Probability
Basics:
- Probability is a measure of the
likelihood of some event happening. (Just about
anything can be an event in this sense.) We measure on a scale from
0 to 1 (or 0% to 100%), where smaller numbers indicate less
likelihood and higher numbers indicate greater likelihood. This
system in an example of how mathematics is used to formalize and
make precise our informal notions about some things being more or
less likely to occur than other things.
Two kinds of
probability:
We can distinguish two kinds of
probability. Mathematical probability is the measure of the
relative frequency of an event occurring. In
addition, we will use the term personal
probability for a statement of someone's degree of belief
that an event will occur.
Determining Relative Frequency
(Mathematical Probability)
Empirical Method
:
If the process under study can be
repeated or simulated many times, we can determine the
empirical probability by keeping track of the
outcomes in our (large number of) trials. The probability assigened
is
P(A happens) = (# times A happened)
/ (# trials)
If the number of trials is very
large, then it is quite likely that this will give us a reliable
estimate.
Theoretical
Method:
Sometimes we can make mathematical
assumptions about a situation and use Four Basic Properties
of Probability to determine the theoretical
probability of an event. The accuracy of a theoretical
probability depends on the validity of the mathematical assumptions
made.
The four useful rules of probability
are:
- It happens or else it
doesn't. The probability of an event happening added the
probability of it not happening is always 1.
P(A happens) + P(A doesn't happen) = 1
- Exclusivity. If A
and B can't both happen at the same time (in which case we say that
A and B are mutually exclusive), then
P(either A or B happens) = P(A happens) + P(B happens)
- Independence. If B
is no more or less likely to happen when A happens than when A
doesn't (in which case we say that A and B are independent), then
P(A and B both happen) = P(A happens) * P(B happens)
- Sub-Events. If
whenever A happens B must also happen, then B must be at least as
likely as A, so
P(A happens) <= P(B happens)
Empirical probabilities will also
follow these rules (for a given set of trials). Becuase people
often have a poor sense of the likelihood of an event, personal
probabilities often do not follow these rules. A collection of
personal probabilities is called coherent if it
does not violate the rules for mathematical probability.
Equally likely
outcomes
- One especially important use of
these probability rules is the conclusions that can be drawn if we
assume that a number of events are equally likely. If there are
only n such events that are possible in a given situation,
and all are equally likely and pairwise mutually exclusive (no two
can happen at once), then each must have probability 1/n.
More complicated situations can be handled by dividing a situation
into a number of equally likely outcomes and counting how many of
them are "of interest" (in the event). The probabilty then is given
by (number of interest)/(total number), just as in the case for
empirical probability.
The Law of Probability :
Law of Large Numbers :
- you can discover the unknown
probability of an event through experimentation. Using the previous
example, say you do not know the probability of drawing a certain
colored marble, but you know there are three marbles in the bag.
You perform a trial and draw a green marble. You perform another
trial and draw another green marble. At this point you might claim
the bag contains only green marbles, but based on two trials, your
prediction is not reliable.
- It is possible the bag contains
only green marbles or it could be the other two are red and you
selected the only green marble sequentially. If you perform the
same trial 100 times you will probably discover you select a green
marble around 66% percent of the time. This frequency mirrors the
correct probability more accurately than your first experiment.
This is the law of large numbers: the greater the number of trials,
the more accurately the frequency of an event's outcome will mirror
its actual probability.
Law of Subtraction:
- probability can only range from
values 0 to 1. A probability of 0 means there are no possible
outcomes for that event. In our previous example, the probability
of drawing a red marble is zero. A probability of 1 means the event
will occur in each and every trial. The probability of drawing
either a green marble or a blue marble is 1. There are no other
possible outcomes. In the bag containing one blue marble and two
green ones, the probability of drawing a green marble is 2/3.
- This is an acceptable number
because 2/3 is greater than 0, but less than 1--within the range of
acceptable probability values. Knowing this, you can apply the law
of subtraction, which states if you know the probability of an
event, you can accurately state the probability of that event not
occurring. Knowing the probability of drawing a green marble is
2/3, you can subtract that value from 1 and correctly determine the
probability of not drawing a green marble: 1/3.
Law of Multiplication :
- If you want to find the probability
of two events occurring in sequential trials, use the law of
multiplication. For example, instead of the previous three-marbled
bag, say there is a five-marbled bag. There is one blue marble, two
green marbles, and two yellow marbles. If you want to find the
probability of drawing a blue marble and a green marble, in either
order (and without returning the first marble to the bag), find the
probability of drawing a blue marble and the probability of drawing
a green marble. The probability of drawing a blue marble from the
bag of five marbles is 1/5. The probability of drawing a green
marble from the remaining set is 2/4, or 1/2. Correctly applying
the law of multiplication involves multiplying the two
probabilities, 1/5 and 1/2, for a probability of 1/10. This
expresses the likelihood of the two events occurring together.
Law of Addition:
- Applying what you know about the
law of multiplication, you can determine the probability of only
one of two events occurring. The law of addition states the
probability of one out of two events occurring is equal to the sum
of the probabilities of each event occurring individually, minus
the probability of both events occurring. In the five-marbled bag,
say you want to know the probability of drawing either a blue
marble or a green marble. Add the probability of drawing a blue
marble (1/5) to the probability of drawing a green marble (2/5).
The sum is 3/5.
- In the previous example expressing
the law of multiplication, we found the probability of drawing both
a blue and green marble is 1/10. Subtract this from the sum of 3/5
(or 6/10 for easier subtraction) for a final probability of
1/2.