by Nada Amin, Flora Amwayi, Neira Hajro, Patrick Kim, Daniel Levy,
Dusten Salinas, Ioannis Tsoukalidis, Sumudu Watugala

Short Answer

      Replace 5 with 1-(1-p)^n/p ~= 9.56 in the final multiplication.

Long Answer

     Simon wishes to calculate the ratio of the expected time for
     Tempus to assemble a watch to the expected time for Hora to do
     so. To this end, he calculates three intermediate ratios and
     multiplies them, but he never actually verifies that they
     multiply up to the ratio of expected times as desired!

     The three intermediate ratios concern:

     1) the number of assemblies per watch
     2) the expected number of steps lost per interruption
     3) the expected number of tries to complete an assembly

     The multiplication of these three values approximates to the
     expected time in steps to complete a watch, only if the expected
     number of steps lost per interruption approximates to the
     expected number of steps per try.

     The approximation holds for Tempus, because he is so rarely able
     to complete an assembly without being interrupted.

     For Hora, however, who can usually complete an assembly without
     being interrupted, the approximation doesn't make sense. A
     "straightforward calculation" shows how much the error in the
     paper is blatant. Multiplying the three values, we find that the
     expected time for Hora is 111 * 5 * 1/.99^10 ~= 502 steps. This
     is better than the time it would take him to complete a watch
     without any interruption (111 * 10 = 1110)!