Left right powermeter data analysis

More and more powermeters nowadays are equipped to measure how power (better said: torque) generated is distributed coming from the left and the right leg. A good example of these meters are the Look Keo Power system and the Garmin Vectors. The Pioneer pedal system even goes further and shows data on different angles of the pedal stroke. What you can actually do with this kind of information – how actionable it is towards increasing performance – is however still questionable. If you read some on this topic you find it mostly in the ‘recovery from injury’ sphere, but you won’t (at least to my knowledge) find any actionable results from scientific literature (let alone the fact that a lot of this literature is often not even actionable).

So why did I want to look at this information? Because I was just interested in it. If you’re interested as well continue to read or otherwise stop here.

Before I show some results of the analysis I will say something about the setup. As you’ve probably already read I ride with a power2max. The analysis I did was based on the ‘normal’ Power2max (not the Type S). This Power2max was updated with left/right balance and temperature compensation a year ago. I use a Garmin Edge 500 head unit and I have Rotor Q rings (52/38). Why do I say this? Because the Q rings might influence the left right readings since power is distributed a bit different than on normal round rings. This does –as seen from some posts- influence power readings and maybe also causes the left right assumption (see below) even more doubtful and influence the results. I am not an engineer, but it would be nice someone could elaborate on that more.

Now, the measurement of the left right balance is done by measuring the stroke across two sections, but essentially it does this by taking the right side. Looking from the right of the bike (where the chainrings are) from 12pm to 6pm (downstroke) for most cyclists most of the power is exerted (right leg) while the left leg basically comes into play on the upstroke (6pm to 12am) . Basically the Power2max makes the assumption that power from this upstroke comes from the left leg. See below a picture (reproduced with permission of Ray Maker / www.dcrainmaker.com).

 

So the way it measures left/right balance is not like a pedal based system. I have tried some left/right drills (as DC rainmaker) and it won’t show 100% left or right power. DC Rainmaker also has done some tests and it came out his rides were dominated by his right leg (47% left / 53% right). Those numbers came from a rather steady (in terms of power) ride.

 

Data used

For the analysis I used a variety of data,

-       1 crit race (40km) with a hill in it (31 aug 2013)

-       1 Les Cingles (3 x Mont Ventoux) ride (4 jul 2013)
-       1 hour FTP test (10 jul 2013)
-       1 Mont Ventoux Climb and another climb (27 jun 2013)
-       1 VO2max 5 min and 20 min FTP (19 feb 2014)
-       2 x 20 min @ FTP (22 feb 2014)
-       A ride with 40min @ FTP (26 Feb 2014)
-       1 hour FTP ride (5 March 2014)
-       5 x 3 min @ VO2max ride (7 March 2014)
-       A 3.50h easy ride (8 March 2014)
-       6 x 3 min @ VO2max ride (14 March 2014)
-       Trainingrace 50km plus extra endurance kms (16 March 2014)
-       4 VO2max intervals of varying length (10, 8 etc mins) (19 March 2014)

I looked at left right balance from 2 dimensions: cadence and power. Why? Because the way in which power is generated by using difference cadence levels might influence the left/right balance. In other words, you might have a different balance for different cadences and for different power levels. Maybe for high power levels a more right balance and for lower powerlevels a more left balance.

 

There is also the dimension time. The idea behind this is that you might change the balance during the ride. For example the longer into the ride the more ‘shift’ to a certain balance or more shifts. I have not done any analysis on this dimension. I focused on the above two parameters as you will see.

First I show you some very simple graphs to tell just some overall relations. In the first graph you see the relation between power and cadence (cadence is binned into cadence ranges).

 

    

In general when cadence goes up, average power goes up. Why is there such a strange spike? Because there are quite some rides at FTP level andit basically shows my preferred cadence. My preferred cadence is around 88-90. After that you see power drop again. Why? Because it contains a mix of higher cadence with lower power and high power (sprinting, etc) with higher cadence. This mix is more ‘favorable’ for the lower power numbers since cadence if (of course) low and power also.

If I relate power output to the power balance I get the graph from below:

 

This graphs looks a lot like a graph which I have once seen on Andy Coggan’s Facebook. The only difference is that his numbers went to 50/50 for higher powernumbers.If you see this graph you can say right leg dominance on the whole spectrum. 

Cadence in relation to the power balance looks like this:

For very low cadence (below 50) show a more ‘evenly’ distribution as well as cadence around my ‘preferred’ or self-selected cadence . After the self-selected cadence the higher the cadence the more dominance to the right balance.

The final graph shows the relation between cadence, power and the power balance.

 

 

At higher cadence with lower power and low cadence with high power there is a shift to the left balance. On the other diagonal this difference is (way) less profound. In the middle of the image the left % goes to on average 49.8% (this is the average of the cadence range 70-90 with corresponding power 250-350).

Conclusion

The overall conclusion of this piece. What to do with it? The honest answer is nothing. Even if the power measurement of left/right was accurate  I still wouldn’t know what to do with it. Basically a lot of questions are unanswered. What is the ‘optimal’ left right balance to produce power? And what is 'optimal'? Maybe the length of my legs make that this is 'optimal', or maybe not. Maybe my position very much influences these numbers and as such it is 'optimal'. Maybe it is 'optimal' for short time durations or for long durations? In a sense I believe these numbers should be investigated in correspondence to the target performance. So when you’re a 1km pursuiter you should focus these analysis on these performances. Obtaining symmetry cannot be the goal of this, but performance is.

I don’t not really see a 50/50 relation. If you have ‘a fluent powerstroke’ these numbers should be 50/50 (that's what I have read from a quote from P2M somewhere). Most of the time relation is skewed to the right leg (meaning more power from the right leg). As said before I am not an engineer, but maybe the way in which power balance is measured AND the Q ring influence these results. Let alone the way lef/right balance is measured, or better said ‘estimated’. A quote from Power2max (Source: www.dcrainmaker.com) “Power2max themselves says there’s little value in it other than from a marketing standpoint (quoted at Eurobike).’’

I do remain curious of others have also done some analysis on the Power2max numbers and come to the same conclusion n. For me it was just a nice to see, but nothing more. When there is not more (real research) on the topic it remains something under the surface and there are more fun things to play with to give better performance results.