Don‘t know where some of these numbers are from, but some are certainly incorrect. This is the official results from their recent testing (I bought the mag, so I am very certain these are real):
HOWEVER, take all of these with a grain of salt. Don‘t believe any statistics you haven’t tempered with yourself. You can hardly take any of these at face value.
For example: between 2017 and 2019, all bike tests where done with standardized wheels (ZIPP 404 Firecrest), to only have the frame results. When considering, that frame manufacturers now develop wheels and wheel concepts to work with their frames (Rapide for SL7, Hollowgrams for Sy6, narrow wide for new Aeroad), then it is detrimental if tour changes things around. Also, how these results translate to real world results, is highly questionable.
Also, very importantly, not all bikes are specced as they should be.
All bikes in this test have disc brakes, but the Pinarello Dogma F12 (211 Watts). Disc brakes usually cost you 3 to 5 watts in a wind tunnel at 45kph, so it gives a „too good result“ in comparison to the other bikes here.
The result of for the Madone smelled foul on face value, and that is because it is:
The bike they used has a non-stock stem/ bar combo that isn’t fully integrated, has 28c tires (that alone probably costs 2-3 watts) and is also the largest frame in the test at size 58. All the other bikes in the test are size 56/Medium, and the Ridley and the Eddy Merckx are even Small. What the f?
They say it is because Pedersen is a tall rider and rides a large frame, but this clearly isn’t Pedersens bike. Firstly, I think he rides a 56, and then, this has 48/35 chain ring, while Pedersen rides a 54/41 for flat races and a 52/39 for hilly ones.
So, really really f‘ing difficult to compare.
The SL7 did the same test at 210W (with it‘s deep wheels and standard spec hand bar/tires etc, obviously playing to its advantage). I am in good contact with a German pro cycling team (not WT) and they recently did aero testing on climbing bikes, and found that the Trek Emonda (2021) with integrated cockpit and their sponsored wheels (I think they use the 50mm DT Swiss) tested marginally slower (2 or 3 watts) than the SL7 S-Works with stock cockpit on the same wheels. Don’t know how the protocols compare, but that is the best I know.