Running Efficiency

Most endurance athletes focus their training on attempting to increase their aerobic capacity (VO2max) and their endurance. Training increases aerobic capacity by increasing the ability to deliver oxygen to muscle fibres and by increasing the capacity of mitochondrial enzymes to generate energy by oxidation of fuel. However our maximum capacity to generate energy by oxidation of fuel appears to be limited by our genes and/or early development.

The traditional approach to improving endurance is the long run. This increases resilience of muscles, tendons and ligaments, and enhances the ability to metabolise fats.   Once we have trained our aerobic capacity to our limit, and we have developed sufficient endurance to sustain us for the duration of our target event, what scope is there for further improvement in performance? The remaining option is increasing efficiency: that is the effectiveness with which we can use energy to produce speed.


Figure 1 shows speed (in metres/min plotted) against rate of energy production (VO2 measured in ml/min/Kg) for three hypothetical athletes. The slope of each line represents the efficiency of each athlete.   The line with medium slope represents the average runner in the sample used by Jack Daniels to derive the VDOT charts in his book ‘Daniels Running Formula’. The steeper line represents an athlete 10% more efficient than average. The less steep line represents an athlete 10% less efficient than average. Greater efficiency indicates a greater speed for a given rate of energy production. Note that the lines are almost straight, indicating that efficiency is nearly constant across a wide range of paces, though there is a minor degree of flattening of each line at higher paces, indicating a somewhat lesser increase in pace for each additional unit of oxygen consumed.

Figure 1: The relationship between pace and aerobic energy production.  These lines are derived from the data used by Jack Daniels to derive his VDOT tables. The middle line (brown) is the data for an athlete who had the average efficiency from the sample studied by Daniels. The upper (blue) line represents an athlete who is 10% more efficient than average.  The lower represents an athlete who is 10% less efficient than average.

Figure 1: The relationship between pace and aerobic energy production. These lines are derived from the data used by Jack Daniels to derive his VDOT tables. The middle line (brown) is the data for an athlete who had the average efficiency from the sample studied by Daniels. The upper (blue) line represents an athlete who is 10% more efficient than average. The lower represents an athlete who is 10% less efficient than average.

If each of the three hypothetical athletes had a VO2mx of 72 ml/min/Kg (typical of an elite distance runner) the athlete with average efficiency would achieve a pace of 350 metres/min at VO2 max while the athlete who was 10% more efficient would achieve a pace of 385 m//min. At 80% of VO2 max (57.5 ml/min/Kg), the athlete with average efficiency would be expected to achieve a pace of 293 m/min while the athlete who was 10% more efficient would achieve 322 m/min. It should be noted that for an athlete with a lower VO2max, the pace at VO2 max and at any given percentage of VO2max will be less, but the relative gain in pace from an increase in efficiency will be similar. In other words, for two athletes with the same VO2max, a 10% improvement in efficiency would result in 10% faster times in races run at any given proportion of VO2max.

Might training produce an increase in efficiency of 10% or more?   The measurements of Paula Radcliffe performed by Andrew Jones for more than a decade provide clear evidence that the answer is yes. In fact, the data shows that Paula achieved a 15% increase in efficiency over the decade from 1993 to 2003*. Her VO2max remained virtually constant at around 70 ml/min/Kg over this period. Thus, a major factor in Paula’s phenomenal marathon record of 2:15:25 recorded in 2003 appears to be the remarkable improvement in efficiency. How might an athlete improve efficiency?  There are two possibilities: increasing biomechanical efficiency and increasing metabolic efficiency.

Biomechanical efficiency

There are three major energy costs of running:

  • overcoming the braking that occurs while on stance;
  • getting airborne;
  • swinging the leg forwards after lift-off from stance.

There is also the cost of unnecessary tension or movements of other body parts, but as is well illustrated by Emil Zatopek and Paula Radcliffe, who both achieved phenomenal performance despite unnecessary upper body movements, the cost of such movements is relatively small, and there is unlikely to be more than a slight gain from reducing them.

Minimizing the sum of the three major costs requires a balance between conflicting effects. At a given cadence (steps/min) braking cost increases as the cost of getting airborne decreases because less time in the air inevitably results in a larger proportion of time on the ground. Although braking only occurs when the point of support is in front of the centre of gravity, braking cost cannot be reduced merely by attempting to land with the foot under the body, because at constant speed the forward-directed impulse generated after mid-stance and the backward directed impulse generated by braking before mid-stance must be equal (after allowing for overcoming wind resistance).  For a give cadence, the cost of braking can only be reduced by spending more time airborne.

During a marathon, many runners spend an increasing proportion of the time on stance as the race progresses. This is likely to result in greater braking and reduced efficiency. It is noteworthy that the well-known picture of Paula Radcliffe at mile 14 on her way to victory in the 2007 New York marathon shows her getting well-airborne. This demonstrates that she had adequately developed reserves of the leg muscle power required to get airborne. Andrew Jones’ measurements demonstrated that her vertical jump performance increased from 29cm in 1996 to 38cm in 2003.

Paula Radcliffe airborne at mile 14 in the New York marathon, 2007.  Photo by Ed Costello, Brooklyn, NY,US

Paula Radcliffe airborne at mile 14 in the New York marathon, 2007. Photo by Ed Costello, Brooklyn, NY,US

The cost of getting airborne can be reduced by increasing cadence, because the body falls a lesser distance during a series so short hops than during a longer hops covering the same distance, simply because a freely falling body accelerates, thereby gaining greater speed the longer it is airborne. The optimum cadence increases with increasing speed, because if cadence does not increase with increasing speed, stride length would necessarily have to increase disproportionately, resulting in heavy costs of getting airborne and also braking. However, there is a limit to the gains that can be achieved by increasing cadence, because the cost of moving the swing leg forwards increases in proportion to cadence (as shown on the my calculations page).

Nonetheless, many recreational athletes have scope for increasing efficiency by increasing cadence. The study by Heiderscheit and colleagues indicates that a typical recreational runner might improve efficiency by decreasing both airborne costs and braking costs by increasing the self-selected cadence by up to 10% . This increase in cadence also reduces stress at the joints by virtue of the reduction in forces required to get airborne and overcome braking.   Heiderscheit reported that a 10% increase in step rate from a self-selected mean step rate of 172.6 ± 8.8 steps/min at a pace of 2.9 ± 0.5 m/s led to an almost 20% reduction in energy absorbed at hip, knee and ankle joints.

It is probable that Paul Radcliffe achieved optimum balance between the cost of getting airborne, braking and advancing the swing leg largely by virtue of fairly intense running, together with hopping drills and weight lifting.   While training near to race pace might optimise neuromuscular coordination, I suspect that the major requirement for optimising mechanical efficiency is adequate muscle power. Although I do not have direct evidence to prove it, I think it is plausible that a small amount of high intensity training will achieve as much gain in mechanical efficiency with less total wear and tear on the body compared with a larger volume of threshold training, simply because training near to maximal effort is more effective for improving muscle strength and power.

Metabolic efficiency

Metabolic efficiency of oxygen consumption is a measure of the amount of mechanical work (and hence speed) that can be achieved from the consumption of a given amount of oxygen. Several factors influence this. The most important is the fact that the efficiency of conversion of metabolic energy to mechanical energy during contraction of a muscle fibre is greatest when the speed of contraction is near the middle of the range of contraction speed that can be achieved by that fibre. When a fibre contracts too slowly it consumes energy developing tension that does little work. Fast twitch fibres have an optimum speed of contraction that several times faster than that of slow twitch fibres. But the speed of fibre shortening during distance running (and also cycling) is better matched to the optimum contraction speed of slow twitch fibres.

It is noteworthy that many of the large muscles that act at hip and knee cross both joints, flexing one while extending the other or vice versa. However during running hip and knee flex simultaneously or extend simultaneously. Consequently, the rate of change in length in these muscle during running is small. Thus type 1 fibres which are well suited to the isometric contractions required to maintain upright posture are also well suited to distance running during which contraction rate is slow.

In the case of cycling, there is direct evidence that efficiency of metabolic to mechanical conversion is greater in individuals who have a higher proportion of type 1 fibres. Although I do not know of any similar measurements in runners, it is very likely that runners with more highly developed type 1 fibres will be more efficient.

The most effective way to develop type 1 fibres is likely to be consistent high-volume training over a sustained period. It is likely that a major part of Paula Radcliffe’s improvement in efficiency was consistent training, with a gradual increase in training volume over a period of a decade. As discussed in my previous post, Paula did a lot of her training at a moderate or high intensity. It remains a matter of speculation as to whether she could have achieved similar phenomenal marathon performances with less damage to her body by a more polarised approach, in which a modest amount of high intensity running was accompanied by a larger proportion of low intensity running. Perhaps she could have achieved similar improvement in metabolic efficiency with a larger proportion of low intensity training over a longer period of time. My own view, based on Skoluda’s evidence that many distance runners have evidence of sustained high levels of the potentially harmful catabolic hormone, cortisol, is that for many athletes a polarised approach offers the best prospect of gradual improvement in metabolic efficiency and hence, the prospect of year-on-year improvement over many years.


Paula Radcliffe’s spectacular 15% increase in efficiency over a period of about a decade, despite an approximately constant VO2max, provides compelling evidence that a worthwhile enhancement of efficiency is possible. It is likely that a combination of high intensity training, hopping drills and weight lifting honed her biomechanical efficiency.  For many recreational athletes, biomechanical efficiency might also be improved by increasing their self-selected cadence by as much as 10%.  It should be noted that optimum cadence increases with speed.

It is also likely that a gradual improvement in metabolic efficiency over a period of more than a decade was also a major contributor to Paula’s improved efficiency.  It is likely that she achieved this by consistent training, with a gradual increase in volume over the years.  Whether or not she might have achieved a similar enhancement of efficiency with a less damaging, more polarised approach to training remains a matter for speculation. Nonetheless, in my opinion, for many athletes, a polarised approach is likely to offer the best prospect of gradual improvement in metabolic efficiency over a period of many years.


*A minor point to note is that Andrew Jones estimated speed at VO2max by assuming a linear increase in pace with increased VO2. This is likely to produce a small over-estimate of actual pace at VO2max, because in reality the curve flattens a little at high values of VO2. Nonetheless, provided all the measurements are made at a similar region of the curve, the error in estimate will be consistent across different measurements. It is pace at around 80% of VO2max that matters most to a distance runner.


4 Responses to “Running Efficiency”

  1. Pete Says:

    Hi Canute!

    Another biomechanical parameter certainly affecting the running efficiency is the runner’s body weight. This could actually be the easiest “trainable” feature, at least as long as one’s body fat percentage is > 10%. One might argue that experienced runners already have achieved their individual minimum fat percentage and therefore cannot improve their running efficiency by loosing weight. This leads to the question whether further weight loss could be attempted even by sacrificing some muscle mass (and possibly leg power). Then again, do experienced runners carry any superfluous muscle mass in their bodies?

    Could the loss of leg power by loosing muscle mass at any circumstances be compensated for by the gain in the running efficiency due to the reduced body weight?

    • canute1 Says:

      Thanks for your comment. I agree that weight loss is potentially beneficial. If absolute VO2max (ml/min for whole body) remains constant, weight loss will lead to increase in VO2max expressed in ml/min/Kg, and it appears that the demand on the body is proportional to VO2 relative to VO2max in ml/min/Kg. However, whether one describes the change as an increase in VO2max or a change in efficiency is only matter of terminology.

      Both the cost of getting airborne and the cost of braking decrease in direct proportion to decrease in weight, so the energy needed to achieve a given pace will be less. The cost of moving the swing leg forward will not change much if the weight loss is due to loss of adipose tissue from the torso, but will decrease if there is loss from leg muscles.

      A modest amount of fat is crucial as a building block for various hormones and for cell membranes; for maintaining levels of fat soluble vitamins and for various useful mechanical functions. The minimum is a subject of debate, as it depends on multiple factors but some athletes appear to be healthy at 5% body fat

      With regard to loss of muscle, the hormonal system puts a very high premium on maintaining glucose supply to the brain and therefore during exercise will consume either fat or protein to conserve glycogen (precursor of glucose). If there is no spare fat, muscle protein will be consumed. A substantial dietary protein intake will minimise loss of muscle, but excessive protein intake creates its own metabolic problems.

      My own view is that if you start with an average musculature appreciable loss of leg muscle is likely to do more harm than good. Perhaps it is advantageous for a distance runner to have less than average arm musculature. I aim for a BMI of 20, with skinny arms but good musculature around the thighs and butt. I think perhaps I could go down to 18 without significant loss of leg strength, but I think that if one goes that low it is important to eat a well-balanced diet including adequate micronutrients such as fat soluble vitamins.

  2. Ewen Says:

    Great post Canute. Paula’s height off the ground in that photo is incredible. Thanks again for a thorough explanation of the variables in fast distance running.

    • canute1 Says:

      Thanks for your comment. Yes Paula’s powerful elevation is amazing, and quite similar to that of Dennis Kimetto, as shown in the slow-motion clip of final few Km of his WR performance in Berlin in October this year.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: