>    restart; read `\c:/Users/Asus/Google Drive/Aek/Card2/NoFeedBack.txt`;

>    # Formula of p_[1,1,..,1](i,j)

>    k:=2:i:=5:j:=3:binomial(i-1,j-1)*(2^k-1)^(i-j)/2^(k*i)*2^k;

27/128

>    B:=For1k(2,5,3); P1(5,3,10);

B := 27/128

27/128

>    #The approximation for k,i at j^*

>    # This is the error from CLT (ii) mentioned in the paper

>    Error1k(3,500);

.488959e-6

>    i:=16000: evalf(Error1k(4,i)*i^(3/2));

.1894136519

>    i:=160000:evalf(Error1k(4,i)*i^(3/2));

.1894144000

>    #It looks like the error of each term is of order 1/i^(3/2)

>