Month: November 2022

I gravel bike the Wiouwash Trail from the public parking area near Bare Bones Brewery in Oshkosh to Hortonville, Wisconsin round trip. The Cykelstrom Bicycle Power Meter App is presented on the video and added graphics are used to show how the app compares to a crank based power meter, a map, and a gauge which shows the grade of the trail.

If you would like to see more Cykelstrom Bicycle Power Meter App videos, please visit @Cykelstrom on YouTube and remember to “Like” and “Subscribe” to our channel.

Trail Pic

Gravel Biking The Wiouwash Trail Cykelstrom Accuracy

The best accuracy for this ride is highlighted in bold.

Sensors	Crank Power Meter (Watts avg.)	Cykelstrom (Watts avg.)	Difference (Watts)	Margin of Error (Watts)
External Speed and Cadence	140.00	130.70	-9.30	+/- 0.9532
GPS Speed and External Cadence	140.00	121.52	-18.48	+/- 0.9534
External Speed Only	142.31	140.32	-1.99	+/- 1.12
GPS Speed Only	142.31	127.98	-14.32	+/- 1.07

The Gravel Bike For This Test Ride

I rode a Diamondback Haanjo Comp as a gravel bike for this ride equipped with a 4iiii Precision Pro dual sided power meter on a Shimano 105 11-speed group set. The cadence sensors were Wahoo Bluetooth cadence and speed sensors. Heart rate information was collected with a Polar H6 chest strap sensor. The smartphone was a MotorolaG6 running Android Oreo and Cykelstrom version 1.36. All of the power meter and sensor data and was collected, saved locally and automatically uploaded to Strava with the Cykelstrom Bicycle Power Meter App.

Discussion Of This Ride’s Results

I have highlighted in bold the results from “External Speed and Cadence” and “External Speed Only.” I leave it to you to decide which is more accurate. I think this points out one of the disadvantages of looking at averages when comparing two methods of measurement.

I would have a tendency to think that the data measured with cadence would be more accurate, however the data shows less of a difference between averages when no cadence information is considered when measuring the cyclists power.

Looking at the shapes of the two distributions shows me two things. First, when considering cadence information the distributions appear to be more similar than when cadence information is ignored. Second, the Cykelstrom power distribution is multi-modal with a couple of peaks. I hypothesize that a wind shift during this particularly longer ride is represented in Cykelstrom’s calculated estimates of power.

Bland-Altman Analysis

Other than the bias differences, these two data sets look very similar from this point of view.

I conclude that the Cykelstrom Bicycle Power Meter App provided particularly good power information to the cyclist when using a cadence senor. For information on how to set up a cadence sensor please see the article Adding Bluetooth Sensors To Cykelstrom.

Strava Activity

Click on the map to visit this activity in Strava.

Follow Cykelstrom Gravel Bike Test Rides on Strava

The Cykelstrom Gravel Bike Power Meter App Is Exclusively Available On Google Play

Gravel Biking the Wiouwash Trail starts from the public parking lot on Hwy S near the Bare Bones Brewery round trip to the Outagamie County line.

If you would like to see more Cykelstrom Bicycle Power Meter App videos please visit: @Cykelstrom on YouTube and remember to “Like” and “Subscribe to our channel.

Trail Pic

Gravel Biking The Wiouwash Trail – Cykelstrom Accuracy

Sensors	Crank Power Meter (Watts avg.)	Cykelstrom (Watts avg.)	Difference (Watts)	Margin of Error (Watts)
External Speed and Cadence	145.51	113.42	-32.09	+/- 1.26
GPS Speed and External Cadence	145.51	106.89	-38.62	+/- 1.22
External Speed Only	145.96	119.17	-26.79	+/- 1.44
GPS Speed Only	145.96	111.32	-34.63	+/- 1.36

Discussion of This Ride’s Results

The result’s from this ride show a bias which is higher than normal. Typically Cykelstrom is within ten percent of the crank power meter when riding gravel. In this case Cykelstrom was within twenty-two percent of the dual sided crank power meter. This may be acceptable for a mountain bike ride in rough terrain, however, I am not happy with this result on a gravel bike riding a rail trail.

So what went wrong?

If we take a look at the power distribution curves, we can see that the distributions are very similar in shape, however there is a large offset from the two data sets. In fact the offset is 32 Watts. Cykelstrom recorded data very similarly to the dual sided crank power meter, however, the offset was due to poor calibration of the app. It is very important to calibrate Cykelstrom with the rider sitting on the bike in a normal riding position on a flat and level surface.

Future Improvements

As the developer of the Cykelstrom App, I am always learning ways to make our app better. I am currently working on a new algorithm so our app will become self calibrating. Here is my call to action. If you would like to help participate in new versions of our app, please send me an E-mail and I will be happy to add you to our ride tester’s list. You will be able to shape the way Cykelstrom evolves and become the bicycle computer or power meter app you always wanted. Here is a link for your convenience: developer@cykelstrom.com

Importance of Calibration (The RED Dot)

A factor that is within a rider’s control to get an accurate calculated power estimate is to calibrate Cykelstrom on a flat and level surface before a ride. After you have securely mounted your smartphone on your bike, sit on the seat in your normal riding position and press the “Red Dot” on the face of the Cykelstrom App. You should see the grade percentage reset to near zero or level. Once calibrated, Cykelstrom will remember this initial position until you press the “Red Dot” again. Recalibration is not necessary unless you change the angle of your phone mount or change bikes.

To learn more about calibrating Cykelstrom please read the article “Setting Cykelstrom Up For The First Time.”

Validating Accuracy For Cykelstrom

Bland-Altman Analysis is a frequently applied technique to investigate the agreement between two methods of measuring the same thing. In this application we are measuring the instantaneous power of the cyclist using the Cykelstrom Bicycle Power Meter App (blue) and a single sided crank power meter (red) in the figure above.

The distribution of dots in the plot below represent the difference of the two measurement methods versus the average of the two measurement methods. If the dots are evenly distributed above and below the zero line there is no bias between the two measurements. In this ride, there is a 32.09 +/- 1.26 Watt bias. The smaller the bias and margin of error the better the two measurement methods agree with each other. We use Bland-Altman Analysis to validate the accuracy of the Cykelstrom Power Meter App. We are constantly tuning our calculated power estimate to be as accurate as possible across mountain, gravel and road bike designs.

The large bias demonstrated by the Bland-Altman Analysis was due to poor calibration. Properly calibrating the Cykelstrom App on a flat level surface will significantly reduce the large bias of the app compared to a dual sided crank power meter.

Strava Activity For Gravel Biking The Wiouwash Trail

Click on the map to visit this activity in Strava.

Follow Cykelstrom Test Rides On Strava

The Cykelstrom Bicycle Power Meter App Is Exclusively Available On Google Play