“Power has changed the language of professional cycling,” quipped physiologist Allen Lim, in a short video for a power meter product he helped to develop. His contention is certainly supported by this year’s Tour de France, where talk of watts and work capacity has saturated the cycling commentator’s lexicon.
But why do we analyse our sport in such forensic detail? Speculation is rife about what constitutes a plausible performance and we now have access to a stream of data to support a range of hypotheses. Cycling’s infamous past has intersected with science and technology to create debating dynamite.
Professional cycling has endured numerous painful culture shifts, perhaps most significant, the suppression of subjectivity, where the drama and uncertainty which has entranced generations of fans has been replaced by objective analysis of power output and efficiency.
The sport faces a challenge as, in contrast to other activities where nuances in technique and moments of inspiration can make or break a competitor, the bicycle fixes the rider into a limited range of motion. This limitation makes the conversion of biological energy into mechanical work very efficient, easy to measure and brutally honest: the rider’s capacity to push on the pedals must meet or exceed the demands made by their competitors and the course.
This relatively simple quantification has made cycling the ‘go-to’ method for scientists to explore the potential of human physiology. You can calculate energy input by measuring the amount of oxygen a rider consumes and the amount of carbon-dioxide he produces. You can gain an insight into what’s happening inside the human by measuring the energy output by recording the speed and force of pedalling with a power meter. You can define the demands of the sport in objective terms by using GPS course data and knowledge of the rider’s physical characteristics and equipment.
The blood, sweat and breath of thousands of undergraduate students, and a smattering of elite athletes, as they have been pricked, poked and analysed whilst spinning away on stationary bikes in laboratories around the world, has created an understanding of performance physiology and a collection of models and statistical norms. But ultimately, this detail all boils down to one simple metric: “The unifying variable of cycling performance” as Cycling Power Lab puts it. This variable is power - the rate of doing work - a rider’s capacity to pedal represents the sum of all energy systems and psychology.
Consequently, cycling may be described in a basic formula requiring energy inputs, energy outputs and energy demands, resulting in a power output. The nature of the sport lends itself to scrutiny as we can use this knowledge to conduct complex analyses of performance.
At its most basic level, power is measured in a unit called a ‘joule’. Following Tuesday’s mountain stage, when commentators wrote about Robert Gesink’s average power output of 409watts on the final climb (as measured on Strava), they are actually describing the fact that Gesink carried out an average of 409 joules of pedalling work per second for the duration of the ascent.
If we go one step further, we can describe how this work is converted into climbing relative to his competitors. Assuming Gesink weighs 71kg (the average of his weight as recorded on Strava vs. the LottoNL-Jumbo team website), the Dutchman needed to produce 5.8 watts (joules per second) for every kilogram of his body mass in order to finish fourth, 1min33seconds behind Froome.
In Gesink’s case, we have the benefit of hindsight as we have access to his actual power data from the climb. However, if Gesink had decided in advance that he wanted to secure the KOM and ascend the Col du Soudet in 39min48sec, he could have modelled this performance ahead of the stage using a website such as cyclingpowerlab.com.
The ‘CPL method’, which has become a popular means to analyse rider performance, combines a number of models of physiology and performance. CPL will take a range of inputs, such as the rider’s best recorded power over various durations, combining this with data concerning the event duration, elevation, rider size, weight, weather, rolling resistance and equipment selection and create a statistical model suggesting the rider’s best possible performance and optimal pacing strategy. If we input Gesink’s pre-stage data into CPL, it estimates a power requirement of 408.8W for the Col - less than one per cent different to that actually recorded by his power-meter and uploaded to Strava.
Knowledge of the ‘cycling equation’ means that we can also work backwards from the performance - the final output - the understand more about the rider. Beginning with a climb of known length and elevation, a rider’s weight and the time they took to climb it, plus assumptions based on norms for efficiency and ranges for friction, we can calculate the power demands. Taking these power demands combined with knowledge of the norms of human physiology from the lab, we can make a reasonably accurate guess as to how much oxygen was required and therefore what the rider’s VO2 max. would need to be.
Unfortunately, this is where power, the unifying variable, becomes divisive. Estimates about how representative these statistical models are range from ±1 per cent to ±10 per cent. Therefore, depending on which validation research you accept, calculations could suggest a performance that ranges from outstanding, but within the range of physiological feasibility at one end of the scale, to ‘mutant’, at the other.
Most scientists accept that there is uncertainty in the analysis and see the results as part of a wider landscape. It may not change their view, but at least there is context. However, the challenge for fans and commentators alike is that the deceptive simplicity of the formula - power supply vs. power demand - and it’s objective results can create a myopic perspective and dichotomous conclusion: a performance is either believable or it’s not. We deceive ourselves that the cycling equation can provide black and white answers when, in truth, we end up with more shades of grey. The input and output may be clear, but the human in the middle is difficult to understand and often unpredictable (just ask Vincenzo Nibali).
Perhaps this objective analysis and statistical scrutiny has failed to suppress subjectivity after all. Drama and uncertainty continue to entrance us in the "world’s greatest bicycle race".
James Hewitt is a performance coach at HINTSA Performance, Geneva