Is speed and cadence on a stationary bike adequate for Zwift? A real world test

Is speed and cadence on a stationary bike adequate for Zwift?  A real world test

Author’s note: A stationary indoor bike will be accurate in Zwift with a power meter. This article is about comparing a “spin bike” (which is a trademarked term) with a power meter to the same bike with a simple speed/cadence sensor.

From time to time I see questions on using a stationary bike in Zwift.  More often than not, the advice given is “grab a cadence and speed sensor and you are good to go”.  The well-intentioned suggestion utilizes the option Zwift provides of using a speed and cadence sensor with a non-smart classic trainer. 

Essentially Zwift has tested certain non-smart classic trainers and those supported trainers have a lookup table in Zwift. Basically. for X speed, you’re doing X watts. It’s a power curve table.  For Unlisted/Untested trainers they have another lookup table, but probably based more on averages that may or may not be accurate.  There’s a lot more to it, but I’ll invite you to read the excellent Zwift Insider post about virtual power vs going over it here again.

The Challenge with Stationary Bikes

Connecting a stationary bike as a trainer on Zwift can be a problem, though. Here’s why: a trainer has a set resistance (the pressure of the roller on the back wheel) and your speed increases with your wattage via gear changes and cadence, whereas a stationary bike has speed that is dependent only on cadence (no gears).  You can change resistance manually with a knob, but if you maintain the same cadence, you will maintain the same speed and thus, the same Zwift estimated power, regardless of where the resistance knob is set

To put that a different way, as long as you pedal at the same cadence on a stationary bike, your flywheel will spin at the same given speed.  The way you increase the wattage on a stationary bike is by putting resistance on the flywheel, resulting in more wattage to keep the flywheel going at the same speed. 

It’s worth mentioning here that resistance on a classic trainer can also be highly variable depending on how tight or loose the roller is against the tire.  Classic trainers have to have the roller tightened against the wheel to the trainer’s spec in order to be anywhere near accurate in terms of wattage. So it can basically be as variable as a stationary bike would be, although you can’t just reach down and tweak it on the fly.  Anyone who has experimented with that knows that the looser a wheel is against the roller, the higher the reported watts will be, and vice versa.  Hence the reason proper tightening is so crucial.

Back to the stationary bike.  When Zwift is estimating power solely by a speed and cadence sensor and knows nothing about the resistance knob, that estimated power number will be wrong.  It will be wrong because it’s missing an enormous part of the equation, the resistance, something it knows for a trainer, but can not know on a stationary bike.

Spindown for Stationary Bikes

How could Zwift know resistance and have accurate zPower with a stationary bike?  The only way I can theorize it working is to have a spindown test within Zwift.  Set the stationary bike resistance to where you want to ride, do a spindown to determine resistance, then keep the resistance at that setting the entire ride, using only cadence to make changes in power. 

Not a great way to ride, and Zwift would also need to know the stationary bike gear ratio (or use the speed which the stationary bike displays) for its calculations.

Testing Stationary Bike Power Accuracy

Enough explanation, let’s see it in practice.  To illustrate why a stationary bike isn’t accurate in the Not Listed trainer setting, I set up a few testing scenarios.  I set up a single Freemotion stationary bike with a Stages power meter and a Wahoo speed sensor.  I then set up two Zwift sessions, with exactly the same height, weight, bike, wheels, gender, FTP, etc.  I paired the speed sensor, HR sensor, and cadence sensor to one session via ANT+. I then paired the power meter, the same HR sensor and the same cadence sensor to the other session via Bluetooth.  The same pedalling effort would go into both sessions, power would just be output via different sensors.  

To put this into a visual format:

Session 1Session 2
NameHarry LegzScott DeLeeuw
Stationary bikeFreemotion s11.6Freemotion s11.6
In-game bikeZwift carbon w/ 32mm carbon wheelsZwift carbon w/ 32mm carbon wheels
FTP197 197
GenderMaleMale
Height5’ 10”5’ 10”
Weight177 lbs.177 lbs.
Cadence inputStages power meterStages power meter
Heart rate inputScosche Rhythm+Scosche Rhythm+
Power inputWahoo speed sensor, “Not Listed” trainer, 20” wheel settingStages power meter
Connection formatANT+Bluetooth

Note: the FTP setting was lowered to see if either of the tests would result in an FTP gain that wasn’t warranted.

For the first test I set up a MeetUp on Watopia Flat between myself and my identical twin nemesis Harry Legz. I configured the MeetUp to include only the two of us so variables like drafting did not come into play.  My first test was a cadence test.  I would do 2 minutes with no resistance at all at 70 rpm, 80 rpm, 90rpm, and 100rpm.  Then I would repeat these same 70, 80, 90, and 100 rpm tests, but with about 130W of resistance on the knob.

The graph below shows the comparison.  Cadence was the same as they were coming from the same sensor:

No resistanceNo resistance130W(ish) actual power*130W(ish) actual power*
Speed sensor zPowerPower meterSpeed sensor zPowerPower meter
70 rpm wattage130W41W128W134W
80 rpm wattage174W51W170W136W
90 rpm wattage209W61W213W135W
100 rpm wattage276W54W270W124W

*Note: I adjusted the resistance knob on each cadence change to maintain 130W(ish) via the power meter reading.

When I turned the stationary bike’s resistance knob to around 130W(ish) actual power, the results were pretty much exactly the same for the estimated power coming from the speed sensor as they were during the no resistance test.  This is expected since the cadence was the same as the first part of the test, and if the cadence is the same, the speed will stay the same. The graph below shows the data comparison visually.

One interesting observation on the no resistance test: true wattage actually went down at 100 RPM.  Since the watts are measured with a crank power meter, my take on that is the momentum of the flywheel started to take over, pulling the pedal through.  I was pedaling faster, but softer.

As you can see, there is a very large discrepancy in actual power vs estimated power, over 200W at some points in the graph, but what did it do for the ride stats?

Watopia Flat Route, ride time 23 min 24 sec

Speed sensor zPowerPower meter
Distance5.48 miles4.03 miles
Avg Power195W75W
Max Power400W162W
Avg Speed20.1 mph14.8 mph 
Max Speed30.5 mph23.8 mph

For the very same pedaling effort, the speed sensor rider covered a mile and a half additional distance in just 23 minutes!  That is almost 30% more!

Just for fun I also tested what would “MAX” out the 400W estimated power on the stationary bike. That happened around 115rpm.  As long as I maintained 115 rpm, regardless of resistance, I would peg 400W in my Harry Legz session.  (It’s worth noting that I could do 400 estimated zPower watts at around 50W of actual power, if the resistance was dialed down on the stationary bike.)

The cadence test is all fine and dandy, but I wanted to see what difference this made during an actual ride.  Maybe one with an interval workout incorporated into it to see actual larger power hills and valleys and how they would track between both setups.  Once again I set up a single Freemotion stationary bike with a Stages power meter and a Wahoo speed sensor, then rain two parallel Zwift sessions just like the first test.

For the actual ride I set up a one hour MeetUp on the 5.7 mile Watopia Hilly Route for the two sessions. I set the MeetUp to include only the two of us so variables like drafting did not come into play.   

This was my ride format:

  • 20 min warm-up
  • 5 min at 190(ish) watts
  • 5 min rest
  • 5 min at 190(ish) watts
  • 5 min rest
  • 5 min at 190(ish) watts
  • 5 min rest
  • 10 min cooldown

The result of this match-up?  Remember this is input from one bike.  My nemesis Harry Legz lapped me on Zwift KOM just as I was starting lap 3 of Hilly Route, rocketing past me at the finish banner.  He also took second fastest time (I honestly felt bad about that) on Hilly Route. At the end of the ride/workout, he received a nice fat FTP gain, despite only about 15 minutes of a moderate 190W effort.

Let’s look at the data.

Watopia Hilly Route, ride time 1 hour

Speed sensor zPowerPower meter
Distance21.73 miles15.11 miles
Avg Power254W132W 
Max Power400W264W
Avg Speed21.4 mph15.0 mph
Max Speed38.7 mph36.1 mph
KOM time during 110W 90rpm warmup3:29.166:52.69
Villa sprint time during 110W 90rpm warmup34.22 sec44.76 sec

While these numbers look dramatic, the graph of the power difference looks even more dramatic:

This graph does look more interesting at first glance than I thought it would.  zPower is about 100-200W high as before, but the power curves parallel each other more than I first expected until I really analyzed the data.  Why is that?  If you notice the cadence line I was soft pedaling at around 90 or so rpm for the warmup, rest intervals and cooldown.  During the intervals themselves my cadence would go higher to 100-105 rpm.  Since zPower is being estimated by speed, which is directly related to cadence on a fixed gear stationary bike, this makes perfect sense. A jump in 10rpm made a massive difference in zPower because it was making a big jump in flywheel speed. It was simply a coincidence that my higher true power was done at a higher cadence and that higher cadence gave a higher zPower number.

Searching for Sweet Spot

Discerning Zwifters may wonder if there’s a “sweet spot” on the stationary bike’s resistance knob where zPower and true power converge across a usable cadence range (say, 70-100RPM). Or to put it simply: can we make the zPower number be close to accurate on a stationary bike? 

The answer is: not really. Not without actively varying the resistance knob, and having a power meter to check your accuracy. (Of course, if you have a power meter, you don’t need a speed or cadence sensor or zPower anyway.) 

On my first cadence test when I had 130W of resistance on the flywheel and was pedaling at 70rpm, the numbers were nearly identical.  On the second test, if I had maintained about 85rpm in my 190W intervals power would have tracked closely.  I could have also cranked up the resistance to 400w and spun at 115rpm (for the whole minute I could maintain that).  But without a power meter I would have been guessing that the numbers lined up and inevitably what would happen is you’d loosen the resistance knob and drop actual power, but zPower would stay at 400W since your cadence isn’t changing. 

Conclusions

So there you have it.  I had fun doing this test.  It might seem pointless, but I wanted some real world data for when the suggestion is thrown out to just use a speed and cadence sensor with the Not Listed trainer setting on a stationary bike.  Seeing this data I don’t think it’s a good idea.  I get that it’s inexpensive and not everyone has the money for a power meter, but you are missing the resistance part of the estimated zPower equation.  Because of that, the power estimates are at best a guess, and a bad one at that.

None of this is to say a stationary bike is a bad idea for Zwift, a stationary bike can be a GREAT tool for Zwift.  If you have multiple riders in the family, or just want the simplicity of hopping on and riding, a stationary bike may just be the ticket. But please reconsider the incorrect speed and cadence route and set it up with a power meter such as a Stages crank or power pedals like the Assioma UNOs.  There are inexpensive options if you look around.  

I agree with the purists when they say a stationary bike won’t get resistance changes from Zwift, but Zwift will use the power you are putting into the bike to adjust your speed up and down in the game, the same as if you were riding a bike outside that dynamically adjusted gears to keep your cadence the same as you hit uphills and downhills.  Stationary bike or smart trainer, the true power you put into your pedals is what determines your speed in the game, the gradient feel is icing on the cake.  

I wrote another article about an inexpensive stationary bike setup that will work with Zwift that can save some cash over the long haul.  I found my entire stationary bike, with power meter, for $200.  If you are thinking about going that route, it’s worth taking a look.

Cadence and speed sensors on a stationary bike – summary

Pros:

  • Inexpensive
  • Can do rides and give Ride Ons
  • Breaks up the monotony of indoor training
  • Could be a gateway to a great Zwift future

Cons:

  • Not what Zwift intended with the speed and cadence option
  • Isn’t accurate, missing the resistance part of the equation
  • Not useful for personal fitness improvement measurements since zPower is only tied to cadence/speed, not how many watts you are actually producing
  • Can not do training plans as power targets are only tied to cadence/speed, not resistance
  • Will be very disheartening if you ever go to a power meter or smart trainer

Questions or Comments?

Share below!

About The Author

Scott DeLeeuw

Scott is just an average rider who loves using Zwift. Living in Colorado, when he's not riding Zwift, he is out walking his dogs Badger and Luna with his two daughters.

Subscribe
Notify of
guest
53 Comments
oldest
newest most voted
Inline Feedbacks
View all comments
Steven Nelson
Steven Nelson
11 months ago

I have used a Spin Bike since I first started using Zwift in the spring of 2016. I have put 27,040 Kilometers on it while doing Zwift. I do, however, have PowerTap P1 pedals installed on the Spin Bike.  My reasons for getting the Spin Bike are the whisper quiet belt, I needed to be able to ride it while my wife slept. The lack of electronics, Facebook is filled with people talking about all the problems with smart trainers. The safety of a solid front wheel. I have grandchildren walking around and don’t need a dangerous wheel and chain spinning around. There are… Read more »

Mel
Mel
8 months ago
Reply to  Scott DeLeeuw

Hello! I just acquired a Stages SC3 which has a power meter and I’m very excited to get started w/Zwift. Are there any rides you recommend for someone new to the platform?

Mel
Mel
8 months ago
Reply to  Mel

Also, strategies/approaches for accurately adjusting resistance manually on the bike to follow along with the Zwift courses?

Bobby Mac
Bobby Mac
11 months ago

Thank you Scott! You just solved a great riddle for me. I’ve been trying for 130 Zwift races and 2 years now to go from a Zwift C rider to a B rider. I’m also trying to figure out how to motivate Zwift HQ to fix their very broken racing system and have riders riding in their proper categorization. A riders keep crashing C races and screwing up the race dynamics. But this article just made a giant light bulb go off. I’m gonna sell my Wahoo Kickr and buy an exercise bike! I’ll immediately be an A Rider on… Read more »

Bobby Mac
Bobby Mac (@rmacdowell1)
11 months ago
Reply to  Scott DeLeeuw

Apologies Scott. My post was 100% sarcasm. Which doesn’t translate well in mssg board text. I appreciate the time you took to perform this analysis and share it with us.

k k
k k
11 months ago

As I know, Zwift only care about power, if we put https://buy.garmin.com/en-US/US/p/606555 into a spin bike, can we have a accurate measure in the game on the power?

William Fracalossi
William Fracalossi
11 months ago

I appreciate and agree with this article, and especially w the premise that an existing spin bike is way less expensive to get a start on Zwift or Rouvy etc. That being said, I rode around 50 hours, close to 1000 “miles” on my professional quality Spin bike used in a large metro spinning studio before “I saw the light and rec’d a ” proper trainer” for my birthday: (Kickr Snap) long story short, my 400 watt maxes went the way of the dodo bird, and my 370 Zwift assigned ftp as well! I mentally and physically struggled with my… Read more »

Aloveitt
Aloveitt
11 months ago

I use a Keiser m3i which is a great spin bike. Quiet, smooth etc. Keiser does sell a Bluetooth dongle which will send out your power output. I had nothing but trouble with it (would work great then start intermittently cutting out mid way through a ride, super frustrating). Their support was great but we could never get it sorted out. I switched to Assioma pedals and have never looked back.

Brandon Johnson
Brandon Johnson (@brandon_j_585)
10 months ago
Reply to  Aloveitt

Also had the same issue with the Keizer m3i. I returned the M3i and bought the Wahoo KICKR Bike. Best decision ever. The number of F-bombs per Zwift ride dramatically decreased. However, for someone who isn’t a Zwifter, the Keizer bike is a great spin bike.

dan
dan
11 months ago

queue the ZP people trying to justify…. put them on a kickr or neo, and see the reality bite

William Fracalossi
William Fracalossi
10 months ago
Reply to  dan

Agree, Dan. I had an extremely inflated opinion of my abilities based on the numbers ZP assigned to me, which came crashing down in flames on my first “real” ZWIFT ride. Took me two and a half months to climb to 154 FTP. Now I know why there is so much animosity aimed at ZP riders. Too bad we couldn’t have all ridden together in an overlay world without approved trainer riders.

james ginn
james ginn
9 months ago
Reply to  dan

My Tacx Blue Matic gives the exact same numbers as the Elite drive in the gym. So the set up is important.

Jason Smith
Jason Smith
10 months ago

I have a Schwinn IC4. It transmits a power number to Zwift that is calculated using algorithms similar to ZPower. The bike’s console calculates the power number based on cadence and resistance level. It is highly inaccurate. I have seen the bike transmits power numbers that are anywhere from 50-150 watts higher than my P1 pedals transmit on a side by side test. It is true that the only way to have a spin bike transmitting accurate numbers to Zwift is to use power meter pedals, unless of course, you have the money to buy an indoor smart bike

Jim
Jim
2 months ago
Reply to  Jason Smith

Yup I got an IC4 last Sept — great way to start. Moved up from D to B and then when I was averaging 300+ watts in a hour race — looking like I was going to be bumped into A group — I figured that the bike was off. So with power meter pedals — I now average 200 watts on a long ride — so the bike was off by easily 33%. If a person is not into racing and just wanting to cycle around in group rides and workouts then the bike by itself is fine.

Alexssandro Loyola Freitas
Alexssandro Loyola Freitas (@freitasloyola)
10 months ago
Excellent, thanks. I ask if zwift in the future would not include at in equation: speed, cadence and heart rate on the spin bike.
james ginn
james ginn
9 months ago

So why are the true power numbers on a spin bike so low ? I use my Giant defy with speed and cadence sensors and my numbers are the same as on the Elite drive in the gym.

Scott S.
Scott S. (@redhotforjesus)
8 months ago

Three questions…

  1. How does using a HR monitor factor into the equation (using the Zwift Companion app)? Doesn’t the increased HR indicate greater output of energy and would logically be used in determining calories burned?
  2. How does a pedal-based power meter work as opposed to a cadence monitor attached to the crank? I am missing the way in which a PM knows anything more based on how fast the pedal is going around than a cadence monitor would based on the crank revolutions.
  3. Do you know of any crank-based power meters that would fit a Sunny SF-B901 model spin bike?
Robert Kinghorn
Robert Kinghorn
8 months ago
Reply to  Scott DeLeeuw

Hi Scott, Great article. I’m not a zwift rider yet but eager to get on. I’ve been looking for a inexpensive way to get on Zwift but also share a bike with my wife who wants to try some basic virtual spin classes. I ride a dumb trainer now and was considering upgrading to a smart trainer but I’ve been reading through your other article on the Freemotion &AD200 and that would work for both of us. In your experience do all the Freemotion S11.9 have the power meters and any issues connecting with Zwift? Thanks!

Robert Kinghorn
Robert Kinghorn
8 months ago
Reply to  Scott DeLeeuw

Thanks Scott, really appreciate it! I will keep an eye out for sure, currently just spin gym resellers asking around $800.

Robert Kinghorn
Robert Kinghorn
8 months ago

Canadian dollars that is, going to use a Sunny Health with a wahoo cadence until then.

Joe Smith
Joe Smith
8 months ago

Hi Scott,
What about a Raspberry Pi calibrated to the spin bike to accomplish simulated power? There are developers who offer code for the android device.

Joe Smith
Joe Smith
8 months ago
Reply to  Joe Smith

And would that still be considered zpower? Thanks

Tobias
Tobias
7 months ago

just discovered HR2VP (https://bipr.fr/training-with-heart-rate-to-virtual-power) which uses the actual HR to match the power within Zwift. Seems to work surprisingly good when comparing to power meter pedals (https://www.youtube.com/watch?v=wCBe9t0AXdc). Only a delay of 5-10sec between increasing power to displaying and experiencing in Zwift are to be expected.

Already tried it with my Schwinn C8 and this made my impressions much more positive then only Zwifting with completely inaccurate power output

joe smith
joe smith
7 months ago
Reply to  Tobias

Hmmm, maybe it’s close. But I know from experience that my h/r isn’t always associated to my real power. Like Scott says, torque is the only way power can really be measured accurately. Take a look at Assioma UNO for an inexpensive way to add power to your C8, if pedals can be switched out.

Tobias
Tobias
7 months ago
Reply to  joe smith

Completely agree! For me as a hobbyist it was/is simply a much cheaper way compared to an actual power meter attached.

Tobias
Tobias
7 months ago
Reply to  Scott DeLeeuw

that is actually a pretty good question and I was sceptical as well. I saw that the result matched best if I used real results from former FTP tests with power meter (before-pandemic-gym-times). You can use them as reference for the algorithm to do the “magic”. As I’m only using it as a somehow comparison to those former times maybe you could do an actual test using your power meter and this solution. But anyway, it’s not the perfect solution but for me as some how “technical fixated and interested” it does the work and helps me stay motivated. 🙂… Read more »

Nk the resistance way up, which
Nk the resistance way up, which
7 months ago

The “same cadence = same watts” regardless of resistance does not seem to be true on my Scwinn IC 8. The resistance control is magnetic. When I increase the resistance and pedal at the same RPM , my power on Zwift goes up. If i crank the resistance way up, which reduces my RPM ‘cause its hard to pedal, my power on Zwift goes up. Perhaps the magnetic resistance makes the power output more accurate?

Kim Santen
Kim Santen
7 months ago

This seems right to me.

Jim
Jim
7 months ago

I’ve been using a Schwinn IC4 for 4 to 5 months and I improved over that time and its been great, BUT I went from D to upper B and I started thinking too much improvement in FTP and watts. So I just got some Assoima pedals and I’m back to earth. I was pushing 300-330 watts with the bike and now with the same resistance the pedals are telling me 200-230 watts. Time to reset my FTP and get down graded on Zwiftpower…

ChiRider
ChiRider
7 months ago

I setup my dumb Schwinn Evolution with speed/cadence. I was unhappy with the unspecified trainer setup so I chose a nashbar dumb trainer setting. At least now I can keep up with the rest of the some of zwift crowd. But riders are constantly blasting past me and thats ok. But if what you say is true noone could keep up with me at 90rpm. My cadence is usually around 80rpm and I’m 142kgs. watts ~230ish. I’m sure my z power is skewed due to my weight. I’d probably crush power meters. like everything else the estimations wouldn’t apply.

NKSunny
NKSunny
4 months ago

I set up my Sunny stationary bike with speed and cadence on Zwift. I’ve read and re-read your article, and conclusions regarding its inaccuracy. I commute on a fixed gear, so riding without gear changes suits me fine. For the first couple of months I was getting steady and consistent readings, they may not have been accurate, but they were reliable for my exercise and enjoyment of zwift. I would race in D and C category and consistently come in last, so I definitely was not benefitting from an unfair advantage. More recently i’ve been noticing some crazy swings in… Read more »

Nathaniel
Nathaniel
4 months ago
Reply to  Scott DeLeeuw

Thanks for the reply, and point taken. Perhaps for next winter that will be the solution, though will have to invest in some cheap road shoes as well to pair with the Assioma pedals (investment is a bit more than the $100 for sensors, but maybe also manageable). For the time being, and while the weather is still uncertain for another month or so, I was hoping for some insight as to why I might be getting these crazy readings. As mentioned in your article, finding a resistance setting should manage to get readings that are reflective of effort, even… Read more »

Nathaniel
Nathaniel
4 months ago
Reply to  Scott DeLeeuw

Good point, I haven’t.
I’ll hook it up to my fixie this weekend and see what happens.
Thanks Scott for the replies.

Craig
Craig
4 months ago
Reply to  NKSunny

As Scott mentions get the Assiouma UNOs for accurate power. If you value Zwift it’s worth it. My wife’s setup, Peloton with UNOs. Opens up the door to Zwift and real power.

Craig
Craig
4 months ago

+1 for Assiuma’s. I put the UNO on my wife’s Peloton and now she can ride Zwift with very accurate power. If you already have a spin bike, and that’s your indoor setup, then upgrade it to REAL measured power with a setup that will a no brainer. I hear you on the other side, I’m a B rider now and there’s no going back that I know of. Unless you cancel your Zwift account and resign up with a new account.

Free Zwift Trial

Newsletter Subscription

53
0
Would love your thoughts, please comment.x
()
x