Cycling Trainer Mystery

tl;dr version – the Kickr is a solid trainer that I think will make a welcome addition to my training arsenal. I will sell the CycleOps Fluid2 to anyone who is interested.

Somewhere in the middle of last season, I finally caved and purchased a power meter for my bike. I have read time and again that training with power is the best way to improve your cycling ability. Coach Ben was able to secure a Quarq Elsa R for me. Yes, it absolutely felt like an extravagance, but given how much time and energy I put into racing, and how little I spend on other things, it felt like an OK bargain.

Having the power meter on the crank, and using Ant+ to communicate with my Garmin 305, meant that I was getting solid feedback on my training output while on rides, and could structure workouts accordingly.

Last year, I was also very fortunate to be able to retire my ancient cycling trainer when I won a CycleOps Fluid2 trainer at a training camp. It’s a reasonably priced piece of kit, and I was very happy to have it.

When I stared using it this season for training, I immediately noticed that I was suffering from the same problems as I had with my last one. The progressive resistance curve of the device means that you can shift while pedaling, and apply more power, and the resistance increases. It’s nice. My own observations have always been that my on the road watts (as measured from the crank) were always higher with less effort than what I was able to do on the trainer.

The knock-on effect of this was twofold. First, my confidence was impacted because I spend so much of the early season training indoors, and the numbers were just low. Second, the structured workouts from Coach Ben are always tied to my power output; specifically my measured max functional lactate threshold (FLT). The last test I did had a measured FLT of 262W, so we use 250W to account for the test being indoors and the shorter duration of the test. The CycleOps and my old trainer all seemed to really balk at the notion of letting me get over 300W. There is no calibration on these trainers, so the spin down is fixed. Therefore, depending on how tightly you crank down the flywheel, you will have variance in your measured power. You try to do it the same every time, but who knows how consistent you are.

The few times I have used a CompuTrainer, my take-away was always that the road feel of that device was far superior to the trainers I have owned and used. The work required to get up to my FLT and stay there was different. I never felt like my legs would rip out of the sockets if I tried to smash the pedals for excessive power output, or that the trainer would just say “no” under these conditions. This has been my experience with the ClcyeOps and my old trainer.

KICKR-profile-rightSo it was with great excitement and anticipation that I attended a Apex Sufferfest session at Rally Sport this morning. They bill this as the first and largest Kickr studio using PerformancePro. Quite a few qualifiers there. 🙂 See, my wife was super nice about my XMas gift this year, which is to say she understands that trainers are a personal item, and that I have been hemming and hawing about finding a used CompuTrainer or a Kickr. Over and over I went on this decision, reading endless reviews, or the one true review. She told me that gift was “one of those trainer things” but that I should make the choice to make sure I got the right one. The ability to test a device is not one to be passed up, so I happily made my way over this morning.

My first impression is that the device is very well built, and exceedingly easy to get my bike attached. This was much appreciated. I did not get a chance to play around with any of the software, so this write up doesn’t really speak to the functionality and usability of the software. DCRainmaker provides the super ADD review of the device and all software.

What I can speak to is the feel of the device. Riding on this device felt more like riding on the road. The power adjusted in a smooth manner. It did not punish me as harshly for smashing on the pedals. Coupled with the software the studio was using, I was able to experience two modes: 1) power, and 2) real course. The power mode is nasty. You input the power and RPM target, and the device forces you to hit that. If you pedal too slowly, or try to change gears, the resistance adjusts. Changing gears doesn’t matter. Only power output. It was a nice tie in with the Sufferfest video (“There is no Try” in case you are wondering). The real course mode was exactly what I want from a trainer. A file that has the topo and distance of a real course, and the trainer adjusts resistance based on grade. This is exactly what I want in a trainer, and Kickr delivered.

There is, sadly, one hitch, and I need to get to the bottom of this one.


Kickr Measured Power File as Reported To Strava


Quarq Measured Power File as Reported to Strava

If you look at these two files, they are mostly similar. The second one is a zoom in, as the session at the gym was actually two different rides in a two hour period. My watch recorded this as one session. What PerformancePro sent up to Strava was two separate sessions. I have done my best to zoom in on the one long ride to make the comparison similar. Because of this, the averages and max numbers in the left hand column of the Quarq file are not correct, as they apply to the whole ride. [Feature Request for Strava: when I click to zoom on a chart, recompute these numbers.]

We’re looking at the purple lines here. The Kickr reported numbers are much more discrete, as this was a session where I was required to hit power over time. I don’t pedal with that kind of mechanical precision, which is why the bottom chart has a slightly lumpier line.

Here’s the rub. That middle section that is about 8 minutes and a gradual build? On Kickr, as this was a percent of my FLT (which I input as 250w), the range was 225W – 245W over the build. It was going up 5W per 2 minutes. The data reported by my watch, however, was approximately 207W-230W. This doesn’t seem that out of whack, but earlier in the ride I was seeing variances as high as 40W.

After this 8 minute session I asked the instructor to bump me to 260W FLT to see if that would make a difference. It’s hard to tell, but it looks like it brought the banding in a bit closer.

However, the second session (not listed) was a real course, on a flattish (with some gentle incline and then decline) time trial course. If you are so inclined, you can compare the full data file here (click to zoom from 1:10 to the end of the ride) to this second session here. There appears to be about a 20W delta, with my Quarq reporting a lower number.

I know that there is going to be a delta between what a trainer says and what the crank says. What I don’t know is if I am out of the error bars. I also don’t know if I should be sad face or not. According to Kickr, I was around 220-260W during this 40 minute session, and this after I did a 1 hour Sufferfest. I should be elated. According to Quarq, I was mostly between 190W and 220W, though I did go as high as 260W.

Again, either way, I should be happy with these numbers given how much of a pain the transition to living at altitude has been for me. I suppose I will just have to wait until the snow clears off the roads before I can get back outside with something more closely resembling my racing legs. Only then can I really get a sense of what kind of power numbers I can produce on the road over distance.

Overall, I was impressed with the Kickr. Solid build, easy to get in and out of, and the road feel was impressive, and smooth. It’s not cheap. That said, this is my next trainer. Both Kickr and CompuTrainer have ability to tie into software and allow you to ride real course rides, but I find the API-centricity of the Kickr appealing to my inner geek. Comparing the UX of a Kickr to a CompuTrainer is like comparing the Space Shuttle to the Apollo Lunar Module. You almost feel analog using the CompuTrainer. It’s a rock solid piece of kit, but my gut is telling me the Kickr has legs.