Want to know why the InfoCrank is the most accurate power meter in the world?
This three-part video series covers the 2% myth (the accuracy claim made by numerous power meter devices currently on the market) as well as what validity is, what reliability is and why they are important if you want to be able to trust the data from your power meter.
The series is hosted by Brad Hall, Managing Director of both the Veris Racing team and the Exercise Institute, based in Perth, Australia.
Part one of Brad’s guide to what’s important when it comes to the accuracy of your cycling power meter highlights what you need to consider if you’re thinking of investing in one. This episode reveals the 2% myth – the accuracy claim made by numerous power meter devices currently on the market.
Part two of Brad’s guide to power meter accuracy demonstrates why the InfoCrank is the only power meter you can trust. In this episode, Brad talks about why validity is important and needs to be considered if you’re thinking of investing in a power meter.
The final part of Brad Hall’s three-part series looks at reliability and what you should look for if you truly want to be able to trust the data from your power meter. Brad covers all in this last installment.
Brad Hall is managing director of both Veris Racing and the Exercise Institute, based in Perth, Australia. Brad has undertaken extensive research into power meters – how accurate they are in comparison to how accurate they claim to be. Brad took some time out to share his thoughts on what should really be considered if you want to be able to trust your power meter.
Often when we talk about accuracy, one of the big subjects that comes up as a means of proving accuracy is the 2% myth. Now, I call it a myth because certain power meter brands claim to be accurate within 2%.
However, the accuracy itself isn’t well-defined, we don’t actually know what the 2% is referring to – is it 2% of validity, is it a 2% reliability over time? We don’t know. So the concept of ‘within 2%’ is quite difficult to grasp from a scientific perspective.
In research settings, empirical and scientific circles, two important aspects that must be applicable to any study are validity and reliability – and these are two of the ways to measure accuracy.
Validity refers to how accurately a method measures what it is intended to measure. If research has high validity, that means it produces results that correspond to real properties, characteristics and variations in the physical or social world.
For example, if we are measuring 100 watts on a power meter, are we actually measuring 100 watts in the real world? To do this, you would test a power meter against a known measure.
Example of high validity
Example of poor validity
Reliability is defined as the probability that a product, system or service will perform its intended function adequately for a specified period of time, or will operate in a defined environment without failure.
For example, when we measure 100 watts at a given time against a known measure, if we were to measure it again in five days’ time, would we get the same result?
Example of Reliability:
In addition to validity and reliability, there are other forms of accuracy measures, including external validity. External validity refers to your ability to generalise your experimental results across populations, places and time.
What happens when the power meter encounters a temperature gradient? Metal generally contracts/expands, as do the strain gauges that are providing the power measurements. Most power meters run a series of algorithms to overcome these issues of expansion.
The InfoCrank power meter doesn’t. The InfoCrank doesn’t need to run algorithms because the hardware itself is made to only measure force through a specific plane of motion, which is not compromised by temperature gradients.
The above reliability graph demonstrates the current crop of latest generation power meter devices on the market; InfoCrank is the red line. What happens when devices which are already unreliable encounter a temperature gradient, is anyone’s guess.
There is a reason why institutes around the world – such as British Cycling and the UCI – are turning to InfoCrank. Having attended the last cycling science conference in Brussels in 2019, I heard from the bulk of head coaches at pro tour levels of employment. Most had completed higher levels of tertiary education (masters/PhD) as a minimum standard. I believe this demonstrates a need for cycling to hone its ability to measure performance specifically and accurately using tools that are indeed verified to do this. As the sport continues to reach for 1-2% it will require devices that can measure with greater accuracy rather than appeal to hypothetical marketing concepts such as ‘within 2%’.
In some ways, accuracy is the easiest thing in the world to describe. Something is either true or not. 300 watts is 300 watts or it is not!
However, the concepts can be quite confusing sometimes, so I just want to work on one issue today. The question is, “Can you measure the radial forces?”
Firstly, what are the radial forces?
They are the forces that are applied to the crank which presses it inwards or outwards, usually by foot placement at different times during a ride or sprint.
The short answer is that InfoCrank does not measure any forces except the tangential force – that’s what drives the bike forward and constitutes the energy the cyclist applies in order to move the bike.
Technically, this is because our software, strain gauge placement in a Wheatstone bridge formation and bonding techniques all combine in such a way that no other force can register. No; we do not measure radial forces. In fact, we nullify them completely.
Just assume for a moment that we built an InfoCrank that did not nullify the radial (and other) forces? What most people do not realise is that we would still just get a number purporting to be watts. What we would not know is how much of that number was actually watts and how much was something else.
Our rider may sprint and his foot may move outwards which would show up as increased watts but no one knows if it is or not.
Therefore, the only way to measure radial force is to purposefully measure it. Measuring all the forces and then trying to build out algorithms to separate them is one of the main reasons most power meters are vastly inaccurate in real life. Real life is complicated.
With an InfoCrank, no matter how the pressure is applied to the pedal, the power result is the same. It does not differ if your foot is 25mm from the crank or 35mm, or if you have a different pair of pedals. 300 watts is 300 watts – that is your energy put into making the bike go forward.
But what if we built an InfoCrank in such a way that we measured both tangential and also radial force? Now our 300 watts is going to vary up and down and we are going to have no idea how much of that is due to differing foot pressure and how much is driving the bike forward. We are going to have to build an algorithm and assume all sorts of things such as riding style, leg tiredness and seat height to try and isolate out the radial force. What a mess!
So, no we don’t measure radial force or any other force except tangential. If you want us to, we will build a specially instrumented crank or pedal to do so but that would be when you discover that no one really wants to know radial force and certainly they do not want to pay for it…
British Cycling has tested the InfoCrank to confirm its accuracy, with the results surprising even some of the best cycling experts in the world.
At Verve Cycling, we’re often asked what it is exactly that sets the InfoCrank apart from most other power meters. This isn’t a simple question to answer, as there are many more basic ideas that need to be understood before the question can be formulated correctly, however we can go some way to answering this question by considering the practices of British Cycling, one of the teams supported by Verve Cycling.
British Cycling are the owners of more InfoCranks than any other globally and they test each of their cranks for calibration (accuracy) before they go into service. One of their testing protocols involves a static load step test. As the name suggests, the test involves suitably fixing an Infocrank and gradually increasing the load at the pedal axle.
As you would expect from the most successful track-cycling nation of the last 15 years, they have performed this test on different power meters. However, we can only describe the results seen on an InfoCrank, which are concisely depicted by the graph below (showing time in seconds on the x axis and power in watts on the y axis).
Two things stand out in the tests of InfoCrank compared to the best of the others:
The error is consistent (compared to the testing equipment which itself is independently tested and certified to a high degree of accuracy) across the entire range. The standard range for testing is from 0-3000 watts at 60 RPM (about 500nm of torque).
The InfoCrank responses are not only correct, but instantaneous
(We explain why Verve Cycling talks about error as opposed to accuracy here – but suffice to say it is because it is the correct way to characterise the margin between the real or true value and the reported value).
Point 1) above is a well-known feature of the InfoCrank. This consistency is because the Infocrank is designed to track reality via design, as opposed to approximating reality via an algorithm, which is the territory of most other power meters. With the InfoCrank, we have seen errors as low as 0.35% compared to true across the entire spectrum of potential uses. It is also something that we work very hard to maintain on each new variation of InfoCrank.
We also ensure that this error does not deviate due to external forces and so there is no ongoing need to correct design inadequacies by a need for continual ‘zeroing’ – we can say with confidence then that all InfoCranks report virtually the same numbers for the same force.
For some further visual representations that help understand the above please see the below graphs, provided to Verve Cycling by the British Cycling Research and Development team, to demonstrate actual InfoCrank tests compared to true and compared also to other InfoCranks!
What we can see is that when a given load is applied, the InfoCrank actually produces a value which represents that load to within an extremely small margin of error.
Table 1. Data from two InfoCranks tested on the British Cycling rig across a range of powers
Point 2 above was surprising to the Research and Development team at British Cycling when they first observed it. They were used to testing other power devices which took a large percentage of the time (seconds) to stabilise in reaction to the load applied. What they noticed with the InfoCrank was that it stabilised instantly on each load step. This means each data point (every four milliseconds) is accurate and able to be read and reported!
Many power meter users would either quite reasonably expect reality to be tracked moment-to-moment in the case of every power meter (it absolutely is not!), or never think to consider it. Clearly such issues with stabilisation described above will have a massive impact on the quality of the data your power meter provides – if your power meter cannot report values immediately in a static load test imagine how it performs when trying to report values in a pedalling environment. The below graph with comments shows visually how quickly the InfoCrank stabilises throughout a static load test.
Figure 2. Power during a static load step test with a simulated torque of 60 rpm. Showing low latency in response to changes in load throughout the test (circle).
British Cycling comments: There is virtually zero overshoot or undershoot as each load step plateau is reached. This gives us confidence in capturing instantaneous power during training and events.
So, to answer the initial question – what is it exactly that sets the Infocrank apart from most other powermeters? – we should now be able to see a couple of things. Firstly, it measures what it is designed to measure, and it does this consistently, unlike many other power meters that use algorithms.
And secondly, it is able to track reality at a very high frequency, such that when you look at your head-unit you are actually seeing what is happening now, as opposed to 3 seconds ago, or worse still some confusion of what is happening now with what happened 3 seconds ago. Time and again, InfoCrank shows itself as simply the most accurate and repeatable power meter available today.
Indoor cycling now incorporates a fully-fledged competitive domain and has somewhat inevitably followed the well-trodden human path of play becoming game becoming sport. The fundamental nature of sport is debated, but it is in part the process of formalising play just enough to preserve both a sense of fairness and the necessary unpredictability of outcome. This process benefits participants and spectators alike, since if the outcome of sport is either predictable or grounded upon inequity then it ceases to be fun, which is the fundamental nature of play.
But to borrow some legal terminology, has this transformation from indoor cycling to virtual racing occurred mutatis mutandis? That is, has all necessary formalisation been considered and included on the road between play and sport?
Whilst high-profile virtual racing events take place in a singular location including both participants and spectators, by its very nature virtual racing usually occurs in a global network of self-funded pain-caves, interconnected by online platforms such as Zwift. This fortunate convergence of technology has enabled a new order to emerge out of the chaos, and in some ways has revolutionised what we think of as cycle-sport. But many participants and spectators are already sensing unease regarding the question of in-race power validation, which threatens to undermine the sport’s credibility.
As in any new market, there is a proliferation of brands vying for market share by creating the best products they can and which they think the participant needs to enjoy his or her pastime. Also as in any new market, there is much that is different between these products and therefore much room for an unintentional reduction in fairness or too much unpredictability to creep back in and erode trust in those at the top of the leaderboard.
When Pain Cave A represents 280W on-screen for an actual rider effort of 300W, and Pain Cave B represents 300W on-screen for an actual rider effort of 280W, we can see how Rider A becomes very quickly disadvantaged.
It is easy at this point to resort to scientific claims about this and that, but in the spirit of collaboration we can attempt to conceptualise the above challenge in terms we can all agree on. Watts is the unit of power and is the result of measuring objective quantities such as metres and kilograms and then accurately mapping these measurements against time in seconds – legs having both extension (size) and mass and moving in circles at certain speeds are fundamentally analysable within these terms.
There is therefore an objective and correct fact-of-the-matter about how much power a cyclist produces and we can in some sense call this The Truth. Whereas the cyclist experiences The Truth as effort, it is the sports scientist who represents, or more accurately Re-Presents The Truth to themselves mathematically and conceptually.
But if we can say that we’re all interested in The Truth, what then of the variation in numbers displayed by differing power meters? This is sometimes characterised as a difference in degree of accuracy. Yet Accuracy, unless carefully characterised, is not really something that admits of degrees – to say something is 2% accurate is actually in some sense conceptually incoherent. This is because the Accuracy of any stated value relies on the underlying measurements generating those values having two aspects, the measurements must be both:
True – The measured quantity is as close as possible to that quantity’s actual value, and;
Precise – The degree to which two or more repeated measurements show the same results each time.
It is possible for measurements to be True and not Precise, just as my stopped clock may tell the right time twice a day. It might happen to be six o’clock when I check but this is simply coincidence. Equally, it is possible for measurements to be Precise and not True, just as my clock always being set 15 minutes behind, unless I am aware of this I will always be late for my meetings.
Where the Infocrank excels is that its measurements are both True and Precise. To extend the analogy, it is in effect an operational clock that is set at the correct time. When I look at the clock I can know that what I am seeing is an accurate Re-Presentation of The Truth, that is the actual time.
For Accuracy to admit of degrees the way in which the measurements are being taken must in principle be able to converge to True, assuming sufficient sensitivity of the measuring device. Measuring some kind of causally associated yet distant change within the cycling drivetrain (such as deflection in the spider or the tension in the chain) and processing the measurements to generate the power numbers cannot be said to be Accurate, simply because it is not possible in theory for such a measurement to converge to True.
Only when measurements are taken directly, that is of the literal force applied by the cyclist on the pedal down-stroke measured as crank-arm deflection, can the measurement in theory converge to True, and therefore stand a chance of being Accurate. Everything else is some measurement or other run through an algorithm, and this is what you actually see on your graphs, but InfoCranks do not need to conceal the lack of True with an attempt to secure Precise. This is why we can talk in terms of percentage of error from The Truth.
For another view consider the words of Sports scientist and coach, James Spragg, in Cycling Weekly recently: “Before virtual racing came along, power meters needed to be relatively accurate to themselves, to use as a training gauge. As a training tool they work fine. But when you start comparing the values from a Stages crank to an SRM then you start to get issues. Because they were never designed to do that…. Between two SRMs you could have an eight per cent swing. Between two Stages power meters you could have a 12 per cent swing,” he says, referring to data from the 2017 study. This expresses the challenge faced by virtual racing in an alternative way.
Spragg makes a similar point to Verve, but as we have seen already Accuracy has two components, True and Precise. We can therefore see the common misconstruction mentioned earlier expressed in Spragg’s words “power meters needed to be relatively accurate to themselves”, which means on the most charitable reading Precise and which we now understand is only one of the aspects necessary for Accuracy to even be possible. This is why many claims about 2% Accuracy are nonsensical, because they are not being made in relation to The Truth. Most would now concede that “Accurate to themselves” is synonymous with ‘uniformly and consistently wrong’.
What is needed then is a benchmark, and in virtual racing the benchmark has to be accurate power numbers, or the objective and correct fact-of-the-matter, or simply the Truth. Top eSports team Canyon ZCC chose Infocrank precisely because Verve are in the business of truth, and we are playing our part in preserving the sense of fairness and unpredictability of this new discipline, but we understand that more is needed.
This isn’t a difficult problem to solve, in fact it already has been, but what is required first is helping others to understand what Accuracy really is, not what they are told it is. At Verve we firmly believe that the truth is always out there and rarely any more than one step away if you know where to look. So, back to the original question, has indoor cycling’s transformation to virtual racing occurred mutatis mutandis? Quite simply, no, but it is happening more quickly with the help of Infocrank.
Using a power meter enables you to carefully monitor and structure your training sessions in order to gain as much benefit as possible from any given session. When increases in power can be measured and accurately tracked over time, it becomes possible to better understand your body and how it responds to different types of workouts.
Power meter technology is what has enabled cyclists of the modern era to become increasingly specialized and accentuate the precise physical characteristics necessary to dominate in any given discipline, be it road, track, mountain bike or BMX.
Very broadly speaking, cycling events can be split into two categories – power-based events such as the Track Team Sprint or BMX, and endurance-based events such as the Track Team Pursuit or Tour De France. This is not an entirely clean characterisation as a degree of endurance is required in power-based events as well as vice-versa, but where one or the other predominates we can reliably say it falls within its respective category.
Mark Cavendish may be said to be a hugely successful sprinter, but he is by no means a power based athlete in the conventional sense. Surviving a three-week Grand Tour takes exceptional levels of endurance, but we can also say that alongside all other Grand Tour riders, Mark is an exceptional sprinter. After three or four hours of endurance-based racing, he possesses the relative power advantage to cross the line first.
Power output is the best indicator of how a cyclist may perform when they are attempting to maximise their speed, yet absolute speed within cycling is a truly multi-faceted and complicated beast since it is entirely dependent on context. Aerodynamic drag, body position, weather conditions, discipline, duration of effort and terrain are just a handful of factors that any cyclist will wish to understand before he or she decides if a given speed is worthy of respect.
If power output matters, what is actually being measured when measuring a cyclist’s power? In simple terms, the power of a cyclist is contingent upon how much force they can push their pedals round as well as the speed at which they are doing so – maximum power is obtained when the force on the pedals multiplied by the crank’s cadence (RPM) gives the highest figure. We can say:
Pedalling Power = Force On Pedals x Speed Of Pedals = 200N x 3m/s = 600W
N = Newtons
m/s = Metres Per Second
W = Watts.
Any increase in either cadence or force will therefore result in an increase in power. To be able to visualise how much power is being generated, simply divide the Watts figure by 10 and convert it to Kilograms – such a weight will be lifted by 1m in 1 second. So in our example, a cyclist producing 600W is producing enough power to lift 60kg by 1m in 1 second. That’s the same thing as lifting an average sized human female 1m vertically in a second! But what does all this tell us about the power output of the different types of professional cyclist?
Unsurprisingly true power-based cyclists such as track sprinters and BMXers generate the most power. In fact, on the GB Cycling Team the most powerful athlete in 2012 was the BMX rider Liam Phillips. Liam was able to generate more power than Sir Chris Hoy due to his super-fast and explosive cadence. Amazingly elite level BMX riders can generate in excess of 2200W during their 1.5 second run down the 8 metre start hill. They accelerate from 0 – 60km/h in this time, which is as fast as pretty much any supercar available today, and lift the equivalent of 200kg a vertical height of over 1 metre.
To put this in context, that is three average sized fully grown men elevated 3 feet! If you consider this alongside BMX athletes’ ability to hit 15m long jumps at 50 km/h and land them within 10cm every time, you begin to see the true talent of these athletes. They truly are exceptional. They can’t of course maintain this power output for sustained periods, but that is not the point in BMX. They fight first to the bottom of the start-hill, and then to the end of the first straight. Whoever enters that first corner first becomes the man or woman to beat.
At the other end of the endurance and power continuum, you find Grand Tour riders who look to be able to sustain the highest power possible for a three-week duration, as opposed to bursts of 1.5 seconds. Most top Grand Tour riders will produce a fifth of the power of Liam, but they can hold that power for maybe an hour or more on a given climb or time trial.
Bradley Wiggins produced 440W on average for his entire hour record, which is a staggering feat, and for endurance events such as this the mental toughness required to maintain such an output makes the feat even more remarkable. For most mere mortals, holding a power number between 150W and 200W for an hour would be a respectable achievement.
So, we can see then that although power output is an excellent indicator of how fast a cyclist is able to propel themselves, we can also see that power output alone tells us very little, we need to understand the context too.
Most physically fit people could generate 440W, but none of them could do so for an hour. Different professional cyclists produce varying amounts of power within different disciplines to meet the physiological demands of their specific discipline.
But unless you wish to become an elite level BMX athlete, at which point you will need to develop truly exceptional bike-handling skills, it’s sustaining the power numbers that really makes the difference to your performance. This is why it’s always crucial to decide what your goal is ahead of time and train scientifically toward that outcome.
The Verve InfoCrank is the most accurate and repeatable power meter available today, and is used by most of the world’s dominant track cycling nations, as well as triathletes, rowers and eSports cyclists.
Strictly Necessary Cookies
Strictly Necessary Cookie should be enabled at all times so that we can save your preferences for cookie settings.
If you disable this cookie, we will not be able to save your preferences. This means that every time you visit this website you will need to enable or disable cookies again.