Suspicious: no. I treat GPS like a mild curiosity, and when it's off, I think nothing of it. After all, the simple difference in running straight lines vs. drifting around etc. accounts for enough discrepancy that it isn't really worth wondering about. Have you seen DCRainMaker's GPS comparisons between watches? He will wear 6 watches at one time and show you the GPS data... just not something I consider authoritative on precise measurements.
Most runners knew the course: I knew it as well--that racers ought to run on the right side of the road--because of the other runners on the course. I had no inkling that there was any cause for concern if a runner drifted to the inside of curves. Also, I don't recall ever consciously thinking "I'm going to run on the other side of the road because it will be faster even though I shouldn't." As I've stated previously, I have a firmly ingrained instinct to run the shortest possible course available to me. Even then, for very nearly the entire time, there were runners coming at me on the left side of the road, so if I were to estimate how much time I actually spent on the left side vs. the middle, it would be quite a bit less than the theoretically possible amount of distance.
One of the most notable missing links in the entire argument about the runners cutting this short is the fact that you are all assuming the runners must have been optimizing every moment of the race. There is a long list of reasons why that isn't likely, not least of which is the fact that there is just as much photo evidence of us running right next to the other runners in the middle. But hard data aside, the bottle tables were on the far right side, and navigating to those required not only being on that side, but also weaving through the half marathon runners who were all on that side. It's probably no more than a miniscule contribution, but I contend that the opposite contribution was also minimal. Overall, the entire argument here hinges on a premise that is so particular that I just find it statistically unlikely.
Namely this: you claim that the abundant proof of concern is in the fact that all the GPS data from runners is short enough to raise concern AND that photo evidence of runners running in restricted areas further proves the point. But we've already confirmed that a great many runners followed only the righthand side of the course and yet still ended with just about the same distance discrepancy. The most logical conclusion is usually the right one, and it remains true that the most logical conclusion is that the GPS data is skewed and that whatever potentially minimal advantage a runner gained by small amounts of time cutting the curves shorter, that advantage was likely returned in fairly equal measure in other instances.
Remember the numbers--my previous note about a 0.5% advantage was a liberal estimation. In reality, it is likely more around 0.3%. That's margin of error territory.
The rule with such deliberations, as always, is that you will never convince your opposition to change their mind, but as stated, I hope those who ran it and are seeing this thread see enough clarity here to feel justified in their great performances and dismiss these narrow criticisms.