
Stop running after 90 rpm: the cycling cadence myth debunked
Updated November 21, 2026 at 11:14 a.m.
Welcome to Dropped, a new weekly road and gravel cycling newsletter from Velo bringing you the latest ideas we’re always debating about group rides, plus a behind-the-scenes look at our coverage.
I thought we had left the 90s behind us, but it seems I was wrong. No, I’m not talking about the return of baggy jeans. I’m talking about cadence.
For years, keeping my cadence within an arbitrary range wasn’t even on my radar. Then I started playing with a few AI coaches – RestOrTrain and Ride with Dave – which take your data and give you feedback. Both systems kept reporting my cadence as a problem.
At first I ignored it. AI leverages information that already exists in the world and the old advice remains. It was so persistent, however, that I challenged each system to back up these claims with research. Both immediately complied with a vague explanation that maybe my cadence was good.
Okay, so the AI has retreated. This does not prove that the underlying advice is right or wrong. Curious about where I really was, I did every ride last year and found that indoors my average was 71.3 rpm, and outdoors it dropped slightly to 69.7 rpm. According to accepted cycling logic, this is a problem. So why is an arbitrary cadence range still taken as gospel, does riding at 70 rpm actually hurt anything, and what should you do instead?

The 90s and the myth of optimal cadence
This is a professional racing affair. We are currently in the era of short cranks. Before that, they were oval boards, and if you go back far enough you land in the “spin to win” part of the story. Just like short cranks, there is a real physiological reason why high cadence exists, but there is also the EPO era to contend with.
Power is simply force times cadence, which means you can move the workload between two different systems. Pushing a heavier gear at a lower cadence puts more strain on your muscles, while lighter gear changes tax your cardiovascular system. If you’re a WorldTour pro with a huge aerobic engine trying to survive a five-hour race or a three-week Grand Tour, relying on your heart and lungs to save your legs for the final climb makes a lot of sense.
This is also where EPO comes into play. In the 1990s and early 2000s, EPO artificially increased red blood cell counts and increased the rider’s ability to deliver oxygen. Suddenly the cardiovascular system was no longer the limiter it was. Spinning at 95 or 100 rpm became a way to harness that chemically pumped aerobic capacity while keeping the legs fresh for attacks. The problem is that everyday cyclists watched the pros furiously climb climbs on TV and adopted 90+ rpm as the universal gospel, without having a WorldTour aerobic engine, and certainly without EPO.

You must ignore the dogma of cadence
It seems obvious that you should ignore what professionals do, especially those who work on EPO. Still, I wanted to check my understanding with someone else, so I reached out to Ric Stern, performance director and head coach at CycleCoach, and the guy who invented the ramp test.
Stern’s response was simple: let the riders self-select. “I don’t believe there is a universally optimal cadence, and I certainly wouldn’t tell someone they should ride at 90 rpm just because it’s become an accepted benchmark,” he told me. Cadence changes depending on power output, terrain, fatigue and the actual demands of the event.
When I asked him if he would try to change my 70 rpm average, he said probably not. “If you’re comfortable, making the power you want, and aren’t having any obvious issues, I wouldn’t automatically consider 70 rpm a problem,” Stern said. Outside of niche cases like track pursuit runners – or occasionally giving someone a cadence goal on the indoor trainer just to break up the monotony – he doesn’t routinely prescribe cadence exercises at all.
Stern also had some thoughts on AI coaching reporting my data. “It’s relatively simple to identify a number that falls outside of a predetermined range,” he said. “It’s much harder to determine whether that number represents something that needs to change.” As Stern says, “cadence is a variable to consider, not a performance goal in itself.”

A low cadence might suit you
What AI is missing is the same thing that this discussion on places like Reddit is generally missing: nuance. My cadence is low because my goal is efficiency over ultra distances, not sprinting or responding to constant pushes during a race or fast group run. I ride regularly with a runner friend, and while he can drop a wattage bomb and lose me in a sprint, I can let him down just as easily by slowly building up my power and holding it there until he breaks down. I’m only pushing 160-170 watts, but I’m small and on an aero bike. At first he thinks it’s easy, then pretty soon he leaves – and I didn’t even bring a bottle with me.
Different runners and different situations have different cadence needs. At 70 rpm, I’m pushing more torque per pedal stroke than someone spinning at 90, but because my overall power output is relatively low, the actual force on my legs isn’t high enough to require relief at high cadences. As Stern pointed out, “a cadence that works well at 150W may be very different from a cadence that works well at 350W.”
This little detail about the bottle is also important. In a 2009 study on cadence selection, researchers Abbiss, Peiffer and Laursen found that for long-distance endurance racing, lower cadences improve cycling economy and reduce overall energy demand. Another study by Bieuzen et al. Give it a concrete number: Among well-trained cyclists, the most metabolically efficient cadence was 63.5 rpm, even though those same runners naturally chose to hover around 90.6 rpm. Stern noted the same thing, explaining that the cadence that minimizes oxygen consumption is often lower than that adopted by experienced cyclists. When your goal is to travel long distances without burning up all your fuel, going slower isn’t a fault, it’s just that your body is efficient.

What you should actually do
I like to end an article with a bang – a sort of grand conclusion and farewell – but the reality here is less about a new rule and more about rejecting generic rules. There are still times when training at a specific cadence makes sense, whether you’re preparing for the pushes of a criterium, managing knee pain, or just trying to stay awake on the indoor trainer. What doesn’t make sense is to search for an arbitrary number just because an AI tool or automated training asked you to. There is no universal cadence range that you should achieve as a rule of thumb.
Research confirms this. In a 2017 systematic review of low-cadence training, researchers Hansen and Rønnestad found no strong evidence that forcing low-cadence work is better than doing the exact same workout at a freely chosen cadence—and despite the old gym myth about cycling, grinding heavy equipment doesn’t generate enough force to develop maximum leg strength. A 12-week trial by Kristoffersen et al. came to a similar conclusion: trained cyclists performing 40 rpm intervals saw no gains in strength or performance, while the group performing the same intervals at their freely chosen cadence actually improved.
This also fits the way Stern coaches. He sees no reason to force it. Rather, he views cadence as a diagnostic tool. You only need to look carefully at your diet if you are:
- Struggle to maintain power during longer efforts
- You feel excessive local muscle fatigue (your legs give way well before your lungs)
- Having trouble reacting to sudden changes in pace during a run or quick group ride (assuming that’s actually your goal, unlike me)
Even if you hit one of these red flags, Stern wants concrete evidence that changing your cadence improves your output rather than just assuming it will. If you’re comfortable and getting the power you want, stop chasing a number from the 90s.

Is this news to you?
As always, I want to hear from you. Jump in the comments and tell me where your natural cadence is and if a coach, riding buddy, or app has ever tried to tell you to change it. Have you ever ignored the 90 spin rule or are you still convinced I’m wrong?

Gn tech