1924 CHAPTER 59. BASIC PROBABILITY

Thus

eit·(

X−m√n

)= 1+ it·X−m√

n− (t·(X−m))2

2n

+

(1− f

(t·(

X−m√n

)))(t·(X−m))2

2n.

Thus

φ Zn(t) =

n

∏j=1

[1−E

((t·(X−m))2

2n

)

+E

((1− f

(t·(

X−m√n

)))(t·(X−m))2

2n

)]

=n

∏j=1

[1− 1

2nt∗Σt +

12n

E((

1− f(

t·(

X−m√n

)))(t·(X−m))2

)]. (59.18.44)

(Note (t·(X−m))2 = t∗ (X−m)(X−m)∗ t.) Now here is a simple inequality for complexnumbers whose moduli are no larger than one. I will give a proof of this at the end. Itfollows easily by induction.

|z1 · · ·zn−w1 · · ·wn| ≤n

∑k=1|zk−wk|. (59.18.45)

Also for each t, and all n large enough,∣∣∣∣ 12n

E((

1− f(

t·(

X−m√n

)))(t·(X−m))2

)∣∣∣∣< 1.

Applying 59.18.45 to 59.18.44,

φ Zn(t) =

(n

∏j=1

(1− 1

2nt∗Σt

))+ en

=

(1− 1

2nt∗Σt

)n

+ en

where

|en| ≤n

∑j=1

∣∣∣∣ 12n

E((

1− f(

t·(

X−m√n

)))(t·(X−m))2

)∣∣∣∣=

12

∣∣∣∣E((1− f(

t·(

X−m√n

)))(t·(X−m))2

)∣∣∣∣

1924 CHAPTER 59. BASIC PROBABILITYThusC00 es aave((- ves ee)n=I |1-5 it ttt1ota i = L+itThus$7, (t) =j=se ((1-r(t (*"))) (t: (x—m))*)}. (59.18.44)(Note (t: (X —m))” = t* (X —m) (X —m)*t.) Now here is a simple inequality for complexnumbers whose moduli are no larger than one. I will give a proof of this at the end. Itfollows easily by induction.nIZ1+++Zp — Wie Wal < Ye lee — wel: (59.18.45)k=1Also for each t, and all n large enough,se((-6CS2))emar)|sApplying 59.18.45 to 59.18.44,n 1 xj=1 n= (1- ese) +pwherelen|AM