Well, since I enter as a homebrewer and judge both the home and commercial competitions at the MCI, let me take a cut at answering your questions....
First, it is difficult to compare the home and commercial scores directly, in part because they are on two different scales (BJCP-based home comps are scored on a 50 point system; the commercial comps are scored on a 100 point scale), and partly because the criteria for awarding points are different (BJCP puts a lot of emphasis on being "true to category," and lacking in perceived flaws, where the commercial guys are much more interested in scoring that is based on a perceived positive organoleptics and marketability of the product). That said, there are some "rules of thumb" about the numeric score that are similar between the two systems.
Let me focus on the home scores, since the majority of Gotmeaders are home meadmakers. The BJCP has worked long and hard to quantify an assessment system for meads (and for beers, obviously) that bestows a significant amount of objectivity on the judging process. Nevertheless, numeric scales are our best effort at putting a quantitative score on what is essentially a subjectively perceptive process, so IMHO any mead's numeric total score should be regarded as having a variance of about +-3 points on it.
Now regarding minimum scores, at the MCI we try to give credit, rather than subtracting for fault, in all the judging. The bottom line is that while there are some judges out there who really take a perverse delight in finding arcane reasons to subtract as many points from that perfect score of 50 as they can, we try to discourage that practice at the MCI. Other contests are different, so you may see some differences between average scores at MCI and at some other events. In general, a mead has to really be virtually undrinkable in order to receive a net score below 25 at the MCI. That is entirely consistent with the scoring methodology of the BJCP, and we have found (especially in recent competitions) that home meadmakers are for the most part really turning out quite acceptable meads, although some have more perceived flaws than others. So within that range, with a virtual scoring floor of 25 and a theoretical perfect score of 50, we see that many meads end up scoring in the 35-38 range. That doesn't mean they are "average." Instead, it means that most entries are very good, with only minor flaws. If a score for one of your meads falls in the low 30's (i.e. 30-34) it is still a good mead, but could use a little more work to come up to the current perception of very good. From 38-42, the flaws are detectable, but very minor. Many meads in this range will advance from first flights to subsequent judging rounds. If you score 43 or higher, now you have something truly special with only the slightest hints of flaws. Then there are the truly incredible meads out there -- scoring in the 48-50 range -- and that almost never happens. If you scored there, the judge considers your mead a world class example of its category.
Judges do vary, and in general judges with more experience ranking beers tend to be more sticklers for a literal interpretation of the "guidelines," but mead judges tend to allow a bit of latitude in their interpretation.