Actionable Intelligence for Running Part 7 : Running Power Characteristics During a Duathlon

In Part 6 of this series, I inspected data from a VO2 lab test data and graphed it's relationship to my corresponding power to weight ratio for 6 different running speeds. What I discovered what the non-linear nature of VO2 (rising and settling dynamics) and the inability of a linear equation to predict instantaneous oxygen cost from a power to weight ratio. 

In general, what was encouraging to see the was the proportional rise in both VO2 and power to weight ratio as speed increased and this credits the Stryd footpod as a steady state "running cost" predictor even though it is an accelerometer / bouce meter, i.e it does not directly measure mechanical power but algorithmically outputs power based on components of velocity extracted from acceleration data.   

I also ended the post by stating that the Stryd does not account for outdoor wind resistance nor for the effect of temperature and humidity, so using an indoor treadmill based correlational equation is likely to underpredict the true cost of running outdoors, especially against winds pushing past Beufort scale #6.

Readers familiar with my previous posts on the GIANT Duathlon Series will know that this race frequently brings some top athletes from the region to the start line. The race format is 3K run, 25K bike, 3K run. This is my third season as a duathlete.

In this post, I'd like to inspect running power during the two running splits of Race#4 held on March 10, 2017. 

Equipment and Personal Data

Running Shoes : Mizuno Wave Ronin 2 (Pre-2010, yes I hold onto old stuff)

Shoes Weight (pair) : 7.5 oz.

Heel to Toe Drop : 9mm
Footpod : Stryd 

Body Weight (unclad) : 63.5 kg
Training (conditioning) : 8-10 hours weekly

Fig 1 : Mizuno Wave Ronin 2

The Course 

ADULTS
Run - 3.0km (https://ridewithgps.com/routes/19399829) - 2 laps
Bike - 26.9km (https://ridewithgps.com/routes/19399694)
Run - 3.0km (https://ridewithgps.com/routes/19399829) - 2 laps

The Data

Elsewhere, I will describe in a short race report the feelings and effort going through this race. In a nutshell, it's been one of my best performances to date, having placed 8th in my age category. However, it keeps getting more difficult race by race to move up and 1:17:36 is nothing to boast about.

Fig 2 : GIANT Duathlon 2017 Race 4 results (30-39 Age Category)

The Stryd and Stages powermeter will not automatically pair to the Polar V800 as sport mode changes in a duathlon. Further, the V800 does not have a power feature within running. Therefore, I relied on offline data saved on the footpod for post-processing.

Below is a composite plot showing running power and biomechanical characteristics of the race. Note that cycling has been ignored except to trace an average cycling power for that duration.

Fig 3 : Composite plot showing running power, form power, ground contact time, vertical oscillation and leg spring stiffness from the two running splits of the Giant duathlon Race#4 on March 10, 2017.

Focusing in on the two run splits, the performance variables are tabulated below :

Fig 4 : Performance tabulation of key power and biomechanical variables during the two running splits of GIANT Duathlon Race# 4.

Please note that speed was calculated from the distance vs time relationship from the clocked results of the race and not from the footpod.

Insights

All highlighted items - speed, avg. power, power to weight ratio, form power to avg power ratio, cadence and LSS suffered in the second run leg. Considering these facts, the effect of fatigue during short high intensity sprint duathlon is clear to see.

The delta between these variables (those in Run 2 minus those in Run 1) expressed in percentage are as follows :

Fig 5 : Calculated percentage differences in running performance variables in Run 2 compared to Run 1.

1. A 10.5% decrease in run power resulted in a 10.7% decrease in pace in run # 2. Knowing the proportional relationship between VO2 and running power from Part 6, I conclude that the internal running engine ran a bit out of steam. 

Fatigue is multifactorial, not just cardiovascular. There was a short duration decrease in power about 1/4th of the way into run # 2 which in reality coincided with a slight tightening of the right leg muscle where I had to throttle down power to 180W for a few seconds.

However, there were no cramps and no stopping to loosen the legs. The salt instake for the day was good considering the ingestion of both a GU gel worth of 180mg of sodium and an aerodynamic water bottle filled with water + 380mg serving of sodium in 50g of electrolyte mix. 

Some others who have seen my data comment that this is an example of predominantly "metabolic fatigue".

2. Form power, i.e the cost of "perpendicular bouncing" as a percentage of total external running power, was 8% higher in run # 2 than run # 1 (external running power does not account for the swing in upper and lower body limbs).

3. Ground contact time (GCT) was 9.76% greater in run # 2.

4. Leg spring stiffness (LSS), extensively discussed in Part 1, Part 2 and Part 3 of this series, was 2.9% higher in run # 2. Cadence decreased and GCT increased between the two runs, but as previously discovered by experiment, the effect of change of GCT on LSS is greater than is the effect of cadence on LSS (Part 2). 

5. Run effectiveness (m/s over W/kg), a surrogate for running economy, decreased by a tiny fraction of a % in run # 2 compared to run # 1. 

Within each of the splits, the behavior of the RE trend (fraction of instantaneous RE over average RE) was a mildy increasing one for run # 1 and flat for run # 2. Please note that the calculated value of RE is sensitive to the data and abnormal spikes in speed or drop in power will result in higher than usual RE values.

My conclusions from the above data study for sprint duathlon are as follows :

In an ideal scenario :

a) Leg turnover would be similar in the two run legs. Longer swing times during the start of the second run affect GCT which consequently has bigger impact upon LSS than the lowering of step rate alone.

b) Vertical "bouncing" as a fraction of total power would be reduced in the second run so that more of the horizontal component makes up the total power. However, I  must confess my understanding on components of power is not on par with the coaching community. I believe one has to bounce to an extent to generate the potential energy needed to activate the storage potential of the leg spring, so an optimum must be struck between too little bouncing and too much bouncing. Too much bouncing is understandable since energy is being used to elevate and lower the center of mass and perhaps some of that could be acively focused instead on moving forward. Form power is something to continue experimenting with.

c) Nutrition points would be more optimally placed during the race to allow proper absorption by body before the demands of the second run. Ideally, this would allow a more even power to weight ratio between both the run splits. Race strategy is knowing exactly at what points to ingest and that can have a pronounced effect on performance during the second run split.

d) Based on 400m and 800m track results, I have the potential for RE > 1.00. What that will mean for performance in the second run split is something to be tried out. Racing is always learning by trial and error. In these short high intensity events, you always have to push your body past the limit to place well but you also have to throttle things down a notch to first finish ! 

In part 8 of this series, I assess if power readings and other Stryd metrics are affected by where you mount the device on the shoe. Testing was done with the shoes available to me.