where:	l(h,a)	=	number of games lost by each human opponent h versus each agent a
	p(h,a)	=	total number of games played, a versus h
	l(h)	=	total number of games lost by h
	p(h)	=	total number of games played by h