The End of FTP?

evolution of power-based testing methods and metrics in cycling coaching and where we might head next

In the car-park or village hall after a race you're sure to hear riders comparing their powers during the race, or maybe their FTP (Functional Threshold Power). Functional Threshold Power is described as the power output that can be maintained as a steady state for one-hour. This coincides nicely with a 25 mile time trial in the UK cycling scene and the typical duration of a professional time trial, or maybe your typical criterium of roughly an hour's duration. This has obvious appeal from a practical point of view, even if a criterium and a time trial couldn't be more different in terms of their demands!

While a higher FTP might bring bragging rights in the post-race chat, you'll notice that this doesn't always match the race results themselves. Using the time trial and criterium examples from before, a high FTP in a time trial is only as good as the aerodynamics of your riding position, a much harder concept to measure than simple power to weight ratio or w/kg. In a criterium or a road race, the majority of your time is spent either much above or much below your FTP, especially the decisive moments. This 'stochastic' nature of bunch racing as well as the importance of aerodynamics and drafting explains why a rider with a much lower FTP can regularly beat someone who 'on paper' is a much stronger rider.

The popularity of the FTP concept in cycling has led to different testing protocols and determination of training zones based on this principle. Previously, training zones have been based on either a maximal number, such as a percentage of maximal heart rate or broad bands based on physiological thresholds determined in a lab setting using blood lactate or ventilatory measures. The different methods of calculating these physiological thresholds is a whole different article (or two) on it's own, but FTP is based on being an estimate of the higher of the two widely measured physiological thresholds, we'll call it the second threshold.

The most popular form of FTP test involves a 20-minute all-out test effort following a maximal 5-minute effort as part of a warm up/to pre-fatigue yourself. You then take 95% of that power as an estimate of what you could maintain for an hour, i.e your FTP. While in a large group of riders, that value of 95% of 20-minute power might be a good representation of their average one-hour power as a whole group, individually you could be someone for whom the true value is 97% or someone where it is more like 90%. This depends on, for example, the anaerobic capacity of an individual. Taking two individuals with an identical 20-minute power for example, the rider with a highest anaerobic capacity would be expected to have a better 5-minute power whereas the rider weaker anaerobically would be likely to have a greater one-hour power (and therefore FTP). This highlights a flaw with basing training zones on FTP, or or indeed anchored to any single point. We expect training zones to elicit a specific adaptation, but in the case of an anaerobically gifted athlete, due to their over-estimated FTP they may be training too hard in their endurance sessions to get the aerobic benefits desired or come out of those sessions too fatigued to do their next high intensity interval session with sufficient quality.

over other durations. This is very useful for training prescription as well as, for example, pacing in a time trial race. Some analysis software, such as the Trainingpeaks WKO5 software we use, uses this concept to construct a power-duration curve (see picture) showing an athlete's predicted best power calculated from their best Maximal Mean Power output (MMP) achieved in racing, training and testing over a defined period. This is a very useful tool as long as an athlete has recent maximal efforts over a variety of durations, and can also indicate when would be a good time to include some more formal testing either in the lab or using the Critical Power model. The best combination could be submaximal testing in a lab to establish training zones below the Critical Power, alongside field-based Critical Power testing to fine-tune high intensity interval prescription.

Where next?

In recent years the level of understanding of power-based metrics has been driven by the concept of FTP, and various measures derived from it. Training Stress Score (TSS) for example attempts to integrate intensity and duration to compare the overall difficulty of training sessions that on their own might be very different, for example an hour-long interval session versus a 6-hour endurance ride. This is based on a one-hour effort at FTP intensity being worth a score of 100 TSS. A longer, less intense session such as 2-hours at the upper end of zone 2 might rack up the same TSS points but the physiological effects and recovery required will be markedly different between those two sessions. A more useful metric might integrate intensity, duration and other factors such as your recent training history and recovery (monitoring sleep and HRV?) or maybe fuel depletion, core body temperature or stress hormone levels using newly available non-invasive monitors, if they prove reliable, for a more complete picture.

Mass data analysis and machine learning could reveal relationships between training and cycling performance on a mass scale given the huge amounts of data stored on platforms such as Strava. What had a rider been doing in the hours, days and weeks before taking that KOM? And would it work for everyone? A smaller scale example of this is the work of Dr. Stephen Seiler and the concept of Polarized Training. In the first part of this article we discussed setting training zones based upon either a single anchor point such as FTP or on physiological variables such as Lactate thresholds. In his review article Dr. Seiler used blood lactate markers to define three Heart Rate based training 'domains' in elite endurance athletes. The moderate intensity domain, equivalent to an intensity eliciting a blood lactate concentration under 2 mMol/L, the heavy intensity domain, between 2-4 mMol/L lactate and the severe intensity domain, where blood lactate concentration exceeds 4 mMol/L. Analysing training studies in a variety of elite endurance athletes from runners and cyclists to rowers and cross country skiers Seiler concluded that elite athletes in all these sports performed around 80% of their training sessions in the moderate intensity zone while approximately 20% of sessions emphasised the severe intensity zone using for example high intensity interval training. In the run up to competition this polarization increased, with even less time spent in the heavy intensity zone. While this kind of retrospective analysis doesn't prove causation, and by definition elite athletes are genetically different to most of us, this does indicate what might be possible on a larger scale. While artificial intelligence and machine learning is progressing rapidly, the skill of a good cycling coach is to integrate this research with an individual rider's characteristics and be creative, not just at the level of a single training session but over time, to maximise their development.

While the understanding of how to use power meters has come a long way in recent years we are still learning more. The integration of additional sensors to measure aerodynamic characteristics, the body's energy stores, stress level and temperature regulation could offer new possibilities for development of useful metrics to inform your training and racing and maximise performance. Here at Custom Cycle Coaching we use the latest research in cycling physiology to inform our cycling coaching plans. We use both lab-based lactate testing at our Birmingham base, as well as field-based tests to establish training zones and to set training power goals maximise your performance. If you'd like to find out more please get in touch.