Endurance Analytics

Our blog - covering sports analytics and EA products

Coach-rider communication and the need for power

Last year an article in the cycling press asked the question "are power meters killing the art of coaching?" It's conclusion, after some detectable anti-power sentiment, was that coach-rider communication is key and therefore the art of coaching must trump the use of power data, positioned we suppose as the scientific opposition.

Of course the infamous Dr Ferrari to this day includes the caption "coaching is art" on his website 53x12.com and he has a point. Science gives us models of the world, but the best users of those models know how to combine and interpret quite a few, having regard for their limitations and applicability to unique circumstances such as events and riders. This process, intuition, triangulation, call it what you will, is well described as an art form and fits with the role of a coach.

But to suggest that power meters or power data might be somehow diminishing, rather than enhancing or enabling this art, is frankly the wrong question. What makes a coach-rider relationship flourish, as the article points out, is communication. Communication requires language, and we believe that the most powerful language the sport now has is power itself.

Now why is power the key to coach-rider communication?


The language of power based training allows us to be succinct and expressive about the makeup of training sessions, with great accuracy. Imagine, for example, being an engineer before anybody had developed a vocabulary to describe different building designs. You'd be using dozens of words to describe concepts that now need one or two. Training sessions are the same - once a rider understands what is a "2x20 @ 105% FTP" communication is eased and enhanced.


A doctor would have a hard time evaluating a critically ill patient without access to a modern suite of diagnostic capabilities such as blood and ECG data. These tools reveal more about the patient's state than words will ever do and anyone who has ever laid in a hospital bed will recognise that at doctors rounds far more time is invested in checking these diagnostics than asking the patient how he is. In cycling the diagnostic capability of power - pre or post training and pre or post race - simply trumps a lot of verbal communication in it's richness and value. The really important communication then becomes the eventual diagnosis, prognosis or prescription facilitated by the data.

How hard & how fast

Your assessment of RPE is not my assessment, just as your heart rate is not my heart rate. Some people are simply terrible at evaluating how close they are/were to blowing, especially relatively new bike riders. Cycling was crying out for an objective measure of intensity before power came along but coaches and riders must communicate on three major topics: duration, intensity and frequency of training. Only power allows riders to be benchmarked on ability. Power = Speed, with just a few parameters in the middle. Rider benchmarking and goal setting, fundamental parts of the coaching process, are practically impossible without power data. You cannot communicate what you cannot define.

How sick or how injured

Sickness often manifests itself as the inability to achieve a certain intensity while injuries tend to be evident at certain power outputs, torques or cadences. Power data is therefore incredibly expressive when the unfortunate time comes to communicate either of these problems.

Form and history

A rider without a history of power data has no context in which a coach can place him. What is his baseline, untrained level of fitness? How far is he above that, and how much headroom does he have? What types of training has he best responded to? Some coaches would admit that coaching riders without power data feels almost dishonest. Dishonest because it imposes some very real limits to the amount of value a coach can add.

Distance learning

We live in an internet age where a significant number of riders have never met their coach, nor will they ever really need to. Cycling is not football or any other team sport characterised by group, face-to-face training sessions that may or may not end in the pub. Power data is not a web protocol but it might as well be. It opens all kinds of possibilities in terms of two way, coach-rider-coach communication, which nowadays increasingly has to be virtual.

Power is changing the game. Don't be one of the naysayers who claim it's damaging racing or coaching or it might just damage your job.

Ride Review with Post-Event Performance Modelling

The opportunities to use power based performance modelling in a pre-event context, for Goal Setting, Scenario Analysis, Equipment Optimisation (EO), and Optimal Pacing Strategies (OPS) should by now be compelling for all serious cyclists and triathletes. For the price of a few carefully collected details describing a rider's physical attributes, power-duration curve, and bike setup, it is now possible to first predict a given rider's time on a given course and then to apply scenarios on the underling model to make the same rider go faster. Analytical intelligence equals free speed...if you want to put the work in.

While the pre-event context has always been the most compelling there are also significant opportunities to use the same performance modelling techniques in a post-event context. Nobody wins a race by looking back at the day's data but but you can, perhaps, use it to win the next one...if you know what to look for and have the right tools. We have invested a lot of effort recently creating performance modelling tools for the post-event context and this article outlines what they might be able to do for you.

It's not (just) about the bike (data)

Pre-Event modelling combines data about a rider, his or her bike and equipment, the course, and the weather. Weather can mean historical trends or the actual race day forecast. We then simulate the rider's performance on the given course, under the given weather conditions. A simple simulation, based on the maxium power a rider should be able to sustain for the course duation gives us a time-on-course prediction. Many variations of power output, sector by sector, allow us to find an Optimal Pacing Strategy. Variations of rider parameters or equipment allow us to see the effect of scenarios. Variations of weather allow us to build a complete picture of which equipment is "probably" the fastest, on a "normal" day.

Post-Event modelling replaces weather forecasts with actual, observed weather. It also takes the ride power and speed data (via a .fit. .srm. .tcx, .pwx or .gc file) as a precise description of how the ride was executed. This gives us analytical possibilities and opportunities to combine and compare "model" versus "actual" performance not possible in the Pre-Event context.

Performance Modelling is more than race planning

Some people see performance-modelling as a way to predict race-day performance and then create a plan. That is Pre-Event modelling, but there is more to offer. Post-Event modelling, with the extra second-by-second information provided by the ride file, has some powerful applications. Let's review some of them:


The Virtual Wind Tunnel has never been more available

In our experience the most versatible, and now most used, field test protocol available for determining a rider's aerodynamic drag is the "Virtual Elevation" method a.k.a. "Chung" method (after it's inventor Robert Chung) a.k.a. the method of the "Virtual Wind Tunnel". Some very serious operations routinely use this method, in a velodrome setting, to provide riders with drag data and aerodynamic optimisation advice. In simple terms this methodology takes ride file data (power and speed) to build a "virtual elevation" (VE) profile for the ride. If you vary CdA, or let a computer do that for you, until the VE profile closely matches the actual course elevation profile of the same ride, then the CdA you finish with can be a very good estimate of the rider's actual CdA.

Post-Event Modelling, by combining the rider data, course data, ride data, and environmeental conditions, already has all of the information necessary to estimate CdA using the VE methodology. There is one caveat - wind. Rather than assume zero wind, as velodrome operators can, making VE available for any rider requires us to factor in the wind. Since we already have weather data, at times of day covering the event, and we know rider's time's of day on all sectors of the course, from the ride file, then we can do a reasonable job of factoring in air density and effective winds. VE under observed winds isn't perfect, but if estimating CdA is important to you it is better than nothing...a lot better!

Physics + Power Data = Power & Energy Attribution

Imagine a ride file says that a rider averaged 200 watts on a short climb. Now imagine we model that rider's performance on the same climb, we take all of his parameters, then plug his time into a "power given speed" model to estimate the power required to complete that climb at the given speed. If we're using the right parameters, including VE CdA, then our estimate will be 200 watts. But we have more information... Those 200 watts are composed of some watts required to overcome gravity, some to overcome aerodynamic drag, some to overcome rolling resistance, and a few others. By combining the actual power data from a ride file with theoretical power demands from a performance model we can achieve what we call a "Power Attribution" or "Energy consumption study". (Remember that energy, in joules, is just watts multiplied by time in seconds). This is an interesting capability of Post-Event Modelling.

Pacing Quality Evaluation - How well did you execute?

The identification of theoretically Optimal Pacing Strategies is one of the most compelling applications of Pre-Event Performace Modelling. But how can we know how well a strategy was executed? Sure, we can compare race normalised power with the prediction from the plan, and we can eyeball ride power and speed data relative to the sector by sector plan, but we felt that some better tooling was required.

"Pacing Quality Evaluation" is a software based comparison of actual, sector powers and speeds, from a ride file, verus a theoretically optimal plan calibrated to the rider's power-duration curve as exhibited in the ride it'self. Having looked for some magic number to describe pacing quality we eventually settled on the one number that counts...time! This analysis can also highlight something very interesting - good riders can be very good at pacing. Often the way a rider executes, with the freedom to vary power on every little hill or under every little gust of headwind, results in a time that is better than what we might predict from an "optimal" strategy simple enough to understand and execute. Nevertheless, Pacing Quality Evaluation is a fantastic gross error check and has some great applications in learning lessons and executing better in the future.

Rider Benchmarking with Virtual Competitor

We have always emphasised the application of performance-modelling in the key coaching disciplines of rider profiling, performance benchmarking, and goal setting. It is a powerful thing to be able to compare a rider's actual performance against some theoretical competitor, a better version of the rider's future self, or even a worse version of the rider's past self. A virtual competitor study, in the post-event context, makes this possible.

Real World What-If Analysis

Have you ever agonised about the race you or one of your riders lost, perhaps by a matter of a few seconds? Wouldn't it be great to know the size of the winning (or losing) margin, not in terms of time but in terms of parameters you can work to improve such as power, weight, or drag? Post-Event what-if analysis does exactly this because we can re-simulate performance given certain improvements. No more agonising.

Post-Event modelling is now available at our usual Performance Modelling link. Simply select "Post-Event Modelling" near the bottom of the page. CdA estimates via the virtual wind tunnel are open and free. The full Post-Event report, with all of the above analytical features, can be configured below.

Intelligent Event Selection : Comparative Advantage & CPLPercentile

The concept of "comparative advantage", attributed to the economist David Ricardo in 1817, says that it is better to capitalize on your strengths than to shore up your weaknesses. Back then Ricardo was pointing out the futility with which England might compete with Portugal on wine production and how the English should instead leverage their textile industry to trade cloth for wine, but there is far wider relevance across many aspects of life. In competitive sport, and especially in a sports like cycling and triathlon where there are a multitude of different events and courses available, the idea has a lot to offer.

In our blog article on the measure of a cyclist we outlined 8 metrics of a cyclist's ability. Strengths or weaknesses in each one of them will tend make a rider more or less suited to specific events. For example, a rider with a high FTP/Weight ratio will probably be good in events with long climbs. A rider with high AWC/Weight (Anaerobic Work Capacity) and fast AWC recharge may be better in events with lots of short, punchy climbs. A rider who is relatively heavier but also powerful is likely to have a higher power to drag ratio "Watts/CdA" and excel in flatter time trials. The full set of metrics will interplay to make every rider uniquely better or worse than another rider in every event.

Imperceptible Differences

Identifying comparative advantage is not so easy. Of course we can immediately say that a Fabian Cancellara is better suited to time trials than summit finishes in the Dolomites. And that a Nairo Quintana is unlikely to win a world championship time trial. In between these extremes, however, there is a very large population of more "average" riders whose morphology is similar, whose weight is visually similar, and whose only differences lie in the subtleties of their power numbers, their actual weight, and their aerodynamic drag. This population includes you, your competitors, or if you are a coach then your athletes and their competition.

It is not just riders who are often imperceptibly different. The same goes for courses. A typical, rolling, neither extremely hilly nor extremely flat course may suit one rider more than another, or it may not. To the naked eye and the human imagination there is no way to tell. It's a similar problem to deciding between a light climbing bike or an aero time trial bike for a rolling course. Apply some Performance Modelling though and the answer, rider versus rider, course versus course, just falls out in front of you. All you have to do is simulate both riders then compare their time on course estimates.

Comparing Riders

We have always advocated using Performance Modelling to compare any two riders on a particular course. For example, given metrics for "Rider A" it is very easy to simulate his/her performance on the course in question, then to simulate an alternative or competitor "Rider B", and decide who has comparative advantage. If you were managing a team then you might use this sort of approach to recommend one rider over another for a certain event. If you are a coach then you might use this approach to estimate the shortfall between your rider and a competitor. The latter "gap analysis" is the foundation of the key coaching practises of rider profiling, performance benchmarking and goal setting.

Comparing Events

A logical extension of Performance Modelling is to change the problem, forget about the specifics of any one alternative rider or competitor, and ask "which of several events should this rider enter to maximise his chances?" This is a question equivalent to the problem (for time triallists) "on which course can I maximise my ranking?" or (for triathletes) "which course gives me the best chance of qualifying for Kona?" Until recently we didn't have an answer to that, but now we do.

CPLPercentile is a metric we have developed to indicate a rider's comparative advantage on a specific course. Very simply: Every time you use CPL to simulate a given rider on a given "official" course we show a time on course estimate. Below that we show a "CPLPercentile" which is a number from 1 (best) to 100 (worst) indicating the simulated rider's ranking relative to the theoretical performance of every other rider in our database. Without going into numbers that is now a very comprehensive population of average and not so average riders. The actual value of CPL percentile doesn't matter. It could be 25 (stronger than average), 50 (average), or 75 (weaker than average). The reason it doesn't matter is because it's simply an indication of where the rider might finish in a typical field of typical but anonymous competitors. What does matter though, and what is a very powerful indiation of which event a rider should favour, is the way CPLPercentile changes across events.

Intelligent Event Selection

By example: Let's say we simulate a rider in Event A. The CPL percentile is 50. This means that with a large number of typical competitors the rider might finish exactly mid-field. We dont' care aboout that the absolute level because we dont know who the field are. Now we simulate the same rider in Event B. The CPL Percentile is 45. Again we dont care about the absolute level. What we do care about is that the percentile is 5 lower. The rider's attributes are giving him/her a comparative advantage in this event. The lower the percentile the higher the comparative advantage. Logically the rider would choose Event B to maximise his/her chances. It's that simple.

In conclusion, CPLPercentile is a new metric to indicate rider's comparative advantage across events. It has become possible thanks to the large population of riders in our database, whose attributes we compare 100% anonymously, and a fair deal of computing power. Finding the event where a rider's metrics give comparative advantage really is as simple as checking the CPLPercentile for two or more candidate events and favoring the one with the lowest CPLPercentile. CPLPercentile is visible absolutely free from our Performance Modelling resource. We would of course be keen to see any questions or feedback.

Mean Maximal Gradient, Predictive Cadence and Gearing Analysis

Bike riders and coaches of all levels understand the importance of informing themselves about the demands of a forthcoming event or race. At it's most informal this process might be a quick glance at the route and profile, while at it's most rigorous a formal analysis of demands may include power analysis, pacing and equipment options, nutrition and hydration strategies. CPL's Event Models are ideal for this level of detail.

One of the factors frequently overlooked, or simply not considered in sufficient detail, is analysis of gradients and gearing. Sure, it's easy enough to look at a profile and see the length and average gradient of climbs, maximum gradient is often indicated too, but is this enough?

Most Gradient Analysis Is Misleading

A number of software products will readily show you a bucketed summary of "Total Distance at Gradient", either for a particular climb, or for an entire course. By "bucketed" we mean they will break the road down into small sections, calculate the gradient on each, then add up the total distance in sensible gradient "buckets" or bands. For example, 300 metres of road at 1.5% gradient goes in the 1-2% bucket, another 200 metres at 1.8% goes into the same bucket (the total distance in bucket total is now 500 metres), and so on.

The main limitation of this approach is that it fails to consider how close or how far apart these sections of gradient will be hitting the rider, or their order on the road. 1 kilometre of 10% gradient sounds pretty tough, but if it's really 100 metres of 10%, on 10 separate sectors, with several hundred metres of flat or event descent in between, then it becomes clear that this sort of analysis is at-best misleadiing and at-worst useless.

There is a paralell here with power and heart rate analysis. 10 minutes above FTP or above threshold heart rate tells us little about a ride unless we know how that time was distributed. Was it continuous?. Was it just a few seconds here and there, many times over? Gradient analysis has exactly the same considerations and what we really want is something equivalent to "Mean Maximal Gradient".

Max Distance At Average Gradient

A better way to do gradient analysis is in terms of "Max Distance At Average Gradient". Define some gradient buckets, as above, but this time scan the entire course looking for the maximum distance with an average gradient falling inside that bucket. By example, a 100k course from point A to point A is going to have 100km at a 0% average gradient. If the biggest climb is 10k at 5% then the max distance at 5% average is 10k and the max distance at -5% average may also be 10k. If that climb has a steeper section peppered with short pitches varying between 6 and 8% this analysis will reveal the max distance at averages of 5-6%, 6-7%, and 7-8%. Total distance in any of these buckets, by comparison, could be misleading or useless.

"Total Distance At Gradient" is an easy analysis to make, in fact it can be done manually, "by eye". "Max Distance at Average Gradient" is much tougher for humans, but relatively easy for a computer, which is well suited to scanning and calculating all continuous combinations of all sectors across an entire course.


CPL Pre-Event Models now include an analysis of Max Distance At Average Gradient in addition to Total Distance at Gradient.

Predictive Cadence & Gearing Analysis

The logical goal of gradient analysis is to make informed decisions about gear choice, given some range of acceptable cadence. But there are no good "rules of thumb" about the right minimum (climbing) or maximum (descending) gear for a certain course or certain gradients. Cadence is a product of road gradient and speed on the road. Speed on the road is a product of a rider's sustainable power output, weight, and to a lesser extent other components of resistance such as aerodynamic drag and rolling resistance. A 60 kilo pro who can hold 350 watts on a climb is not going to need the same minimum gear as an 85 kilo amateur with 200 watts.

In practise this means that the only way to make good decisions g is by relying on predictive event modelling. You want to know 1) what are the maximum and minimum average gradients encountered for at least some minimum critical difference. A 500 metre pitch works well for most riders. And 2) What cadence would a certain rider achieve on these pitches, through a range of different minimum and maximum gearing options. Only then can gearing decisions be considered optimium, adequate, or in some cases low enough to avoid catastrophe!

All CPL Pre-Event Models now include an analysis of minimum and maximum cadences achievable with a range of gear combinations including chainrings: 34, 39, 42, 50, 52, 53 and sprockets: 27, 25, 23, 13, 12, 11.

Think power is complicating cycling? Think again.

People make a lot of excuses when they dont want to pay for something. It's a waste of money. I dont need it. What would I want with one of those? We've heard all of these procrastinations from cyclists and triathletes on the verge of buying a power meter. Another objection that we hear a lot is "it's too complicated". The purpose of this article is to highlight why power is in fact simplifying, not complicating, cycling.

"The amateur wants the gimmick but the expert sees the beauty in simplicity"

There are a lot of gimmicks in cycling. Go-faster wheels, frames, and components, for example. Some of them are worth your money, but some of them aren't. And sometimes you can only take that decision when you know what magnitude of improvements you are looking at. At CPL we use performance modelling to evaluate potential gimmicks in the simplest possible language. We use science to reduce cycling to a set of energy demands (climbing, aerodynamic resistance, etc), a set of scenarios relevant to you, and then the simplicity of the answer - speed effects and time savings.

Performance questions

At CPL our guiding principles and objectives correlate well with the idea of "performance questions", a term we first heard from the E.I.S.

  1. What data do we have or can we get?
  2. What questions do we have?
  3. What really matters? And the answer is almost always "speed!"

There is one phrase of which we all should be aware when discussing the merits of equipment, strategies, or riding scenarios and that phrase is "it depends". Too often "it depends" means "I dont know", even when it's coming from a supposed expert. One of the reasons we created CPL was a chronic tiredness with "it depends" and a need to properly answer performance questions.

Seek clarity, not fuzziness

A lot has been written about the finer aspects of the coaching process, in cycling, in triathlon, and wider. But for sports which are supposedly so advanced some really important parts of the process are still in the dark ages. Only power gives riders a concrete evaluation of current ability and concrete assessments of progress. But accurate goal setting requires power based predictive performance modelling of the type that we've developed. Fuzzy goals lead to fuzzy preparation and fuzzy performances. Fuzziness is hard to understand, whereas things that are clearly defined promote simplicity.

Mind your ifs and buts

On the quest to become, or train, a better bike rider there are a lot of if's and but's. "If we drop 5 kilos, what's the benefit?". "If we find 20 watts, what's the effect?" Too many if's and buts make life complicated. At CPL we prefer simple answers which is why we apply performance modelling for scenario analysis. We coin the term "Mass Delta" to provide a simple metric of how much faster a given rider will go, on a given course, with a reduction in weight. "Power Delta" is the equivalent, when the "if" is increased power. These sorts of capabilities, rooted in power, undoubtedly simplify cycling.

The new language of cycling

Power has not just changed the language of cycling, it has created it. Coaches and riders can now communicate with an accuracy and expressiveness that was never before possible. We now have names for training patterns, like "2x20" and "FTP", that convey the same details to us as "suspension" or "cantilever" might convey between engineers. Sensations of form or lack of form can be communicated in terms of ability to hold a certain power for a certain duration. And injuries can be described in terms of the power output or torque that causes the problem. Many of us have have come to take this language for granted, but how often do we stop to reflect on just how much power has in fact simplified our lives?

Put a number on it

The things that win races - times and speeds - are measured in numbers, as are finishing positions. If these criticial aspects of success can be defined with numbers then why not the things that make them possible? Form, fitness, weight, endurance, explosiveness, etc. Plan your race or your season, dont't speculate. Power and all of the numbers that follow are there to simplify life, not complicate it.