No, not at all. It depends on the calculated 2011 biases, which I haven't seen yet. And you are right, that even these biases do not fully capture the individual benefit of pacemaking. I thought all the caveats would be obvious, but I'll make them explicit now.
First, I'm just interpreting figures calculated and posted at the ARRS website. Although I have some confidence they use a sound statistical approach to comparing what cannot be easily compared, I can't speak for the accuracy and uncertainty of their computed race biases, without performing the calculations myself, or analyzing the data and the methods (which I haven't done and will not do).
Second, I haven't seen specific race biases for 2011, except for Boston. The ARRS will compute an exact race bias, based on their methods and criteria, the actual race results and some predicted results. We likely won't see these results for 2011 until January 2012. I specifically said we would have to wait until the website is updated. For example, if everyone in Frankfurt ran 30 seconds faster than expected (e.g. due to good weather, and some course changes), then Kipsang's time becomes 2:04:11 -- still a great run, but now ranking below Wanjiru (2:04:00), and Geb (2:04:10), in the ARRS adjusted all time performance rankings. Same thing with Makau. (I just noticed that Kipsang's 2010 Frankfurt performance of 2:04:57 is ranked by the ARRS as the third best all time adjusted performance, with a corrected time of 2:04:36).
But I don't have the 2011 data. In the meantime, I made some calculations with the course average, which is the average race bias for all of the previous years for which there is data (e.g. 1973-2010). But don't confuse the specific 2011 race bias, which I haven't seen anywhere yet, with the average race bias over all previous years. One is specific to that day, and represent variable effects, like weather, while the other is an overall average, which tends to show constant effects, like course difficulty.
And third, we are dealing with statistics, which in and of itself, has some inherent uncertainties. Averages tend to dull out individual variations:
a) The bias is an computation of the averages of selected expected times, subtracted from the actual times. The expected time is some theoretical calculation, based on previous recent performances, and an assumption of equivalent fitness. Besides any possible doubts that we can model performance to produce an accurate prediction in the best case, several individual factors such as injury, illness, training, race day nutrition, lack of sleep, etc., may mean the prediction is overly optimistic, or too pessimistic.
b) Even the specific 2011 race bias will be an average over many individual values. It would be unrealistic to think that individual race conditions (e.g. weather) will affect everyone equally. For example, a hot marathon will affect heavier runners more than lighter runners. Similarly, pace-making will benefit a handful of runners, but not everyone in the sample. So the bias will not fully reflect the advantage of a paced race versus a non-paced race.
So *only* if we assume 2011 is typical for Frankfurt, Berlin, and New York, then we can statistically argue that Kipsang has the best all-time performance, even if it fell short of the world record.
Are you saying that you think Wilson Kipsang is the de facto world record holder when you take course/conditions into account?
You can't underestimate the role of pacemaking. Remember that Wanjiru led the pace from very early on in Beijing. He did a lot of extra work that guys like Makau and Kipsang didn't do. G Mutai had an unofficial pacemaker in Boston in Ryan Hall and he didn't really go to the front alone until after 30k today.