WEBVTT

00:00:00.140 --> 00:00:01.030
Hi.

00:00:01.030 --> 00:00:03.740
Today, we're going to do a
really fun problem called

00:00:03.740 --> 00:00:05.520
geniuses and chocolates.

00:00:05.520 --> 00:00:09.380
And what this problem is
exercising is your knowledge

00:00:09.380 --> 00:00:12.230
of properties of probability
laws.

00:00:12.230 --> 00:00:15.340
So let me just clarify
what I mean by that.

00:00:15.340 --> 00:00:18.160
Hopefully, by this point, you
have already learned what the

00:00:18.160 --> 00:00:20.980
axioms of probability are.

00:00:20.980 --> 00:00:25.290
And properties of probability
laws are essentially any rules

00:00:25.290 --> 00:00:27.880
that you can derive
from those axioms.

00:00:27.880 --> 00:00:31.760
So take for example the fact
that the probability of A

00:00:31.760 --> 00:00:34.830
union B is equal to the
probability of A plus the

00:00:34.830 --> 00:00:37.640
probability of B minus the
probability of the

00:00:37.640 --> 00:00:38.700
intersection.

00:00:38.700 --> 00:00:42.450
That's an example of a property
of a probability law.

00:00:42.450 --> 00:00:44.860
So enough with the preamble.

00:00:44.860 --> 00:00:47.930
Let's see what the problem
is asking us.

00:00:47.930 --> 00:00:51.040
In this problem, we have
a class of students.

00:00:51.040 --> 00:00:55.610
And we're told that 60% of the
students are geniuses.

00:00:55.610 --> 00:00:58.000
70% of the students
love chocolate.

00:00:58.000 --> 00:01:00.370
So I would be in
that category.

00:01:00.370 --> 00:01:03.470
And 40% fall into
both categories.

00:01:03.470 --> 00:01:06.550
And our job is to determine
the probability that a

00:01:06.550 --> 00:01:10.060
randomly selected student is
neither a genius nor a

00:01:10.060 --> 00:01:11.400
chocolate lover.

00:01:11.400 --> 00:01:14.330
So first I just want to write
down the information that

00:01:14.330 --> 00:01:16.800
we're given in the problem
statement.

00:01:16.800 --> 00:01:21.580
So if you let G denote the event
that a randomly selected

00:01:21.580 --> 00:01:25.990
student is a genius then the
problem statement tells us

00:01:25.990 --> 00:01:30.360
that the probability of
G is equal to 0.6.

00:01:30.360 --> 00:01:33.890
Similarly, if we let C denote
the event that a randomly

00:01:33.890 --> 00:01:37.390
selected student is a chocolate
lover, then we have

00:01:37.390 --> 00:01:42.240
that the probability of
C is equal to 0.7.

00:01:42.240 --> 00:01:45.340
Lastly, we are told that the
probability a randomly

00:01:45.340 --> 00:01:49.610
selected student falls into
both categories is 0.4.

00:01:49.610 --> 00:01:52.590
And the way we can express
that using the notation

00:01:52.590 --> 00:01:58.320
already on the board is
probability of G intersect C

00:01:58.320 --> 00:02:00.610
is equal to 0.4.

00:02:00.610 --> 00:02:05.770
OK, now one way of approaching
this problem is to essentially

00:02:05.770 --> 00:02:10.120
use this information and sort of
massage it using properties

00:02:10.120 --> 00:02:13.310
of probability laws to
get to our answer.

00:02:13.310 --> 00:02:16.710
Instead, I'm going to take a
different approach, which I

00:02:16.710 --> 00:02:18.120
think will be helpful.

00:02:18.120 --> 00:02:19.860
So namely, we're going
to use something

00:02:19.860 --> 00:02:22.250
called a Venn diagram.

00:02:22.250 --> 00:02:28.320
Now a Venn diagram is just a
tool that's really useful for

00:02:28.320 --> 00:02:31.750
telling you how different sets
relate to each other and how

00:02:31.750 --> 00:02:33.310
their corresponding
probabilities

00:02:33.310 --> 00:02:34.770
relate to each other.

00:02:34.770 --> 00:02:38.660
So the way you usually draw this
is you draw a rectangle,

00:02:38.660 --> 00:02:41.420
which denotes your sample space,
which of course, we

00:02:41.420 --> 00:02:42.620
call omega.

00:02:42.620 --> 00:02:45.860
And then you draw two
intersecting circles.

00:02:45.860 --> 00:02:51.070
So one to represent our geniuses
and one to represent

00:02:51.070 --> 00:02:52.430
our chocolate lovers.

00:02:52.430 --> 00:02:55.970
And the reason why I drew them
intersecting is because we

00:02:55.970 --> 00:03:00.730
know that there are 40% of the
students in our class are both

00:03:00.730 --> 00:03:02.540
geniuses and chocolate lovers.

00:03:02.540 --> 00:03:07.610
OK, and the way you sort of
interpret this diagram is the

00:03:07.610 --> 00:03:12.440
space outside these two circles
correspond to students

00:03:12.440 --> 00:03:16.000
who are neither geniuses
nor chocolate lovers.

00:03:16.000 --> 00:03:20.050
And so just keep in mind that
the probability corresponding

00:03:20.050 --> 00:03:22.330
to these students on the
outside, that's actually what

00:03:22.330 --> 00:03:24.200
we're looking for.

00:03:24.200 --> 00:03:28.710
Similarly, students in this
little shape, this tear drop

00:03:28.710 --> 00:03:31.150
in the middle, those would
correspond to geniuses and

00:03:31.150 --> 00:03:32.400
chocolate lovers.

00:03:32.400 --> 00:03:34.820
You probably get the idea.

00:03:34.820 --> 00:03:36.620
So this is our Venn diagram.

00:03:36.620 --> 00:03:40.560
Now I'm going to give you guys
a second trick if you will.

00:03:40.560 --> 00:03:43.030
And that is to work
with partitions.

00:03:43.030 --> 00:03:47.240
So I believe you've seen
partitions in lecture by now.

00:03:47.240 --> 00:03:50.730
And a partition is essentially a
way of cutting up the sample

00:03:50.730 --> 00:03:52.480
space into pieces.

00:03:52.480 --> 00:03:55.500
But you need two properties
to be true.

00:03:55.500 --> 00:03:58.810
So the pieces that you cut up
your sample space into, they

00:03:58.810 --> 00:04:02.260
need to be disjoint, so
they can't overlap.

00:04:02.260 --> 00:04:06.610
So for instance, G and C are
not disjoint because they

00:04:06.610 --> 00:04:10.000
overlap in this tear
drop region.

00:04:10.000 --> 00:04:12.750
Now the second thing that a
partition has to satisfy is

00:04:12.750 --> 00:04:16.300
that if you put all the pieces
together, they have to

00:04:16.300 --> 00:04:18.670
comprise the entire
sample space.

00:04:18.670 --> 00:04:22.940
So I'm just going to put these
labels down on my graph.

00:04:22.940 --> 00:04:28.700
X, Y, Z, and W. So X is
everything outside the two

00:04:28.700 --> 00:04:31.300
circles but inside
the rectangle.

00:04:31.300 --> 00:04:35.180
And just note, again, that what
we're actually trying to

00:04:35.180 --> 00:04:38.380
solve in this problem is the
probability of X, the

00:04:38.380 --> 00:04:40.890
probability that you're neither
genius, because you're

00:04:40.890 --> 00:04:43.740
not in this circle, and you're
not a chocolate lover, because

00:04:43.740 --> 00:04:46.040
you're not in this circle.

00:04:46.040 --> 00:04:50.330
So Y I'm using to refer
to this sort of

00:04:50.330 --> 00:04:52.270
crescent moon shape.

00:04:52.270 --> 00:04:55.680
Z, I'm using to refer
to this tear drop.

00:04:55.680 --> 00:04:59.640
And W, I'm using to refer
to this shape.

00:04:59.640 --> 00:05:03.590
So, hopefully, you agree that
X, Y, Z, and W form a

00:05:03.590 --> 00:05:05.530
partition because they
don't overlap.

00:05:05.530 --> 00:05:06.890
So they are disjoint.

00:05:06.890 --> 00:05:09.160
And together they form omega.

00:05:09.160 --> 00:05:12.690
So now we're ready to
do some computation.

00:05:12.690 --> 00:05:15.730
The first step is to sort of
get the information we have

00:05:15.730 --> 00:05:19.660
written down here in terms
of these new labels.

00:05:19.660 --> 00:05:27.110
So hopefully, you guys buy that
G is just the union of Y

00:05:27.110 --> 00:05:32.740
and Z. And because Y and Z are
disjoint, we get that the

00:05:32.740 --> 00:05:35.620
probability of the union is the
sum of the probabilities.

00:05:35.620 --> 00:05:39.620
And, of course, we have from
before that this is 0.6.

00:05:39.620 --> 00:05:43.220
Similarly, we have that the
probability of C is equal to

00:05:43.220 --> 00:05:47.920
the probability of Z union W.
And, again, using the fact

00:05:47.920 --> 00:05:50.660
that these two guys are
disjoint, you get this

00:05:50.660 --> 00:05:51.850
expression.

00:05:51.850 --> 00:05:54.950
And that is equal to 0.7.

00:05:54.950 --> 00:05:58.790
OK, and the last piece of
information, G intersects C

00:05:58.790 --> 00:06:02.810
corresponds to Z, or our tear
drop, and so we have that the

00:06:02.810 --> 00:06:06.450
probability of Z is
equal to 0.4.

00:06:06.450 --> 00:06:09.830
And now, if you notice,
probability of Z shows up in

00:06:09.830 --> 00:06:11.520
these two equations.

00:06:11.520 --> 00:06:13.760
So we can just plug it in.

00:06:13.760 --> 00:06:16.820
So plug in 0.4 into
this equation.

00:06:16.820 --> 00:06:20.830
We get P of Y plus 0.4 is 0.6.

00:06:20.830 --> 00:06:23.050
So that implies that
P of Y is 0.2.

00:06:23.050 --> 00:06:25.110
That's just algebra.

00:06:25.110 --> 00:06:27.090
And similarly we have point.

00:06:27.090 --> 00:06:30.800
0.4 plus P of W is
equal to 0.7.

00:06:30.800 --> 00:06:34.770
So that implies that
P of W is 0.3.

00:06:34.770 --> 00:06:36.630
Again, that's just algebra.

00:06:36.630 --> 00:06:39.880
So now we're doing really well
because we have a lot of

00:06:39.880 --> 00:06:40.840
information.

00:06:40.840 --> 00:06:43.930
We know the probability of Y,
the probability of Z, the

00:06:43.930 --> 00:06:48.300
probability of W. But remember
we're going for, we're trying

00:06:48.300 --> 00:06:52.260
to find the probability of X. So
the way we finally put all

00:06:52.260 --> 00:06:56.250
this information together to
solve for X is we use the

00:06:56.250 --> 00:07:00.890
axiom that tells us that 1 is
equal to the probability of

00:07:00.890 --> 00:07:02.860
the sample space.

00:07:02.860 --> 00:07:05.330
And then, again, we're going
to use sort of this really

00:07:05.330 --> 00:07:10.090
helpful fact that X, Y, Z, and W
form a partition of omega to

00:07:10.090 --> 00:07:14.310
go ahead and write this as
probability of X plus

00:07:14.310 --> 00:07:17.800
probability of Y plus
probability,

00:07:17.800 --> 00:07:19.540
oops, I made a mistake.

00:07:19.540 --> 00:07:21.680
Hopefully, you guys
caught that.

00:07:21.680 --> 00:07:23.522
It's really, oh, no.

00:07:23.522 --> 00:07:24.235
I'm right.

00:07:24.235 --> 00:07:25.790
Never mind.

00:07:25.790 --> 00:07:28.030
Probability of X plus
probability of Y plus

00:07:28.030 --> 00:07:32.040
probability of Z plus
probability of W. And now we

00:07:32.040 --> 00:07:34.860
can go ahead and plug-in the
values that we solved for

00:07:34.860 --> 00:07:36.140
previously.

00:07:36.140 --> 00:07:43.480
So we get probability of X plus
0.2 plus 0.4 plus 0.3.

00:07:43.480 --> 00:07:45.600
These guys sum to 0.9.

00:07:45.600 --> 00:07:49.010
So, again, just simple
arithmetic, we get that the

00:07:49.010 --> 00:07:52.720
probability of X is
equal to 0.1.

00:07:52.720 --> 00:07:57.100
So we're done because we've
successfully found that the

00:07:57.100 --> 00:08:00.060
probability that a randomly
selected student is neither a

00:08:00.060 --> 00:08:03.520
genius nor a chocolate
lover is 0.1.

00:08:03.520 --> 00:08:05.600
So this was a fairly
straightforward problem.

00:08:05.600 --> 00:08:08.180
But there are some important
takeaways.

00:08:08.180 --> 00:08:11.090
The first one is that
Venn diagrams are

00:08:11.090 --> 00:08:12.470
a really nice tool.

00:08:12.470 --> 00:08:16.450
Whenever the problem is asking
you how different sets relate

00:08:16.450 --> 00:08:18.890
to each other or how different
probabilities relate to each

00:08:18.890 --> 00:08:21.870
other, you should probably draw
Venn diagram because it

00:08:21.870 --> 00:08:23.530
will help you.

00:08:23.530 --> 00:08:28.870
And the second takeaway is that
it's frequently useful to

00:08:28.870 --> 00:08:33.760
divide your sample space into
a partition mainly because

00:08:33.760 --> 00:08:35.320
sort of the pieces
that compose a

00:08:35.320 --> 00:08:37.929
partition are disjoint.

00:08:37.929 --> 00:08:40.599
So we will be back soon to
solve more problems.