4.5 General Probability Rules Rules of Probability Rule 1. Rule 2.

0 P ( A) 1

for any event A

P ( S ) =1

Rule 3. Complement rule: For any event A,

Rule 4. Addition rule: If A and events, t"en

are dis!oint

or

B ) = P( A) + P ( B)

Rule #. $ultipli%ation rule: If A and independent events, t"en

are

and

B) = P ( A) P ( B)

&nion '"e union of any %olle%tion of events is t"e event t"at at least one of t"e %olle%tion o%%urs.

Figure 4.1# '"e addition rule for dis!oint events: P(A or or C)*P(A)+P( )+P(C) ,"en events A, , and C are dis!oint.

Addition Rule for -is!oint .vents If events A, , and C are dis!oint in t"e sense t"at no t,o "ave any out%omes in %ommon, t"en
P(

one or more of

A, B, C ) = P ( A) + P ( B) + P (C )

'"is rule e/tends to any number of dis!oint events.

Figure 4.10 '"e 1eneral addition rule: P(A or )*P(A)+P( )2P(A and ) for any events A and .

3eneral Addition Rule for &nions of ',o events For any t,o events A and ,
P( A

or

B) = P( A) + P( B) P( A and B )

Figure 4.14 5enn dia1ram and probabilities for ./ample 4.36. ./ample 4.36 -ebora" and $att"e, are an/iously a,aitin1 ,ord on ,"et"er t"ey "ave been made partners of t"eir la, firm. -ebora" 1uesses t"at "er probability of ma7in1 partner is 6.0 and t"at $att"e,8s is 6.#. P(at least one is promoted)*6.0+6.#26.3*6.9 P(neit"er is promoted) * 1 2 P(at least one is promoted) * 1 : 6.9 * 6.1.

Conditional Probability and Probability Trees '"e ne, notation P( A | B) is a %onditional probability. '"at is, it 1ives t"e probability of one event under t"e %ondition t"at ,e 7no, anot"er event. ;ou %an read t"e bar < as =1iven t"e information t"at.> -efinition of Conditional Probability ?"en P( A) > 0 , t"e %onditional probability of 1iven A is
P ( B | A) = P ( A and B ) P ( A)

./ample @et8s define t,o events:

t"e ,oman %"osen is youn1, a1es 14 to 29

* t"e ,oman %"osen is married

'"e probability of %"oosin1 a youn1 ,oman is

'"e probability t"at ,e %"oose a ,oman ,"o is bot" youn1 and married is
P ( A and B ) = 7,842 = 0.075 103,870

'"e %onditional probability t"at a ,oman is married ,"en ,e 7no, s"e is under a1e 36 is
P ( B | A) = P ( A and B ) 7,842 = = 0.348 P ( A) 22,512

$ultipli%ation Rule '"e probability t"at bot" of t,o events A and "appen to1et"er %an be found by
P ( A and B ) = P ( A) P ( B | A)

./ample 4.34 Alim is still at t"e po7er table. At t"e moment, "e ,ants very mu%" to dra, t,o diamonds in a ro,. As "e sits at t"e table

loo7in1 at "is "and and at t"e upturned %ards on t"e table, Alim sees 11 %ards. Bf t"ese, 4 are diamonds. '"e full de%7 %ontains 13 diamonds amon1 its #2 %ards, so 9 of t"e 41 unseen %ards are diamonds. 'o find Alim8s probability of dra,in1 t,o diamonds, first %al%ulate
P( P(

first %ard diamond ) = 41

se%ond %ard diamond < first %ard diamond ) = 40

$ultipli%ation rule no, says t"at

bot" %ards diamonds ) = 41 40 = 0.044 .

Alim ,ill need lu%7 to dra, "is diamonds.

Probability Trees $any probability and de%ision ma7in1 problems %an be %on%eptualiCed as "appenin1 in sta1es, and probability trees are a 1reat ,ay to e/press su%" a pro%ess or problem.

./ample. %onsider t"e problem of flippin1 a fair %oin t,i%e. Bn t"e first flip or sta1e t"e out%ome %an eit"er be a "eads or a tails. '"is is e/pressed in t"e tree dia1ram belo, by movin1 from left to ri1"t ,it" one bran%" 1oin1 up or do,n to represent t"e t,o out%omes. Doti%e t"at t"e probabilities of t"e t,o out%omes are ,ritten ne/t to t"eir respe%tive bran%"es, bot" bein1 .# for t"is e/ample. '"e ne/t set of bran%"es represent t"e ne/t toss of t"e %oin. '"e ne/t toss %an also be eit"er a "eads or a tails. '"ese bran%"es e/tend from eit"er of t"e first out%omes bran%"es be%ause t"e ne/t toss %ould be a "eads or a tails no matter ,"at t"e out%ome of t"e first toss ,as. Doti%e also t"at t"e probabilties for t"e se%ond set of bran%"es is also .#, t"is means t"at t"e probability of "eads or tails on t"e se%ond toss does not depend on ,"at "appened on t"e first toss. '"is is an e/ample of t"e first and se%ond tosses bein1 independent. Bne event does not %"an1e t"e probability of t"e ot"er event "appenin1.

'"e probability of any final out%ome of t"e e/periment li7e EE, or "eads on bot" tosses is found by multiplyin1 t"e bran%" probabilities it too7 to rea%" t"e final tree out%ome on t"e ri1"t. Al,ays remember: multiply t"e bran%" probabilities. '"e probability of any ot"er final out%ome ,ill be found in t"e same fas"ion, li7e 'E, tails follo,ed by "eads ,ill also be 1F2 times 1F2 1ivin1 1F4 also.

Fi1ure. 'ree dia1ram for ./ample. '"e probability P( ) is t"e sum of t"e probabilities of t"e t,o bran%"es mar7ed ,it" asteris7s (G). ./ample '"ere are t,o dis!oint pat"s to (professional play). y t"e addition rule, P( ) is t"e sum of t"eir probabilities. '"e probability of rea%"in1 t"rou1" %olle1e (top "alf of t"e tree) is
P ( B and A) = P ( A) P ( B | A)

= 0.05 0.017 = 0.00085 .

'"e probability of rea%"in1 (bottom "alf of t"e tree) is

,it"out %olle1e

P ( B and Ac ) = P ( Ac ) P ( B | Ac )

= 0.95 0.001 = 0.00095 .

About 9 "i1" s%"ool at"eletes out of 16,666 ,ill play professional sports.

Doti%e. In t"e last tree t"e tree ,as dra,n ,it" t"e 1iven information in t"e first set of bran%"es. In a problem des%ription ,at%" for ,"at is t"e H1iven t"atH information. '"ese items ,ill be t"e first bran%"es of t"e tree, and t"en t"e ot"er events ,ill be t"e ne/t set. If you follo, t"is 1uideline you "ave a better %"an%e of %onstru%tin1 t"e tree %orre%tly.

ayes8s Rule If A and are any events ,"ose probabilities are not 6 or 1,

P( A | B) =

P ( B | A) P ( A) P ( B | A) P( A) + P( B | Ac ) P ( Ac )

Independent .vents ',o events A and t"at bot" "ave positive probability are independent if
P ( B | A) = P ( B )