Hopefully this will put most of the arguments to rest:
https://www.ncbi.nlm.nih.gov/pubmed/29781934
(Note that power and leg spring stiffness are calculated from the same raw accelerometer measurements used to determine GCT, etc.)
.
Do you know of any studies that measure the accuracy of the Garmin Running Dynamics sensors in either the Tri HR Monitor strap or the Running Dynamics pod?
It would be my guess that Garmin isn’t too far behind Stryd in figuring out the algorithm to determine run power using GCT, stride length, vertical ratio, and vertical oscillation
Notable: The study didn’t address or test the accuracy of power…the main thing people care about with Stryd.
It focused on all the other metrics.
As I noted, power is calculated from the same raw accelerometer data as everything else. It is hard to imagine how these other metrics could be reasonably precise and accurate, yet power be way off.
Now precisely which power is being calculated might be a different story…
Do you know of any studies that measure the accuracy of the Garmin Running Dynamics sensors in either the Tri HR Monitor strap or the Running Dynamics pod?
I have seen a few comparisons against Stryd, but not against a force plate (or mat…we just a runway at work for about $20k that is long enough to collect a handful of strides, thus avoiding possible targeting issues).
It would be my guess that Garmin isn’t too far behind Stryd in figuring out the algorithm to determine run power using GCT, stride length, vertical ratio, and vertical oscillation
Garmin started reporting power 6 months or more ago. Not surprising, really, since calculation of running power is rather trivial, at least if the input data are “clean”. A bigger issue, really, is deciding what power is most relevant to calculate.
Please correct me when (not if) I’m wrong:
the practical usefulness of a watt in cycling is really it’s strong correlation with a Kcal. so, if the “watt†being displayed by Stryd is also just as strongly correlated with a Kcal, how is that not all that really matters?
As I noted, power is calculated from the same raw accelerometer data as everything else. It is hard to imagine how these other metrics could be reasonably precise and accurate, yet power be way off.
Now precisely which power is being calculated might be a different story…
I’ve employed plenty of analysts who took perfectly good data and then effed it up in an algorithm. Good inputs do not automatically lead to good outputs.
I guess I look at it this way: there are more well-trained/highly-qualified PhDs working for/on Stryd than on other “wearable tech”, including anything developed/sold by Garmin, Apple, etc. The odds that they would stuff up a relatively simple calculation therefore seems pretty low to me.
Can you distill this paper down into something more readable for us mere mortals Andy?
Despite having a doc and a masters myself I struggle with this as it is outside my area of expertise. Downloaded the paper and had a good read but struggled.
Is the OptoGait the gold standard by which other systems are normally compared?
Have been tempted to try out the Stryd but not quite sure how it will work for me.
Cheers.
Posted this to the Stryd Community FB group earlier this week:
“Just an FYI: I am at the American College of Sports Medicine meetings at the moment. I saw a number of posters evaluating Stryd, mostly (IIRC) from Furman University. The executive summary is that Stryd power data are highly reproducible across days/surface/devices, and is closely correlated with VO2. This is of course expected based previous reports and Stryd’s own data, but independent confirmation is still nice to see.”
To anticipate Ray’s commentary: none of the posters I saw seem to have addressed the accuracy of the power data, only the precision/reproducibility and association with VO2. I assume that is because they were from smaller institutions that didn’t own a force plate (or better still, a $70k force plate treadmill, like Stryd built in-house). The way that the instrumented mats or walkways are catching on, though, I expect that will soon change.
Posted this to the Stryd Community FB group earlier this week:
“Just an FYI: I am at the American College of Sports Medicine meetings at the moment. I saw a number of posters evaluating Stryd, mostly (IIRC) from Furman University. The executive summary is that Stryd power data are highly reproducible across days/surface/devices, and is closely correlated with VO2. This is of course expected based previous reports and Stryd’s own data, but independent confirmation is still nice to see.”
To anticipate Ray’s commentary: none of the posters I saw seem to have addressed the accuracy of the power data, only the precision/reproducibility and association with VO2. I assume that is because they were from smaller institutions that didn’t own a force plate (or better still, a $70k force plate treadmill, like Stryd built in-house). The way that the instrumented mats or walkways are catching on, though, I expect that will soon change.
Let’s assume for sake of argument that it is accurate. How does this tool change run training? In the context of running, where this correlates to vO2, how is this better than just using a Daniels vdot table or another analogue for vO2?
Let’s assume for sake of argument that it is accurate. How does this tool change run training? In the context of running, where this correlates to vO2, how is this better than just using a Daniels vdot table or another analogue for vO2?
Purely as a layman triathlete, I really don’t understand this argument. An accurate and precise powermeter for running seems to give me the same benefit as one for cycling, i.e., I can gauge my level of effort independently of the grade I’m running on. There’s very little ‘flat ground’ around here - everything I run on is slightly uphill or downhill. Targeting x pace is often a fool’s errand, and gauging short intervals by HR is useless, so why wouldn’t I want to be pacing by power instead?
Let’s assume for sake of argument that it is accurate. How does this tool change run training? In the context of running, where this correlates to vO2, how is this better than just using a Daniels vdot table or another analogue for vO2?
Purely as a layman triathlete, I really don’t understand this argument. An accurate and precise powermeter for running seems to give me the same benefit as one for cycling, i.e., I can gauge my level of effort independently of the grade I’m running on. There’s very little ‘flat ground’ around here - everything I run on is slightly uphill or downhill. Targeting x pace is often a fool’s errand, and gauging short intervals by HR is useless, so why wouldn’t I want to be pacing by power instead?
Because for most people, training by power wouldn’t benefit them and could hurt them. For example, your hilly runs - Running downhill at the same wattage as uphill would mean you’re running faster. Unlike cycling though where you’re clipped in and making similar muscle movements, your quads would take the full force of your weight and if it were a race, you could easily bury yourself with muscle damage due to fast downhill sections.
AFAIK, these power meters also don’t take in to account wind differences like cycling PM’s do – something that could make a pretty big impact on pacing as well.
The only real place for them is for trail running and form evaluation / corrections.
In addition, power means nothing in terms of a running race where most people are concentrating on a specific time (and correlated pace). You call pace a “fool’s errand”, and as an experienced marathoner, I call pace the only metric that matters if you’re going for a clock time.
Given that power correlates to vO2, and in running it’s easy to run by feel or pace depending on course and conditions, what’s the gain in having a power meter? In cycling it’s obvious. What is it in running? I can go to a track and do intervals off pace perfectly. And in a race I am RACING, not trying to hit some power number. If I need to cover a surge, I will cover the surge if I feel I can. Or I can sit on a certain power and lose contact? WTF, no way.
I train completely by power on a bike. It works. In running no one has convinced me there is any benefit, and every single person I know who tried quit it fairly quickly for that reason.
AFAIK, these power meters also don’t take in to account wind differences like cycling PM’s do – something that could make a pretty big impact on pacing as well.
It takes nearly gale-force winds to alter the energy requirements of running by more than a few percent.
Given that power correlates to vO2, and in running it’s easy to run by feel or pace depending on course and conditions, what’s the gain in having a power meter? In cycling it’s obvious. What is it in running? I can go to a track and do intervals off pace perfectly. And in a race I am RACING, not trying to hit some power number. If I need to cover a surge, I will cover the surge if I feel I can. Or I can sit on a certain power and lose contact? WTF, no way.
The most common benefit people seem have found from running power is using it as a pacing tool in longer events. That is especially true in hilly races, but even in flat ones it can be helpful to have a real time, stable, and accurate indicator of your effort, something that, e.g., ordinary footpods or (worse) GPS often fail to provide.
AFAIK, these power meters also don’t take in to account wind differences like cycling PM’s do – something that could make a pretty big impact on pacing as well.
It takes nearly gale-force winds to alter the energy requirements of running by more than a few percent.
Not even close to true. Look at any track meet. Wind aided times in excess of 2m/s are considered to not be considered for record keeping for sprints. That’s about 4.5MPH.
It absolutely is true.
First, the counter-example you offer is essentially meaningless, since the standard being applied in choosing that cut-off could be a difference of, say, no more than 0.00001%.
Second, a wind speed of 2 m/s at ground level is actually a fairly strong wind.
Third, there are plenty of data in the scientific literature going back at least 50 y supporting my statement.
It absolutely is true.
First, the counter-example you offer is essentially meaningless, since the standard being applied in choosing that cut-off could be a difference of, say, no more than 0.00001%.
Second, a wind speed of 2 m/s at ground level is actually a fairly strong wind.
Third, there are plenty of data in the scientific literature going back at least 50 y supporting my statement.
You said Gale force winds. 4.5MPH isn’t Gale force. Having had 10 years of experience running track, it happens fairly often that that limit is exceeded especially in the spring. I’d rather believe the IAAFs rulings on what is wind aided alongside numerous accounts of 100 and 200m wind aided times which are statistically faster than non wind aided times.
It’s not hard to believe that a fairly open course running into a 5-10mph headwind would have performance suffer. I don’t need 50 years of studies to tell me that wind effects performance. Numerous Boston marathons have shown it does (depending on directions of the wind, either aide or hinder performance). If the stryd power meter doesn’t account for wind, then it’s not calculating real power usage in the same way a cycling power meter does.
I said nearly gale force, not gale force.
Since you brought up an exact value of 2 m/s, though, let me be a bit more specific:
Even a wind speed of 5 on the Beaufort scale (which on open terrain - an airport taxiway, for example - would correspond to a ground level wind speed of ~2.5 m/s) would only alter the energy requirement of running by a few percent.
This conclusion is based in part on studies in which the oxygen uptake of runners has been measured while they ran on a treadmill placed in a wind tunnel.
“In God we trust. Everyone else must bring data.” - W. Edwards Deming