0

I have a Bryton 420 bike computer and my friend has a Coospo BC200. You might think Bryton is a better brand and would produce more accurate data but it doesn't. When I upload my ride to Strava I see power data all wrong. E.g. it shows 90w average power when I am not even pedaling on a descend while Coospo calculates 0w on same situation. What do you think about it?

I also wonder if bike computers use heart rate data to calculate power data. Do they generate more accurate data with a heart rate sensor connected?

P.s. I know these devices are too bad at calculation power data without a powermeter but they're too expensive for an amateur like me.

Here is the link of our ride if you want to check: https://www.strava.com/activities/9451045533

9
  • 1
    Comparing the rides - you had a HRM on and Ali did not. For the whole ride, your heart rate never got into Tempo / Zone3. Heart rate peaked at 129 BPM around 400 metres from the turn-around at the top. The ride was a casual cruise not a hard effort right ? Here's a ride with some hard effort to compare: strava.com/activities/8616632940/heartrate
    – Criggie
    Jul 14, 2023 at 23:55
  • 1
    I also made you some segments to compare - but after a wait there's no other rider's efforts. Is strava unpopular in Turkey ?
    – Criggie
    Jul 14, 2023 at 23:55
  • 1
    It's kind of popular in Istanbul but not in my area. Actually bicycle riding is not popular as a sport in my area. Regarding our effort I have an ongoing knee issue since so many years so I don't want to push myself much. Also I'm only 51 kgs despite being 175 cm tall so I can't generate much power.
    – Ender
    Jul 15, 2023 at 5:48
  • 1
    I've checked the segments you created on our ride. Our average power and speed are exactly the same on the climb. On the descend though, average power of my friend is just 19 watts while mine is 82 watts. Strava algorithm thinks this guy is so weak that he needs to generate reasonable power to move the bike at a moderate speed on a -2,6% descend. In real world, it's not that hard. Anyway... Thanks for your help. Could you delete the segments as we are done with our analysis? :)
    – Ender
    Jul 16, 2023 at 6:42
  • 1
    this might be better in Bicycles Chat than comments. I'll redo those names now.
    – Criggie
    Jul 18, 2023 at 8:33

4 Answers 4

3

Bike computers do not measure power data. They will need to connect to a power meter for this. I am not aware that any of them estimate power data from speed, either.

When you connect a power meter to the bike computer, it will record that data and upload that to Strava. If you don't have power meter data, Strava will estimate the power it would have taken you to complete the segment based on, at minimum, "weight, speed, and elevation change."

Speaking of calories burnt, if you have a power meter, this estimates the work you did to the drivetrain. Strava estimates the amount of energy your body burnt to do this work. I've covered this elsewhere, but it assumes a figure for gross mechanical efficiency; people vary enough that you should probably assume that there's an error margin of 5-10% on either side. Anyway, if you have a power meter and a heart rate monitor, I do not believe the HRM adds anything to the estimate for calories burnt. If you have an HRM but not a power meter, then Strava will estimate calories based on your heart rate during the ride, but this has a wide error margin. I recently ditched my power meter, and comparing similar rides done with and without a PM, my HRM-based calorie estimate could be higher than the power meter estimate by as much as 50%. If you have neither heart rate nor power, then Strava does still report calories, but this should be from its estimated power.

Speed is taken from the GPS track. GPS sensors have poorer accuracy than dedicated speed units, although most people don't use these (these will either use an accelerometer or magnet + reed switch to count wheel revolutions, and you'll need to measure the diameter of the wheel + tire). Weight is your own weight plus your bike's weight, so if you haven't accurately weighed yourself and your bike and updated that in Strava, that will be off. I remember that there are additional simplifying assumptions. For example I believe it may assume a constant temperature (and therefore air pressure), and zero wind. Or at least it may have done this in the past. It would have to assume your drag coefficient from your height and weight, and make some assumptions based on your bike. It would have to assume a certain rolling resistance, etc.

Once, there was a device called the Polar PowerCal which estimated power based on heart rate. It was US$99. It was offered around 2012, when power meters were quite a bit more expensive than they are now. Each person's maximum heart rate can vary quite considerably, so the accuracy of the PowerCal estimate probably also varied considerably. In fact, the PowerCal originally contained a function to calibrate it against an actual power meter. So, someone could certainly rig an algorithm to estimate your actual power from your heart rate data. I don't know how you would program this into your computer, but someone surely could. However, it would be quite a bit less accurate than actual power meter data. Power meters are fairly widespread. The juice probably isn't worth the squeeze.

5
  • Regarding “GPS sensors have poorer accuracy than dedicated speed units”, from my understanding, GPS is better for long-term speed/distance but worse at instantaneous speed/distance. The GPS only updates every second or so, so it’s bad at instantaneous stuff, but on the other hand the wheel magnet etc will accumulate errors over time. Would you happen to know anything about this?
    – MaplePanda
    Jul 15, 2023 at 2:55
  • Do you think Strava would generate more accurate power data with a speed sensor?
    – Ender
    Jul 15, 2023 at 5:44
  • 1
    @MaplePanda the wheel sensor is as accurate as its calibration. GPS not only gets the location every second or so, but also is not perfectly precise, so it will be very accurate when riding in straight line and lose accuracy in sharp turns. I don't know if any manufacturer uses GPS to calibrate the wheel sensor.
    – ojs
    Jul 15, 2023 at 6:11
  • 2
    Using the terminology from en.wikipedia.org/wiki/Accuracy_and_precision I would say that wheel sensor is very precise but its accuracy depends on calibration. GPS on the other hand is accurate but not always precise.
    – ojs
    Jul 15, 2023 at 6:14
  • 1
    @ojs I agree with your conclusion. Thanks for the analysis.
    – MaplePanda
    Jul 15, 2023 at 14:52
5

Without a power meter, power data is generally pretty meaningless (with the exception of perhaps average power over the entire ride). There are a lot of factors which determine how much power you need to output in order to sustain a given speed at a certain point along your ride. The ride tracking software needs to make a lot of assumptions about all of these continually variable considerations, leading to major inaccuracies in the estimated power. It might be accurate enough for you to know whether you were pedaling soft/medium/hard, but definitely not the single-watt accuracy you get out of a real power meter.

8
  • The CooSpo has support for cadence sensor and it wouldn't be rocket science to detect pedaling vs freewheeling from accelerometer data. It wouldn't of course make the rest of the power readings any more reliable.
    – ojs
    Jul 14, 2023 at 19:08
  • So does Bryton...?
    – Ender
    Jul 14, 2023 at 19:11
  • @Ender the power reading without power meter is still a bad guess, Bryton just puts even less effort to it
    – ojs
    Jul 14, 2023 at 19:17
  • 2
    Point of info: it's Strava that's estimating the power using its own internal algorithm, which makes a lot of simplifying assumptions. It is using the Bryton's GPS track, that's all.
    – Weiwen Ng
    Jul 14, 2023 at 19:17
  • 1
    @Ender as I explained, Strava estimates your power from its algorithm. I don’t know of any bike computer that does this. I don’t know why you and your friend have very different power estimates, because you didn’t provide enough info to answer the question. However, if you’re on an MTB and he’s on a road bike, and if Strava knows this, then that could be part of the reason. If this is an actual concern, you could update the question. Otherwise, people have answered the question that you originally asked.
    – Weiwen Ng
    Jul 14, 2023 at 21:50
4

To quote Strava:

When data from a power meter is not available, Strava can estimate power using information about an athlete's weight, speed, and elevation change.

As far as I’m aware there is no public information available how exactly Strava is calculating the estimated power, but they have to make some very rough guesstimates about your friction losses, rolling resistance, aerodynamic drag, weather conditions and so on. You could be riding 40km/h for an hour with a strong tailwind or a super aero time trial bike and Strava would probably give you huge power estimates because it has no better way of knowing.

Power meters are built into the pedals, crankset or rear hub, connect to your bike computer using Bluetooth or Ant+ and cost several hundred Euros. Their measurement accuracy is usually in the ±3% range or better.

2
  • 1
    Why does it ask for bike weight then?
    – Ender
    Jul 14, 2023 at 19:41
  • 3
    @Ender: So it knows the system weight (bike+rider) which is very useful for estimating power on ascents. On steep-ish ascents you are mostly fighting gravity and it’s pretty trivial and accurate to estimate power for an average road bike on an average road if you know the system weight, gradient and speed. On flat-ish terrain there are so many other factors.
    – Michael
    Jul 14, 2023 at 19:47
3

I checked a few old rides and Strava does indeed take pedaling into account on descents, it is just bad at making accurate point-in-time estimates and is often a bit offset in the track, so there definitely are gaps between Strava assuming power input and actual pedaling in my older tracks before I got a power meter.

Here's an example:

enter image description here

I think you can see that there is a huge gap of several seconds between the power spike and actual pedaling. It's a 2021 ride and they might have improved the algorithm but I'd say it is pointless to rely on data at zoomed-in level. I'm pretty sure there was no 400W power-spike when I started pedaling at the bottom of a descent into a flatter section. No way, map data can't provide fine enough resolution so that you could estimate what is actually going on on the road.

However, the average numbers can be trusted, in my opinion. That's not so hard to calculate and works well on full rides or segments of at least 1 minute or so, similar efforts are in the same ballpark as those measured with my (single-pedal) powermeter, unless they are tampered by wind. It might be that Strava gives you 50 watts more than you had to input for a given speed (riding into ha tailwind), it levels out if you have a route with equal amount of wind exposure, like riding the same straight both directions with head- and tailwind or a loop on open roads with hardly any wind cover.

2
  • Generally Strava calculates about %15-20 less average power comparing to powermeter data. There are some experiments on the youtube.
    – Ender
    Jul 16, 2023 at 7:00
  • 1
    Doesn't look as bad for me on climb segments, but as others have explained, depending on your setup and riding style, there are many variables Strava just can't take into account and hence, it'll always be a guesstimate. Still works to compare relative effort for your rides only because it'll be consistent within those boundaries.
    – DoNuT
    Jul 16, 2023 at 11:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.