The problem is that cadence doesn't tell you that you are putting extreme stress on the knees or not . You can be putting a ton of stress on your knees at any cadence, by simply pushing too hard at that cadence. Cadence is simply how fast your feet and cranks are moving, and has absolutely nothing to do with the stress/power being applied to the drivetrain through the knees. A "mashing" cadence of 50 is fine at 50 watts, and a cadence of 90 is stressful at 1,000 watts, yet traditional cadence wisdom would state the opposite. Without knowing the watts, you and the cadence meter are simply in the dark about the stress being applied to the knees.

If you REALLY care about your knees (possibly for some good reason), forget about cadence and just get a power meter. If you don't get a PM, and you end up hurting your knee(s) because you couldn't properly monitor the stress to them, then the insurance deductible for a knee MRI will be equal to, if not more than, a power meter, so why not just plunk down the money for a PM now, and avoid the future injury and knee Dr. altogether? You'll thank me later.

That I certainly agree with. I can do that. I simply don't for the most part.

That I do not agree with entirely because it does not account for my own, subjective experience as a rider. I know that I pedal with less force, on average, when I pedal at a higher RPM. How do I know? I know because it's me on the bike and I can feel it. As such, I can indeed use a cadence meter RPM alone to achieve that outcome, albeit with limited precision. Obviously, it is possible for a rider to know something about their own riding characteristics without purchasing a power meter to confirm it.

As RPM alone won't tell you the force applied to your pedals, neither will watts alone. Power, as you know, is FORCE X DISTANCE / TIME. And, in that equation, distance over time is really a function of RPM. So, with the watts output alone, you don't know whether the power that you're generating is mostly a function of the force of your pedaling or the rate at which you are pedaling.

If I'm not mistaken, power meters actually use pedaling RPM as an input to be used in calculating the power data that they provide. Power meters are, themselves, cadence aware even if they're not showing the cadence to the rider or the rider is not paying attention to the cadence.

Don't get me wrong, I'm a fan of power monitors. Once they come down in price by 50% or so, I may even get one.

Do power meters report pedaling force directly, separate from watts? I would certainly that would be possible.

So just stick with your cadence meter then, since it's half of a power meter. It's just that the half you're missing is the important half. :)

It's the important half IF the power meters report pedaling force directly. Do they? Do they also report RPM directly?

LOL, yeah they literally just changed its name yesterday to Wahoo SYSTM to make it sound less intimidating.

Power meters broadcast power and cadence directly. They don't usually broadcast pedal force directly.

You don't need to measure either cadence or power to know whether or not you are stressing your knees. You can simply feel it. If you are riding hard and deliberately mashing big gears then your knees (and all major leg muscle groups) are going to be more loaded than riding equally hard, but spinning lower gears. This much should be pretty obvious to anyone who has been riding bikes for any length of time. Of course riding at 50W is easy at any cadence because you are barely ticking over. But once you are putting out any kind of serious sustained effort then higher cadence will be easier on your legs in terms of muscle loading. Although there does come a point where spinning uncomfortably high puts stress on your joints from the sheer velocity. I sometimes feel this in cadence drills at very high cadence (150+ rpm). But within a normal comfortable cadence range, leg loading becomes proportionally lower with higher cadence for a given power level. Power = Force x Cadence.

I ride with a power meter and it always amazes me how perception of "power" can be quite misleading at times. We "feel" pedal force more than we "feel" pedal speed and so our brains tend to associate a high pedal force with a high power output, regardless of cadence. But quite often I find I can generate more power from a lower gear, even though it "feels" easier (at least until your HR starts to climb!). The classic example is dropping into the small chainring at the start of a steep climb. It feels instantly like your power has just dropped off a cliff as you spin up and then you glance at your power meter and find that you are actually putting out more power than you were just before you down-shifted. But your senses are confused both by the reduction in pedal force in the lower gear and your deceleration as the slope kicks in. The fact that you are spinning much faster goes pretty much unnoticed by your brain until your lagging HR responds and you start breathing more heavily.

I'm going to be honest here - that does not count as evidence that you are riding at your optimum cadence. It is simply evidence that you ride pretty fast. :)

Interestingly, if you go back and read the article I linked from Cinch Cycling, I think it mentions this as an example where a rider might choose to start in a slightly higher gear, the intent being to avoid burning too hot on the start of the climb and not being able to finish strong.

If it isn’t that particular article, I apologize; I’ve read several articles about cadence on their site and it may be a different article.

Anyway, the point being that a climb is a situation where you might choose to gear a bit higher to avoid putting out too much power. But to your comment, yes, climbing in that lower gear probably does let you increase power.

Otto

I'm going to be honest here, I don't really think anyone is producing evidence here, just a bunch of assertions and suppositions..

Well, here’s Formenti’s paper on internal work. See table 2, page 8. And I remembered two of the values incorrectly: 90 rpm work ran about 0.6 W/kg, not 0.7 and 110 rpm work ran around 1.0 W/kg, not 1.1.

https://kclpure.kcl.ac.uk/portal/fil...12500.full.pdf

Otto

Riveting I went for a ride after reading you most recent comments and noodled on power vs cadence the the whole way. I believe that I've warmed up to your perspective substantially. Here's how I see it now:

Yes, one could fire up excel after a PM ride, juggle some numbers, and know their pedaling forces. And, similarly, it would be programmatic child's play for the PM manufacturers to show pedaling force in real time if they cared to. At the end of the day, however, I don't think that most folks are interested in managing the magnitude of the their pedaling force. And those that are interested in managing it are likely nursing existing knee injuries, not trying to prevent new ones. Heck, I don't even know what a desirable level of pedaling force is from the perspective of knee health.

It seems reasonable to me to assume that for each rider, in each circumstance, there exists a "Goldilocks" RPM that maximizes power output. I think that a precise definition of "mashing" is not merely pedaling slowly with high force but, rather, pedaling at an RPM lower than the individual's Goldilocks RPM. The consequences of mashing in this context being:

1) By definition, sub-optimal power generation and;

2) Pedaling at an unnecessarily high pedal force for a particular power level.

One doesn't normally ride targeting a particular magnitude of pedal force for the sake of knee health. However, when one is pedaling sub-Goldilocks (mashing), they can certainly increase power output while lowering pedal force. And that's clearly a good thing for knee health regardless of the magnitude of the Goldilocks pedaling force.

As a hypothetical, imagine Jane cycling at her Goldilocks cadence of 90 RPM yet pedaling with a massive amount of pedal force compared to average folks. Is Jane mashing? I would say not. Rather, I would say that Jane is merely a very strong rider. Is Jane taxing her knees similarly to a 40 RPM rider pedaling at the same level of pedal force? I suspect so although I'm sure there are some physiological nuances to it that I don't understand. Most importantly, is there a cadence other than 90 RPM at which Jane could generate the same or better power out put at a lower pedal force? I think not.

So yeah, of the ride parameters that riders actually manage, power is the parameter of greatest interest for knee health. Or so it currently seems to me.

There are two ANT+ standards used by power meters: the PWR standard transmits power and cadence. The CTQ standard transmits crank torque and cadence. With the CTQ standard, the head unit knows the crank length.

In addition, some power meters simultaneously transmit on a non-ANT+ mode. Typically, this is a high frequency mode at 64 Hz, and it transmits crank torque and sometimes (but not always) crank position.

Typically, riders modulate their power far more with crank torque (or pedal force) than they do with cadence. Here is a plot taken from a ~five hour ~160 km ride, that included some flats, a couple of rolling hills, and one big climb (and descent), showing power, cadence, and crank torque in Nm. (Each dot is a one-second observation; the ride was ~5 hours long so there are ~18000 dots in each panel). As you may be able to see, the linear correlation between power and crank torque is much higher than between power and cadence. While it's true that cadence should be interpreted in the context of crank torque (or pedal force), the correlation between power and crank torque is so strong that you can basically just look at power instead of crank torque. Indeed, because cadence varies much less than power or crank torque, it's a poor candidate for a "control variable." Cadence is a red herring.


https://cimg3.ibsrv.net/gimg/bikefor...95a274af48.png

I was responding to someone claiming I lack evidence for determining my "optimum" cadence. Even if that study suggests that there's such a thing, it says absolutely nothing about applying it on an individual level.

I do know I don't fall into the range of the sample on age or frequency of exercise. I'm quite a bit higher on both of those than the highest test subject.

This makes sense as most riders will operate within their preferred cadence window for most of the time, changing gear as appropriate to manage their crank torque. So for example they may ride at a cadence of say 90 rpm over a wide power range, say 100-300W. You can see that in your plots as a vertical cadence band spanning all power levels.

It was interesting to watch Joss Lowden's hour record breaking ride yesterday. Cadence came up in the commentary. She chose to ride at a target cadence of just over 90 rpm, while most previous record attempts have been at 100+ rpm. So she chose a "relatively" low cadence by these standards. Her reasoning was to keep her HR slightly lower i.e. less cardio stress and more muscular load. It certainly worked for her in this case!

Watch from 32:30 for the discussion about cadence. Quite on topic.

Thanks for posting these plots. One thing you can note in the rpm vs ct plot is that there is a large oval with a high density of low to middle level torque with a median rpm around 90 and a second oval with a high density of medium to medium high torque with a median value around 70 rpm. Might the first one be mostly flats and the second one be mostly climbing?

Otto

You know that the pedal wouldn't move in these scenarios right?

This is simple physics, the pedal would not move since it is in equilibrium (the same weights at the same distance from the center of rotation)

I don't have time to watch the video, but I looked it up--the gearing was 64X15. 90 rpm on such a big gear for an hour! Now that's an outlier!

Exactly. Any force produced by the muscles would go into spinning the cranks (moving the bike). Because the two sides are in equilibrium, no force is required to raise the opposite leg.

It's a very big gear, but the aero drag is super low. Way lower than a normal rider. So road speed is very high (30 mph) relative to the power output. This is not a gear she would ever use on the road.

That's ridiculous. There has to be a disequilibrium in order for the leg to ascend, which requires that some force be put into the system, either by lifting the ascending leg with its own muscle or by pushing it up through the efforts of the descending leg (or more likely, some combination of both). Either of those will require the expenditure of some energy through muscle, and as pointed out above, the body is inherently inefficient in that it requires the expenditure of more energy to create the force than is actually applied. Peter's point above is fundamentally correct, that some of the energy put into the system gets released back as the ascending leg descends (it was stored as potential energy), but there is definitely energy lost in the creation and transfer. This amount of energy may be relatively small, but when repeated many times a minute, cumulates.

You're basically arguing that if two items are balanced on a scale, no force is needed to throw them off balance. While the force needed to do so may be small, it is not zero, and the more you want them off-balance, the stronger that force needs to be. Also, the heavier the items, the more force needs to be applied to increase to lift them. Imbalancing a pair of 1 ton weights by one ounce will not cause the lighter side to rise by as much as if you imbalance a pair of 8 oz. weights by the same amount. Heavier legs are going to require more energy to be lifted than lighter ones, there is no perpetual motion magic introduced because of balance between the sides.

..

Yup. There's a negative relationship between cadence and gradient, so as the road got steeper cadence dropped and crank torque increased. Because I have speed and cadence, we can also calculate gear ratios.

I disagree with the conclusions that you've drawn from your, very interesting, graphs:

1) When I look at the graphs, I see a rider deliberately holding cadence to roughly constant, preferred values which forces power to vary linearly in relation to crank torque. As such, I don't feel that power winding up being linearly related to crank torque says much of anything about the value of RPM data since the rider essentially contrives that outcome.

2) I don't feel correlation is particularly meaningful in this context. As you likely know, correlation is a statistical tool useful where a clear mathematical relation between the variables is not well understood. It's great for stuff like stock performances etc. With power meter data, the relationship between the variables is well understood and dirt simple.

3) To an extent, I think that cadence having a non-linear, steeply sloping trend speaks to the importance of cadence. It suggests that being in the right cadence zone has a larger impact on power generation than does crank torque. This is complicated, of course, by the rider possibly intentionally targeting certain cadences rather than naturally floating towards cadences that are the most efficient.