General:
========
In Section 12 we write that two independent separable statistics are used, and that they
are identically distributed as a one-variate normal random variable N(u, u) for 
u = n*mu*C^(-1)*mu. 



mu*C^(-1)*mu = u/n = 9.3498065782031568501973e-13



n = 2^39 * 7:
=============
u = 3.5980773675668169780516 (computed in the "old" way. Ignore this 
unless you are verifying an old draft of the paper)

u = 3.5980773675668169704622 (computed like in paper)
z = 1.0335996763286862643819
n = 3848290697216
log2(n) = 41.807354922057604108138
alpha = 0.99269207541645714573241
beta = 0.088194138395904420113568
(1-beta) = 0.9118058616040955798661
(1-beta)^2 = 0.83138992925558710201175
(1-alpha) = 0.0073079245835428542675885
log2(1-alpha) = -7.0963225389711979463649
2*log2(1-alpha) = -14.19264507794239589273
# of K^hat returned = 2^39.807354922057604108138
# of keys to test = 2^41.807354922057604108138



n = 2^39 * 7.2257:
==================
u = 3.7140896621182213402888
z = 0.98061363072909915519076
n = 3972370584411
log2(n) = 41.853137357524727288349
alpha = 0.99257519589049966079368
beta = 0.078041603343014413075699
(1-beta) = 0.92195839665698558693108
(1-beta)^2 = 0.850007285166319572281
(1-alpha) = 0.0074248041095003392063172
log2(1-alpha) = -7.0734313212376363545243
2*log2(1-alpha) = -14.146862642475272709049
# of K^hat returned = 2^39.853137357524727291819
# of keys to test = 2^41.853137357524727291819