I don't think Cannondale comparing bikes in their stock forms (yes, including their stock wheels) is as unfair as some of you are making it out to be. If you really think it through, there is no way to make a 100% "apples to apples" comparison of aero performance across brands other than to (1) test them as they are sold to the customer (i.e. stock), which has the problems mentioned above, e.g. one bike may be sold with more aero wheels than another or (2) try to "standardize" them with the same wheels and perhaps even other components, which would be unfair to certain framesets depending on what wheels/other components you chose, and would ultimately lead you to demand testing of all possible bike/component combinations.
Problems with approach #1: Pretty obvious. Bike A is sold with uber-fast aero wheels (Wheels A). Bike B is sold with slower wheels (Wheels B). Bike A + Wheels A tests faster than Bike B + Wheels B (i.e. as each bike is sold to the customer). Now if Bike B + Wheels A tests faster than the Bike A + Wheels A combo, we say "aha! now we know that the real winners here are Frame B and Wheels A." But that's all we know. We're going to have to test all combinations across framesets, wheels and tires (see below) to know the truly fastest "bike system." And even if you know that Frame B and Wheels A are the winners of our 2-bike face-off, does that mean that Bike B (with the faster Frame B) is "faster" then Bike A? Well no, not when you look at what the customer is buying. In order for Bike B to be faster than Bike A, the Bike B consumer must buy Wheels A and ditch the Wheels B that came with their Frame B. So we end up going back to the idea of trying to tease out the fastest combination of frameset, wheels and tires across all manufacturers, and that hardly seems the same as the question "which bike is fastest?"
Problems with approach #2: So now we say, let's "standardize" things and test all the bikes with Wheels A. Well what happens if Bike C (sold with Wheels C) is slower when you pair up Frame C with our currently "fastest" Wheels A? At some point, different wheels will all interact differently with different frames, different tires, and so if you try to "standardize" these components, you are going to help one frame and hurt another. Really what we would all ultimately like to see would be an "infinite combination" testing, which would lay out for you every permutation of frameset, wheel and tire (including different widths) - but that is just terribly unrealistic.
P5 | 9070 Di2 | SISL2 SRM | Enve 7.8/FFWD Falcon | Roval 321
T1 | P2Max Type S | Rolf Prima FX58/Zipp Super 9 Disc