A quantification of the reduction in effort when running in lighter shoes

Are lightweight running shoes worth the money?

Yesterday, I attempted a little experiment to see if lighter trainers really could really make running easier. Was the £51.99 I spent at SportsShoe on my lightweight Asics Gel Hyperspeed 5 really worth it?

For those of you who don't want to read the detail, my preliminary answer is: Yes! The Hyperspeeds gave me a about a 1.8% reduction in effort and therefore probably also a similar improvement in running time compared to my previous running shoes.

I have previously run all of my marathons in a pair of Saucony ProGrid Ride 3 (£39.99 from SportsShoe) and I have not had any problems with them. But, they are reasonably heavy (UK 9 ~350g) and after missing my 'good for age' at the London Marathon (3 mins on the loo just over halfway resulting in a disappointing 3:15:42 finish) I decided I wanted to maximize my chance of a fast time 3 weeks later in Prague. So, at the last minute and with no time for any testing I bought a pair of Hyperspeeds. They are much lighter than the ProGrids and my first impressions were that they felt good (3 km test jog a few days before the Prague Marathon). At that point I weighed-in at about 62 kg having put on about 1.5 kg since the London Marathon and I figured that being towards the lower end of the weight spectrum I was unlikely to suffer too much from the Hyperspeeds reduced cushioning. At Prague the race went well and I finished feeling great in 3:07:52 - a fair improvement on London. The Hyperspeeds were definitely light and I had the impression that they might have been solely responsible for the speed improvement. Now, 2 weeks on and well into my recovery (two days ago I just got a PB on my 5 km of 19:07 in my trusty ProGrids) I thought I should compare the two shoes directly with one another.

Method

The test I designed was fairly simple and based on the principle that within a session heart rate reports physiological effort. As you increase your muscular effort, your heart has to pump more blood. If all you are doing is the same type of exercise (i.e. running) with no massive change in posture, muscle groups, hydration etc then increased effort equals increased heart rate. Taking speed into account is very important since increasing speed increases the effort and therefore heart rate. I used two techniques for making sure that speed could not confound my results. First I ran at a flat pace, the same for each run (within a couple of seconds) and then I also normalized the data for speed by calculating the number of heart beats per km (more about that form of analysis later). To make sure that wind speed and hills and road surface did not also confound the results I ran on a flat tarmac path running along the course of an old railway track (now the Cambridge Guided Busway) using an out-and-back course. I made sure that intervals between the tests were similar and I started and stopped my watch at fixed points (a white line on the path) whilst running at a constant pace. The route was 1 km out and 1 km back with the turn around at a 'Cycle Dismount' post where I swung 180 degrees around it using my right arm (you have to make these things fun!). I stuck to the same running-line and I did not drink or take any loo stops between runs. I did five runs in total. The first was a warm-up run in the ProGrids and it also served as a test of the pace that I felt comfortable with (I had done a hard day gardening, I had raced the previous day and it was warm...). I then did each subsequent run alternating between the shoes. I had intended to do six runs, but a loo stop called a halt to the experiment.

I wore my Garmin 305 Forerunner watch and heart rate monitor strap with the autopause and autolap function switched off. I left the recording mode set on the automatic setting. I used a display showing average heart rate, average pace, instantaneous heart rate and instantaneous pace in order to get each run at roughly the same pace.

Results

Data from each of the five test runs is shown below (Table 1). The Time was that recorded by the Garmin, the distance was always 2 km (not the noisy satellite data from the Garmin). Since the Garmin Forerunner only reports average heart rate to integer accuracy, which is not good enough for this study, I have had to use a slightly more complicated technique to extract the heart rate data. I have used a form of signal processing that relies on averaging multiple values in which noise dithers values around a mean (I am guessing most people don't realize that noise is often very useful when manipulating data which is reported with low resolution). I summed the product of the integer heart rates, for each of the Garmin's ~1s sample periods, and the length of the sample period. This produces a mean value for heart rate that has more than integer accuracy.

Num	Shoe	Time	Speed	Heart rate	'Physiological effort'
		mm:ss	km per hour	beats per min	beats per km
1	ProGrid	09:42	12.37
2	Hyperspeed	09:38	12.46	125.4	604
3	ProGrid	09:40	12.41	127.7	617
4	Hyperspeed	09:39	12.44	125.8	607
5	ProGrid	09:39	12.44	127.9	617

Table 1. Data for the 5 sequential 2 km runs with alternating shoes. The first run was a warm-up from resting and, since the first run always shows a lower heart rate, was excluded from subsequent analysis. 'Physiological effort' is the term used to describe the number of heart beats taken per km covered. For this data sequence no other correction was made to this metric - although my previous work has shown that it typically rises during longer runs. Click on the shoe name to take you to the Garmin Connect data.

From Table 1 shows that all of the test runs were completed to a 1 s accuracy. The average heart rate for the ProGrid was 127.8 bpm compared with 125.6 for the Hyperspeed. Correcting for the small differences in running time I have calculated that the average heart beats per km for the ProGrid was 617 beats per km compared with 606 beats per km for the Hyperspeed. This equates to  ~1.9% reduction in physiological effort which, if sustained over a marathon distance, might be expected to result in a similar percentage time improvement (~3min 30s for a 3hour 10min runner). Equally, such a reduction in effort may be expected to drop my PB at 5 km from its current 19:07 to 18:46. Obviously only two tests with each shoe is insufficient for testing of statistical significance, so a note of caution is required (and I would not dream of publishing this in a scientific journal as it stands!).

Figure 1. Bar chart of the heart beats per km from the four test runs.  The y-axis is scaled so that the lowest value (510 beats per minute) sits approximately where I would need to be if I wanted to break the current marathon world record. Thus, you can judge to what extent the shoes might contribute towards an attempt by an only just 'good for age' runner on the world record!

For those would prefer figures (Figure 1) shows the heart beats per km for the four test runs. The effect certainly looks consistent. However, more data is still required. I will do another repeat of this test, but I want to try a faster speed. A simplistic physical analysis suggests that the effect of weight should become progressively greater as speed increases since the kinetic energy is proportional to velocity squared. Since the shoe has to accelerate from rest to roughly twice your bodies travelling speed with each stride, the energy input must rise dramatically with speed. To see the results of my next test, done at a higher speed and using a constant effort click here.

Health Warning

The most obvious question is: "Would this work for me?". Well, if you are running a marathon there are lots of things that can stop you getting a PB and lightweight shoes might not help. The most obvious mistakes are poor pacing (flat is best), unrealistic expectations from current training, not taking weather into account, over-hydration and injuries. But, if you are light enough to tolerate minimal shoes and fast enough to benefit from them with a decent running form then the answer is probably yes! There is a well established literature on the topic of shoe weight and energy consumption using treadmills - I have reviewed some of it here. But, there are probably cheaper ways of getting faster - but, they usually involve more training, less eating and generally reduced fun!