1993 RG final
Bruguera - 1.34%
Courier - (-1.00%)
That's right, Courier had a negative AM against Bruguera (though he's not the only player I've had negative numbers for).
That basically means he the cost he paid in unforced errors was greater than what he was able to win through aggressive plays.
But Bruguera's AM was not much higher. On clay this should not be that much of a surprise.
But it does suggest a lower level of play than in more recent finals, where AM's have been in the 20s-30s, correct?
Also, Brignacca provides the average AM's for the 2005 Grand Slams in the paper you linked in the OP. For men, he has them as follows.
Would it be fair or proper to add or subtract a surface factor when comparing performances in matches played on different surfaces? For example, Rafa's AM in the 2008 RG final was 34.7%, but that might equate to a 53.0% performance at Wimbledon if we added the difference between the French and Wimbledon AM averages.