I wrote it before - but no way that prime Boonen or Cancellara put in this year's edition would finish OTL.
Roubaix is just a race where average speeds are extremely volatile and depend on a multitude of factors. I generally see the uptake in speed as well across races but in the very specific case, its just also very selectively taken data points.
Or do we also think:
Cancellara would have finished the 1948 edition OTL - because thats what the data would indicate (without me knowing the 1948 course. But just as illustrative point - the edition that year was also 4+ km/h faster than 1947 and 1949)?
Reasonable people will agree that one data point is not conclusive, unless it is really off the charts. But that is exactly what the apologists do; they focus on single data points and attempt to explain them with the usual nonsense: a once-in-a-generation talent (all six of them), marginal gains, weak competition, the weather, and so forth. As if we are all idiots who have no awareness of those things and they are here to enlighten us.
All of these wiseguys ignore the context:
1. It's not an anomaly, it's a pattern. Whether it's about climbing times or average speeds, records are being broken left, right, and centre.
2. This increase in the level of performance did not happen gradually, it was an abrupt change that occurred around 2020. This is well-documented on this forum, for example here:
https://forum.cyclingnews.com/threads/power-data-estimates-for-the-climbing-stages.8839/page-138
3. Along with the increase in speed, we also saw a remarkable change in the nature of performances:
- The era of specialisation is over; we suddenly see riders who excel at every kind of terrain and discipline, winning combinations of races that were previously thought to be practically impossible to win.
- These same riders also experience almost no changes in form; they are able to perform at an exceptional level the whole year around.
- They are able to do these things while riding a lot faster than the previous generation, breaking record after record.
- All of the elite riders excel at time trialling, regardless of body type. Races are now frequently won through incredible individual efforts.
4. Despite the large increase in the general standard of performance, we suddenly see extremely young riders, even teenagers, win major races.
5. All of this began to manifest itself at the exact same time.
Any proposed reason for what is happening must be able to explain all of those things.
As for the graphic, we can see that the average speed in fact gradually became less volatile. We can also see that it was stable from 2012 to 2019. There was no race in 2020, and in 2021 we had the first rainy edition in a long time, which suppressed the speed. The massive increase in the level of performance since 2020 should be obvious to anyone but the blind and cannot be explained by "marginal gains".