I have spoken with a few of people in regards to this issue, you'd be on the short list. I'd probably remember your voice/name but to be honest it all kind of blends together. I certainly do remember our conversation though!
Based on what I know of our course congestion over the first 9.5 miles (at 9.5 you turn onto Pine St where it's a lot wider compared to the pack of runners) I wouldn't be so quick to say with certainty that a runner couldn't do 7:00 pace even starting at the back. Without question over the first 1k you could not. That's up Pearl St. Much of this street has barricades to keep spectators back. A runner could not move to the sidewalk to get around slower runners. However, when we turn right onto Willard St a runner would be able to move onto the sidewalk. Spectators are sparse in that area, not non-existant but nothing like Battery St, Church St, or Pearl St. On Church St, miles 2.7-2.9 and 9.2-9.4 it would be really hard to pass too many runners, just too narrow.
I also know that when runners get to mile 3.2, coming back through the start zone, you would be able to pass a lot of people. We use cones to split the street into 3 lanes (2 for racing it's an out-and-back section plus 1 for a relay exchange zone). I work this area every year to keep things safe, and although most runners stay inside the cones I'm always yelling at a few runners who are outside the cones. It's not an issue of cutting the course by running outside the cones so we don't pay that much attention to the runners who are sprinting by the slower runners in that area. What we see a lot is this is right after the first relay exchange is made. 2nd leg runners for relay teams with a slow first runner are often outside the cones, some running what I'd estimate to be sub 6's.
I do have a few other examples in my file of decent runners who either missed our start or started near the back (stuck in porto lines...late driving in...started with a friend, etc) and went on to run solid times. Nobody ran sub 3 like Kip did, but a few have been sub 3:10 to make a BQ.
The geometry of VCM would support a runner tripping the start mat then stepping off the course, then rejoining the race near the 10 mile mark. In this regard VCM isn't dissimilar to many road races. We all know this is an outdoor sport and few races are held on courses that are 100% locked down. Bottom line, if someone really wants to cheat there isn't a race out there that you can't figure out a way to do it and possibly get away with it.
For 2010 we had 2 unannounced checkpoint systems in place, but I'm sad to say neither worked 100%. 2010 was our largest finisher numbers ever, up over 10% from any previous year. Kip does show up on one of our systems (audio tape reading of bib numbers at the 12.5 mile mark, furthest point south), but does not show up in backup photography at mile 9.2. For what it's worth, our photographer did miss a number of other runners in the time period I calculated Kip would have come through that area. I wish we had 100% coverage or video tape at 9.2.
Someone else asked if it's possible Kip ran from 13.1 to 20 and may have cut the course elsewhere. That is certainly a possibility. As I've stated a number of times, 13.1 to 20 is the section I believe would be hardest to get between if one were taking a car or bike during the race. 13.1 to 20 was Kip's slowest section, 6:51/mile (6:11/mile to 6:27/mile the other sections). 13.1 to 20 does include the biggest uphill on our course but that's not enough, to me, to account for the discrepancy.
Thanks for the good words on 2011, I appreciate that. I issued 4 DQ's this year. In all cases their first half/second half splits didn't make any sense, plus they were all missing the 20 mile split time. One was also missing on our audio tape bib check at mile 21.5, but the other 3 would have come through 21.5 after we turned off the tape (we only taped up through 8:15/mile). None of these DQ's contacted me afterwards to dispute our action. The bigger issue is the runners whose B-tags malfunction and we have to research to get them in the official results. Had about 30 of those this year, which is cut in half from 2010. It's getting better, but still a learning curve.