Adrian
Actually, if you are looking for good opinion and to weed out bad opinion, one person or a very small panel is the way to go as its much easier to control quality with a small panel. If you want all and sundry to see if any consensus is reached, then a larger panel is good, but only if all segments of abilities etc are well represented. Unfortunately, with the larger panel mags, they list courses which don't achieve a consensus - say rated highly ( a criteria set by the folks in charge) by at least 75% of the minimum number of raters (34 out of 45 in GD's case) or something like that. Instead, they shoot for an even number rather than a list of true notables followed by a list of honourable mentions that may someday make the list proper.
Pietro
As I state above, the goal of a large panel shouldn't be some sort of objective truth through numbers - it should be reaching a consensus. Shit, I don't ever use any set list of criteria as each course is different and will somehow not fit the criteria exactly.
In any case, I much prefer grouping/a star system where the reader can essentially draw his own conclusions within certain parameters rather than a dogfight over numbers.
Ciao