Studying the ascent times realised by professional cyclists in the high mountains is a popular pastime during and after the grand tours.
There are at least 3 points of curiosity:
- Amateur cyclists may be interested to know how they compare when riding cyclo sportives that cover the same roads.
- Velocita Ascensionale Media
(average ascent rate in metres per hour) has been a common metric of riders ability since Dr. Ferrari popularised it several years ago, and...
- There will always be people who seek to use this data to “red flag” certain riders as suspected dopers.
Frankly there are too many variables at play to support the kind of “conviction my numbers” attemped by some (it is incumbent on those who accuse foul play to recognise the burden of
proof beyond reasonable doubt) and we believe there are far more interesting uses for event modelling than second guessing the nature of riders race preparation.
The attraction of ascent time data is due to its purity. The theory says that when a rider is climbing he is moving more
slowly than on the flat and using the majority of his power to overcome gravity, not aerodynamic drag. Thus inaccuracies
inherent in estimating drag and drafting effects are less harmful to the analysis than they might otherwise be, meanwhile
we can usually know a riders weight with a high degree of accuracy and then add 6.8 kilos worth of bike to that number
for some highly accurate modelling.
Observing Ascension Rates
In an ideal world faithful to the modern era of "sports analytics" race organisers would publish riders ascent times –
from the moment they cross a certain point at the foot of a climb to the summit – with the same degree of certainty
and accuracy as splits in a time trial. This would certainly be possible given routine use of transponder based timing systems. Unfortunately they do not and so the process of
evaluating individual riders tends to rely on individual initiative, good TV coverage and a stopwatch. In one interview
Dr. Michele Ferrari
suggested that this ascent timing methodology may be possible up to 60% of the time in major races. From ascent times one can then compute VAM and
estimate power output and watts per kilo using an appropriate course and power model.
Ascension Matrices
The objective of our analysis here is to take the hard work out of ascent analysis by computing matrices of ascent
times for a range of rider weights and power outputs covering the spectrum of body mass and sustainable power output
applicable in the pro peloton. Anyone who has observed an ascent time for a particular rider can then look up his
weight and then match a power output, wattage per kilo and VAM to that ascent time.
These matrices equally serve as estimates of ascent times that ought to be realisable by a rider of a given weight
delivering a given power output and this use of the data may be particularly interesting to amateurs training for
future attempts on the relevant climbs. The message in the numbers is clear –
less weight or more sustainable power
equals faster climbing.
Assumptions
Estimated ascent times are the result of modelling high resolution GPS representations of the climb in question with the following assumptions:
- Weather and atmospheric conditions are assumed to be the ISO “Standard” sea level values with nil wind.
- Rider’s aerodynamic drag (less important on climbs but still relevant) is computed using the anthropometric CdA estimator outlined in our
Aerodynamics Primer
assuming always that the rider is using a standard road bike. The average Body Mass Index of the 300+ professional cyclists in our database is 21.2 (BMI being weight / height^2) hence the
rider height used in this estimator for each weight option in the matrices is (weight / 21.2)^0.5.
- Rider’s weights are supplemented by a bike of 6.8 kilos (i.e. the minimum machine weight dictated by the UCI) plus 1 kilo representing "helmet + shoes + 1 water bottle".
- Coefficient of Rolling Resistance is 0.003, drivetrain efficiency is 98% (see our article on
Drivetrain Efficiency & Marginal Gains to understand why.
- The power options in our matrices assume the ascent is ridden at this average power in a constant fashion with no accelerations.