How to pedal efficiently: “dead centre size”


As of September 2017, new Sweat Science columns are being published at Check out my bestselling new book on the science of endurance, ENDURE: Mind, Body, and the Curiously Elastic Limits of Human Performance, published in February 2018 with a foreword by Malcolm Gladwell.

- Alex Hutchinson (@sweatscience)


[UPDATE 9/17: Check out reader Phil Koop’s analysis of the paper in the Comments section of this post. Definitely worth a click.]

An interesting Norwegian study on pedalling efficiency has been posted online at Medicine & Science in Sports & Exercise, proposing a new measurement to determine whether cyclists are getting the most out of their pedal strokes.

The basic goal of cycling, obviously, is to convert your effort into forward motion of the bike. To do that, experts have long believed that a quantity called “force effectiveness ratio” (FE) should be optimized. FE basically tells you how much of the force you’re applying with your foot at any given moment is directed perpendicular to the crank. For example, when the crank is parallel to the ground and you’re pushing straight down, FE is very high. But when the pedal is at the very bottom of the cycle, if you’re still pushing down instead of back, your FE will be lower. Averaged over the whole pedal stroke, a typical FE might be about 50%.

This makes sense in theory, but no one has been able to show that better cyclists have higher FE (at least if you’re comparing elite and sub-elite, as opposed to complete novices):

From an energetic-mechanical point of view, this should be the best “technique” parameter because the energy used for producing the ineffective, static component of force will not contribute to external power. However, it is not necessarily so that man is able to produce the highest FE at the lowest metabolic cost: the coordinative challenge of generating power while creating a rotation of the crank by extending the lower extremity may require additional, apparently ineffective, energy expenditure.

The Norwegian researchers propose instead a parameter called “dead centre size” (DC). They look at the two weakest parts of the pedal stroke — the top and bottom — and compare how much useful force you’re generating to the average useful force over the whole pedal stroke. A typical value is about 25%. The idea is that the better these “low points” in the stroke are, the smoother your overall stroke must be, so you’ll avoid wasting energy accelerating and decelerating and so on.

So they did a study with 21 competitive cyclists, using force-sensitive pedals and high-speed video and so on, and sure enough found that DC was much better at predicting the overall efficiency of the cyclists than FE. So what does this mean? Well, it’s consistent with the idea that you shouldn’t worry too much about trying to generate power on the upstroke, since that’s a hopeless task. Instead, focus on keeping the whole cycle smooth, not letting power dip too far at the top and bottom. One thing I couldn’t tell was whether it’s possible for anyone without a high-tech laboratory to get a measurement of DC, in order to see whether they’re improving their form over time. It would be pretty cool if local bike shops were able to offer the service.

3 Replies to “How to pedal efficiently: “dead centre size””

  1. I can’t read the paper, lacking a subscription, and have some dumb questions. Efficiency would be measured in units oxygen consumed per watt? And this would be measured with a face mask & apparatus on the input side and an ordinary power meter on the output? That would imply an indoor setup on rollers or a trainer. One wonders how well observations on pedaling style made under such circumstances will apply to the road. And out of curiosity, how big was the spread of efficiencies?

  2. I’ve now had a chance to read the paper (thanks, Alex!) and can answer my own questions.

    1. Efficiency is indeed measured exactly as you would expect. (An aside: the authors are contributing to a substantial body of literature, to which they refer, and naturally their intended audience would be familiar with basic parameters like this.)

    2. The cycling was done on road bikes fixed to trainers. This seems the best practical method, but it is nonetheless rather dissimilar to actual cycling. One reason is that the bike doesn’t move freely side to side, as during actual riding. The other is that the pedal resistance has an unnatural viscous feel, as though one were pedaling while braking. This is a potential concern in a study of pedaling technique. Finally, each participant was studied at a single cadence, namely that which was freely chosen at the mandated power level. Cadence is significant because it interacts with pedaling style and efficiency. Although freely chosen, these cadences are not exactly comparable because the available gearing provides possible cadences that are much more coarsely discretized than the measurement precision. The reader should be aware that chosen cadences very widely depending on factors beyond work rate.

    3. The researchers say that average efficiency was 21.7% +- 1.2%. They do not specify whether these bounds are absolute or sample standard deviations, but either way they are significant: 1.2 is about 5.5% of 21.7.

    Cycling coaches do not discriminate very clearly between the two measures, FE and DC, that the paper studies. Coaches often talk in terms of effectiveness (force perpendicular to crank) but their actual advice is to push through the top of the pedal stroke and pull through the bottom; this advice corresponds more closely to DC than FE. The study can therefore be viewed as a scientific investigation of existing practice.

    Anecdotally, I find that the pushing part of this advice is more useful than the pulling, and that it matters most when the cadence is too low. One sometimes finds that when working hard at a cadence that is slightly too low, a burning sensation in the leg muscles signals that one is losing the battle to clear fatigue poisons faster than they are generated. One can sometimes reverse this by working through more of the pedal stroke. This usually results in a small increase in cadence (and therefore power) at the same time that comfort increases! One speaks of “getting on top of the gear.” I find that pushing is more important in this process than pulling. Another low cadence situation is when I am riding a single-speed bike, and thus have no choice of cadence. When the cadence is low, I can be working well below my comfortable rate but find that uncomfortably large pedal forces are required to lift the cadence. In this case, concentrating on working through more of the pedal stroke can sometimes raise cadence and power without requiring peak pedal forces to be increased.

    The study is correlational; it merely demonstrates that of the 21 riders studied, those that measured high in efficiency under the test conditions tended to score high in the DC statistic. Naturally, the burning question for any rider or coach would be: can I improve my DC statistic through coaching and training? And if I improve it, will my efficiency improve?

    These questions can in principle be answered scientifically. One would require both riders and coaches in the study to establish a double-blind. Results could be measured by comparing improvements (if any) in time trial performance between DC-coached riders and placebo-coached riders. (A flat TT is the riding condition most similar to the test conditions.) Such a study would be expensive, but it is possible it has already been done. Much cycling research is conducted by bodies such as national programs that have a material and proprietary interest. A positive result would make the methodological issues I mentioned earlier moot.

  3. When walking or running each leg is only effectively used for half of the time, the same should apply to pedaling when one is searching for the perfect flat TT technique, in this way each leg can be given total concentration when generating and applying torque. In reality there is only one dead spot sector, it’s at TDC (11-1 o’c)and if this can be eliminated by replacing it with close to maximal torque, you will have 180 deg. of highly effective crank torque between 11 and 5 o’c and a possible FE of 60 to 70+ % over that half circle, which with both legs can give 60 to 70+ % over the entire circle. Is this possible, yes it is and while using standard equipment.

Comments are closed.