> 00:01:50,350
On average, each million
dollar increase in salary

47
00:01:50,350 --> 00:01:52,765
was associated with a 39%
increase in winning percentage.

48
00:01:52,765 --> 00:01:56,686
That would be something
related to the slope

49
00:01:56,686 --> 00:01:58,735
and the slope was definitely not 39.

50
00:01:58,735 --> 00:02:01,353
The average winning percentage was 39%,

51
00:02:01,353 --> 00:02:02,692
we know that wasn't the case either.

52
00:02:02,692 --> 00:02:05,248
The model indicates
that teams with coaches

53
00:02:05,248 --> 00:02:06,781
who had a salary of zero millions dollars

54
00:02:06,781 --> 00:02:09,971
will average a winning
percentage of approximately 39%.

55
00:02:09,971 --> 00:02:12,717
Yeah this is the closest statement

56
00:02:12,717 --> 00:02:14,264
to what we just said,

57
00:02:14,264 --> 00:02:15,740
that if you believe that model,

58
00:02:15,740 --> 00:02:17,285
and that's a big if,

59
00:02:17,285 --> 00:02:19,452
if you believe this model,

60
00:02:20,916 --> 00:02:23,375
then this model says
someone making zero dollars

61
00:02:23,375 --> 00:02:24,966
will get 39%,

62
00:02:24,966 --> 00:02:27,446
and this is frankly why you
have to be skeptical of models.

63
00:02:27,446 --> 00:02:28,688
They're not going to be perfect,

64
00:02:28,688 --> 00:02:31,249
especially in extreme cases oftentimes,

65
00:02:31,249 --> 00:02:32,247
but who knows.

66
00:02:32,247 --> 00:02:34,873
Anyway, hopefully you found that useful.