This year, this citation looks to be the most appropriate, as the final stage of this edition of the Giro will be set in Verona. Right after the stage that will see the group finishing at Passo Fedaia, Marmolada, riders will fight on their own on the hilly ITT of Verona (Sunday 29th). Interestingly the course is a carbon copy of 2019 and, luckily, we have the GPS and power data from 26 riders competing in 2019. What did we learn from the 2019 data? I would like to propose here a few findings from that analysis. The analysis resulted in a scientific publication, and the interested reader can find more info in (⊕Zignoli et al. (2021)Zignoli, Andrea, Francesco Biral, Alessandro Fornasiero, Dajo Sanders, Teun Van Erp, Manuel Mateo-March, Federico Y Fontana, et al. 2021. “Assessment of Bike Handling During Cycling Individual Time Trials with a Novel Analytical Technique Adapted from Motorcycle Racing.” European Journal of Sport Science, 1–9.).

The course

The 17.4k-long course can be split in two distinct sections: the first is on broad and straight boulevards, followed by a 4.5k climb (~5%) with a series of steps, on narrow road. After the intermediate time is taken, at the highest point on the Torricella Massimiliana summit (9.5k) everything changes. In fact, the second section consists in a 4k fast descend, on wide and straight roads, with large hairpin turns. The descent continues with a nice and smooth negative slope along the city streets, with few sharp bends but generally broad, straight urban avenues. The riders will then enter the Verona walls, and they push all the way to the finish in the Verona Arena.

Weather conditions: dry VS wet asphalt

Arguably, the weather conditions will play a crucial role in the race. While back in 2019 the temperature was pretty high, this year we might experience a very different scenario. In the following figure, the weather forecasts for Verona, May 29th 2022, Windfinder (accessed on May the 27th). On the side, I also report the forecasts for the day before, Saturday the 28th. These suggest that the roads might be still wet and slightly slippery on the race day. Wet conditions might be very detrimental for the tire-road friction coefficient, which is the limiting factor to the maximum grip forces that riders can ask to their bikes. A wet road is more slippery (lower friction coefficient) than a dry road, and hard braking actions and maximal cornering speeds (hence maximal lateral accelerations) are much lower. Simulations revealed that the same virtual rider could loose 1’03” in a technical 5k descent on a wet VS a dry road (please see (⊕Zignoli 2021Zignoli, Andrea. 2021. “Influence of Corners and Road Conditions on Cycling Individual Time Trial Performance and ‘Optimal’pacing Strategy: A Simulation Study.” Proceedings of the Institution of Mechanical Engineers, Part P: Journal of Sports Engineering and Technology 235 (3): 227–36.)).

Weather forecasts in Verona, May 28th 2022 Windfinder.

The data process: beyond the Watts

A 3D model of cycling can give us insights on what is needed to excel in the Verona ITT. Data from 2019 were also used in a more technical paper discussing the potential applications of such analyses (the interested reader is invited to read the paper (⊕Zignoli and Biral 2020Zignoli, Andrea, and Francesco Biral. 2020. “Prediction of Pacing and Cornering Strategies During Cycling Individual Time Trials with Optimal Control.” Sports Engineering 23 (1): 1–12.)). The analysis of the power output can provide useful information about the intensity and the effort required by the riders to compete in the stage, but a 3D analysis has so much more to offer. In the following figure, an example about the level of details of a 3D analysis is given for an hairpin turn in the Verona ITT.

Speed and power estimated values for an hairpin turn in Verona ITT. Negative power values indicate braking actions.

The gg-plot: the acceleration signature

If accelerations are available or retrieved from race GPS data, then they can be easily plotted in the gg-diagram (also known as gg-plot, or adherence ellipse). The “g” in the “gg-plot” is for “gravity on Earth,” the little g, which is also used as measurement unit for an acceleration (1g=9.8$m/s^2$). This graphical technique is inherited from motorsports, but it does a good job in highlighting the influence of corners and course geography on the cycling performance. A highly dispersed gg-plot indicates that a rider is willing to sustain high lateral and longitudinal accelerations. Conversely, a very flat and narrow plot, is typical for a rider who is not pushing to the limits. Two corresponding examples are provided on the margin pics, where a comparison between wet and dry conditions is made (x-axis left/right lateral accelerations, y-axis forward/backward longitudinal accelerations).

How a gg-plot might look like for a rider in wet conditions in the downhill section of the Verona ITT circuit.

How a gg-plot might look like for a rider in dry conditions in the downhill section of the Verona ITT circuit.

To have an idea of how gg-plot can be used, I leave here an example from F1 racing. Of course, F1 drivers are subject to much higher lateral accelerations (e.g. in this figure, >5g). Approximately, cyclists might experience maximal lateral accelerations of 1g while cornering.

How a gg-plot might look like during a F1 race.

Lessons from 2019

Power output and speed patterns

Arguably, power output and speed are the most common variables used to characterize a road cycling performance. In the next figure, the average power output sustained by the riders involved in the study is reported together with data dispersion. In the figure, clear distinction between the different phases of the race can be made. The vertical blue dashed lines are delimiting the fast descending technical section.

Average power outputs for the Verona ITT, courtesy of the teams involved in the study (riders n=26). Shaded area highlights dispersion of the data around the average. Vertical lines encolsing the fast descending section.

An histogram can also be used to present the kind of power output that the riders sustained on average in the 2019 race (median power resulted in 353 W).

$Power output distribution, courtesy of the teams involved in the study (riders n=26). Shaded area highlights dispersion of the data around the average (thick red line) and median (thick dashed red line).$ Power output distribution, courtesy of the teams involved in the study (riders n=26). Shaded area highlights dispersion of the data around the average (thick red line) and median (thick dashed red line).

It is interesting to see that during the descent, high power output values were also registered, and this is due to the corners: riders wanted to accelerate out of the corners to gain maximum speeds. Speeds are also provided in the following figure. Oscillations are not due to a particular pacing strategy, but rather to the corners, who impose specific limits to the maximum speed that can be sustained by a cyclist.

Average speed for the Verona ITT, courtesy of the teams involved in the study (riders n=26). Shaded area highlights dispersion of the data around the average. Vertical lines encolsing the fast descending section.

Association between final rank and area of the adherence ellipse. From the study conducted on the 2019 data.

Consideration on adherence: “no risk, no reward?”

Even considered that power output delivery is of absolute importance in time-trial performance (also considering normalization techniques, e.g. on body weight or allometric scaling), in technical races might not be enough to secure the victory. Data from 2019 revealed an interesting association between the area of the gg-plot and the final rank in the race. This indicates that the riders who were chasing the victory were also those willing to sustain the highest accelerations, especially downhill. We might speculate that lateral accelerations are proportional to the degree of risk of falling or crashing. If we agree with this preposition then, unfortunately, data from 2019 suggest that: “no risk, no reward.”

There is an old quote that sounds like: “you can’t win the Giro in the downhill, but you can definitely lose it.” Like it or not, a lot of risk will be involved in the downhill section at the Verona ITT. The association found in 2019 between the lateral accelerations and the final rank in the stage does not imply that a fast and reckless descent will secure victory in 2022, but it clearly indicates that delivering more Watts in the climb and in the flat sections will not be enough. As the GC contender will likely be careful in negotiating every corner, the stage win might as well go to an outsider.

It’s everyone hope, that riders will find a perfect road surface, and no obstructions along the course. As the riders cannot seat on the tube while descending (position has recently been banned by UCI), we will likely witness slower peak speeds compared to 2019. However, when it comes to lateral accelerations, it’s the cornering speed that matters. Arguably, rainy conditions and wet roads might have no influence on maximum speeds on the straight sections, but on maximum speeds on corners, and on perception of safety. With disk brakes used to the limits of overheating (albeit we do not expect crazy-high atmospheric temperatures), riders will have to nicely negotiate the corners, the ‘S’ bends and the few roundabouts at the limits of physics. Some riders will attempt to stay on the aerobars as much as they will be able to, others will use their bodies as wind-brakes and stay on the hoods.

With tires inflated like crazy, lateral forces that will keep the riders on the road will be generated on tiny-tiny contact patches, in those small-small interfaces between road and rubber. In these regards, less tyre pressure might provide additional grip, but of course more rolling resistance in the straight sections. Mechanics will walk a fine line between safety and performance here.

Verona ITT: the perfect combination of power output and bike handling abilities

2022-05-28