Question Re: Accuracy - GPS vs Accelerometry (?)

morpheus

I searched, found the thread on the NYT article.  Didn't bother reading it cuz most of you suggest that I go review DCR's take on it.  (and I'm doing that as I type here)

But, what is the general consensus on which is better for accurate pacing?

I ask because my wife bought me a Garmin FR70 + footpod for Father's Day.  I didn't have the 310xt at the top of my list cuz I just got the Joule bugs worked out and I didn't think that I need redundant gear.   Now, if the GPS is the preferred method, then I can return the FR70 and footpod, and for a few extra bucks (~$75), pick up the 310xt. w/out HR. 

What would you do and why?

Tom Glynn

Read the DCR article on the foot pod. It was very interesting, basically stating it was pretty much as accurate as the GPS. IIRC, he tested it on a straight course and the results of GPS, foot pod and known distance and all were very close.

He also linked to a piece of free software that allowed you to pull data from your device and compare GPS data to footpad data. I did that and found that data did tend to bounce around a bit, all in all, the two methods of distance measurement were pretty accurate.

As a side note, I now also believe my home treadmill tends to slow down a bit over the course of a workout. When I run on a very expensive gym or hotel treadmill my speed as measured by the footpod is extremely close to the treadmill and if I keep the same speed on the treadmill, my footpod speed stays the same. At home, I find that after 20-30 minutes at the same speed on the treadmill, my footpod tells me I'm slowing down. After reading the DCR's article and testing the footpod, I believe my footpod data over my treadmill.

morpheus

Re: "...as accurate as the GPS."

so, GPS wins, I think.

Tom Glynn

GPS in a straight line works! The more twists and turns, the less accurate it is. Go to a track and run a couple miles and you will see.

Either way, these devices are all pretty accurate. Go run, have fun and don't stress about it!

Steve West

Another point about GPS is that the instantaneous rate (Current Pace) is rarely correct due to the nature of the physics behind the system. If you use Lap Pace, once you travel more than about 0.25 miles (on a relatively straight course) it becomes reasonably accurate.

Timing errors of 0.000,001 seconds (1 microsecond) correspond to a 300 meter error in distance from the satalite with the degraded signal. There are a number of causes of degrded signals that can vary greatly such as atmoshperic, ionospheric, or local multipath conditions. Current Pace is calulated from your current location relative to your last know location. If either of those is off by evan a few feet, your pace will be wildly wrong.

I have my 310 set to auto lap at 1 mile and vibrate every 0.5 miles. The distance alert lets me know when I can trust the pace, at least until the next auto lap.

One of these days I am going to write up a primer on GPS and how it works, or not, for us in the fitness world....

morpheus

So, in layman's terms...

Elevation, switchbacks, tight turns, and ionospheric/asteroid storms can wreck my 'recorded' workout data, right? @steve

I have track access but run alot on tree covered rolling hills in addition to some trail running. Sounds like the footpod might be best for me if I want to control pace.

morpheus

Wait, looks like it's not uncommon to add the footpod to the 310xt?

That way you can use the footpod for real time pace and get the mapping from the GPS?

[Deleted User]

I have the 910xt with a footpod. There was a recent update to fix the problems with GPS instantaneous pace, but I tried it this morning and it still bounces around a lot. I don't trust it for that purpose. I use the footpod for instantaneous pace - once calibrated it is right on. I calibrate at threshold pace so it is most accurate at that pace, but it is still accurate enough at other paces. Plus you can use it on an indoor track (or any track since GPS does not handle those tight turns well). I rarely run on a treadmill, but I would not be surprised if your running mechanics change somewhat on the treadmill so that a footpod calibrated on the road would not be quite as accurate on a treadmill.

[Deleted User]

@Chris - that is what I do with the 910xt and it works great. Not sure why more people apparently don't do that!

John Withrow

@Chris-- I use the 310xt and use the foot pod. But I use the footpod mostly so I can focus on keeping a high cadence... I really like having the piece of mind of having a backup for my PT computer during a race. My Garmin Edge 500 died during a race once and it was nice to have a backup to at least have power and cadence for the rest of it.

[Deleted User]

I think a foot pod provides very good accuracy provided that you calibrated properly, and are consistent in your running mechanics and such. I do notice sometimes that the accuracy of accelerometer-based systems tend to be off when running on different terrain, such as grass vs. pavement.

morpheus

Right, I assume it learns my regular stride and deviating from that could cause some issues.

Most of my paced interval workouts are on roads (and a track) where my stride shouldn't change much. Not sure how bad it'll screw up the trail runs, but I do those more for volume and zen stuff as opposed to 'hit the intervals'. So, if accelerometry gets screwed up on crazy terrain like trails, and GPS gets screwed up on switchbacks/elevation of same trails, then I'm just throwing the trail factor out of it.

Peter Greagg

I use an xt310 with footpod, with the footpod for cadance monitoring.
I do my Z4 and 5 intervals on a main road with a very wide cycle lane (where I run).
I use lap pace to help me dial in my pacing and do half of the interval in one direction and then turn around and retrace my steps so the small gradient of gthe road isn't an issue.
It seems to work fine for me.

Dino Sarti

I have a foot pod and use it with my 610 and 910 depending on what's up, but I really only use it to monitor cadence. On long runs I am watching cadence and pace.

How does one calibrate it and use it to over ride GPS for pace and distance?

Tom Glynn

@Dino,

calibration is done at a track over a two lap course. There is a section in the manual on how to do it.

http://www8.garmin.com/manuals/Forerunner305_FootPodQuickReferenceGuide.pdf

I'd recommend running at threshold pace or the pace that you do most of your running at to make it as accurate as possible.

MattAaronson

My personal experience with the footpod is that the calibration varies with pace...i.e. if I caibrate it to a 6:30 pace then it's pretty good within 45 sec of that. But then when doing recovery intervals at 8:00 the calibration is way off. Probably that's because my stride and running mechanics change when I go from a z4 to a recovery pace.

I did not read the articles you're all talking about. But one thing is for sure, which is that accelerometers can be incredibly accurate and precise. Among the many applications where very accurate and precise accelerometers are used is internal navigation systems in airplanes. Of course the accelerometers in a Garmin footpod are not exactly the bleeding edge of the technology. But then again, nor is the GPS that's in a consumer's Garmin watch.........

Al Truscott

I suspect the "accelerometer" simply makes an assumption about stride length, and uses that plus the frequency of strides to determine speed. Calibration just equates the number of strides taken to cover a known distance, and uses that to deteremine the stride length. My experience is the same - the calibration works great for the speed at which is was originally calculated, but will become increasingly inaccurate when going faster or slower. My Garmin 305 allows me to manually set a calibration number, which I could do if I wanted for different speed workouts, but since *every* workout includes intervals of varying speeds, why bother...

All this relates to trreadmill use, which is the only place I ever use the footpod.

morpheus

Based on what u guys have mentioned, I may calibrate it at/for z4/z3 work during the build. Then hit the track for another recal at a slower pace, like what will be used during the IM run.

[Deleted User]

Note that you don't have to use a track. I mapped a prefect mile on my regular running route on map my ride, and use that to tweak my calibration. I usually start my first mile repeat there, and tweak the calibration until it is right on. Once done it is so much better than the GPS instantaneous pace - for me anyway. I've also used the software mentioned earlier in this thread to get me in the right ballpark.

kimmydct

x2 on what Matt said about the pace you use the calibrate. I have found that as well.

William Jenks

You are all correct about needing to calibrate it near the pace that you intend to use the footpod.

On my older generation one, I had to re-calibrate it when I changed the position of it significantly on my shoe (e.g., after changing the battery or just putting it on another shoe even). This implies to me that the measurement is a little more complicated than just counting cadence, but I can't be sure.

With my more recent generation one, I have not noted quite as much difference when I've taken it off and re-mounted it.

mancona

I also found the foot pod calibration varies by pace. Also, my experience is that it will calibrate differently on different shoes and in different places on the shoe and therefore if you really want it to be accurate you should re calibrate every time you change shoes.

In the winter I calibrate on the shoes I will wear most of the time and use it on the treadmill so I have data.

In the summer, I use GPS for speed/distance but I like to keep the footpod so I have cadence.

Joseph Lombardi

 However; If you are talking about accuracy in a pool  then the accelerometer is more accurate than the GPS for lap swimming in a pool (just thought I would throw this in given Garmin introducing a Pool Swim specific device...very sweet).