For instance, if games are played to ascertain the most effective number and kinds of craft for which to ask appropriations from Congress at next session, the solution, unless a money limit were fixed, would be impossible.

In other words, the amount of money to be expended must be one of the known or assumed factors in the problem.
As this amount can never be known, it must be assumed; and, in order that the whole value of the games may not be lost, in case the amount assumed were incorrect, it is necessary to assume a number of possible sums, the upper limit being above the probable amount to be received, and the lower limit below it, and then work out the answer to the problem, under each assumption.
Of course, this procedure would be laborious, but most procedures are that bring about the best results.

Suppose that such a procedure were followed for, say, a year, and that a number of plans, all worked out, were presented to Congress when it met: plan No.

1, for instance, consisting of such and such craft showing (according to the results of the games) the best programme, if $100,000,000 were to be appropriated for the increase of the navy; plan No.

2, if $90,000,000 were to be appropriated; plan No.