Login required to started new threads

Login required to post replies

Prev Next
Re: How to verify & validate running power [milesthedog] [ In reply to ]
Quote | Reply
Great analogy and very applicable to the testing I have done with Stryd to date. The power will vary depending on your form efficiency. So, on a treadmill using good form you could get X, but if you fatigue and form breaks down at the same pace, you would get Y...I have seen this quite consistently.

Ideally, yes, a power meter could detect shoe differences, which is something I have not seen with Stryd, but I have detected differences in Leg stiffness/Form power by changing shoes.

As per testing, you can't rely on a treadmill for validation using multiple runners, unless they are all at the same weight and the treadmill is calibrated after each use. The notion of a treadmill with force plates is ideal.

Having been testing it for a while, I have seen mostly positives. It may not provide an accurate measure, but it has proven to be consistent and reliable in its readings. Compare it to a power meter that is off by 50 watts. If it is off by 50 watts across the board, it doesn't really matter if the training zones are adjusted for such a value. It may not provide an accurate reading (+50) but it is still a valuable tool.

I wonder if we go back 10-15 years and search the early bike power conversations if they look eerily similar to this one?
Quote Reply
Re: How to verify & validate running power (Stryd vs Garmin) [stumpyx13] [ In reply to ]
Quote | Reply
See my post directly prior to yours. Stryd does capture that variation. And, having done a decent amount of biomechanic analysis and research, you’d be surprised, but there really isn’t a huge deviation between the way people run. I mean, you don’t see one person with a glut-sized hip flexor and a hip flexor sized quad - our isomorphic muscle composition does limit biomechanic variation. We are just hyper sensitive to seeing variation within our own “in group” (note, I don’t say species because we often have difficulty seeing the same differences across race and gender).

And, the vast majority of adults from industrialized nations run in a very similar way: push, swing the leg through from the hip with little knee bend, and then reach out with the lower leg to heel strike - really not much variation there. And most triathletes, even pros, run this way or a hybrid between this and running the way children run and elite runners run: foot off the ground relatively soon after landing, significant knee bend relative to pace, foot lands close to under the hips (none of that swinging from the hips motion seen in the prior example). I can run both ways or a hybrid and my power changes and corresponds to my heart rate and perceived effort, and having done a fair amount of work hooked up to a gas cart, I’m confident my metbolic cost would change correspondingly.

wovebike.com | Wove on instagram
Last edited by: milesthedog: Oct 6, 17 5:11
Quote Reply
Re: How to verify & validate running power (Stryd vs Garmin) [Andrew Coggan] [ In reply to ]
Quote | Reply
Andrew Coggan wrote:
stumpyx13 wrote:
, I'm saying that I wouldn't trust the dimensional number that comes out of the thing due to the variance among the population (just seeing how many different ways there are to run, I don't think the sample size would be large enough to capture this)


??

That's like saying you don't trust gas mileage measurements because some cars are more economical than others.

In the present context, the equation for calculating mass-specific power is acceleration times speed. That equation applies regardless of how someone runs, just as calculating gas mileage by dividing distance covered by fuel consumed applies to all ICE vehicles regardless of their fuel economy.

I think the better analogy is that I wouldn't trust something to predict my car's gas mileage if I just gave them a few variables like engine size and compression ratio. Sure I might get a decent prediction, but some assumption would have to be made about heat loss, etc. In your example I could have a some data saying that based on these inputs, I get this gas mileage and get a regression to predict gas mileage for other cars based on these same inputs, but this new prediction will have variance and not be exact.

And I get that the basic equation is set, but you have to "translate" foot motion to center of mass acceleration. This requires some type of model or assumption based on how your foot is moving. If this model is tuned based on gas analysis and calorie consumption, then the efficiency of caloric power to mechanical power needs to be known or calculated. This again requires some type of model, this time tuned via a statistical study most likely. No matter what, it's saying that there is input data which consists of foot motion and I want to a model to output power, and this will always have some variance. Maybe they do have enough data to show that the true variance is small but I don't think that'll ever be published.

I like analyzing things - http://engineeringfitness.org
Quote Reply
Re: How to verify & validate running power (Stryd vs Garmin) [milesthedog] [ In reply to ]
Quote | Reply
milesthedog wrote:
See my post directly prior to yours. Stryd does capture that variation. And, having done a decent amount of biomechanic analysis and research, you’d be surprised, but there really isn’t a huge deviation between the way people run. I mean, you don’t see one person with a glut-sized hip flexor and a hip flexor sized quad - our isomorphic muscle composition does limit biomechanic variation. We are just hyper sensitive to seeing variation within our own “in group” (note, I don’t say species because we often have difficulty seeing the same differences across race and gender).

So that's my ignorance for not knowing this. Perhaps running motion variance is low and MSE is low for the regression from foot motion to some notion of power. In that case, I'd trust the number more but still hold skepticism until I actually see the data used (which will probably never happen).

I like analyzing things - http://engineeringfitness.org
Quote Reply
Re: How to verify & validate running power (Stryd vs Garmin) [stumpyx13] [ In reply to ]
Quote | Reply
 It swinging the pendulum of reason back in your direction, that low rate of error has gigantic association with rage of injury. So when I purposefully change my mechanics and see huge swings in power from Stryd, that tells me their model is converging on the actual biomechanics and capturing that injury-inducing low variance.

side track: The ultimate proof of a good model, in my opinion, will be to also create a stick figure model (side view) that can show the runner in real time with movement of both legs from a single foot pod. I’ve studied this a good bit and while Ive dropped the ball this past six months in communicating with folks interested in carrying this out, I think it is possible, because in addition to isomorphic reasons for low variance in movements, related, movements (assymetric and all) of one leg are highly predictable based on movements of the other leg.

wovebike.com | Wove on instagram
Quote Reply

Prev Next