Motion Based Question

[Veronica]

How "off" is their altitude data? I know I've seen references to MB over reporting climb.

Looking at some of Jo's comparisons in her journal between her Ciclosport and her Garmin, it looks like Garmin can be off by as much as 45%. That's a huge difference!

Veronica

[maillotpois]

It's not really a Garmin issue, because the barometric adjusted Garmin data run through Garmin training center is pretty spot on - consistent with my old HAC4 data for the same rides and with other people's Polars, etc.

The problem is putting the data into MB without checking the new adjust feature in MB (I don't know what the new button is but there is now some option to allow MB to track more closely with the barometric data somehow. I don't use it so I don't know what it's called. We have a teammate who would always log our training rides and then broadcast his MB data and it would be at least 15% higher - really depended on the ride but it was always WAY higher. When he checked the adjust button on MB it went a lot closer to the GTC damat, but still maybe 5% high.)

[jobob]

I'm seeing that discrepancy between the elevation data from Garmin Edge 305 and my Ciclosport when I only run the Garmin data thru Training Center.

Through MotionBased, bah, forget about it! (unless you want to impress yourself with how much you really don't climb :rolleyes: ) I didn't know there was a way to correct it though - I'll look for that adjust button.

The data on my Ciclosport tends to agree moderately well with Bikely mapping, and if anything might be a bit on the low side (maybe 10% lower than elevation gains from Polar). So as long as it still works, I'll take elevation gain off my Ciclosport.

[OakLeaf]

I just went looking for it again. Up at the top right of your Activity Dashboard there's an icon called "Activity Options." Click that. Then for a Garmin Edge with barometric altimeter, check "Auto-Correct GPS Signal Errors" and uncheck "Use MB Gravity Elevation Correction."

I can't attest to the accuracy of that vs. SportTracks correction vs. Training Center, though.

[jobob]

Cool, thanks, I'll give it a try.

[SadieKate]

Based on the statistical average of 1 ride (today's drop run in which I cleaned rock gardens I never had*):

SW/Ascent/Descent

SportTracks / 1,569 / 4,118 (using data smoothing)
GTC / 1,634 / 4,182
MB with both autocorrection and MB gravity / 2,029 / 4,614
MB with auto correction / 2,195 / 4,747
MB with all elevation corrections disabled / 2,263 / 4,815 (check mark cleared in auto correction and grayed out in MB gravity so that the page shows "corrections disabled)

I think I'll stick with GTC and SportTracks (for my Edge 305).

*just had to get that in there! :D

[SadieKate]

Also, GTC and SportTracks report a max speed of 21.5 and 21.1. respectively. MB reports 27.2. Bwahahahaha!

[OakLeaf]

MB specifically says to DISable the Gravity correction when you're using Edge 305/705 (with barometric altimeter). What happens when you enable the autocorrection but disable Gravity (looks like the one thing you didn't try)?

Also, what level of smoothing do you use in SportTracks? I have mine off for elevation, off for cadence, on at 3 (of 150) for speed, on at 10 for HR. I think I'm going to increase the HR smoothing, but it seems to work pretty well for speed - just eliminates the obvious spikes and not the momentary fast descents.

SportTracks worked so well to correct distance when I was having that software bug, that I'm inclined to trust its elevation data as well (with smoothing off, especially important in my terrain of short, steep hills).

[SadieKate]

MB specifically says to DISable the Gravity correction when you're using Edge 305/705 (with barometric altimeter). What happens when you enable the autocorrection but disable Gravity (looks like the one thing you didn't try)?Yes, I did. It was just mis-typed list. See above. I verified again.

[SadieKate]

Here's a ride jobob can use for comparisons and demonstrates the erratic nature of MB:

Crater Lake Century
SportTracks / 6,222 / 6,230 (using data smoothing)
GTC / 6,218 / 6,225
MB with both autocorrection and MB gravity / 8,053 / 8,053:eek:
MB with auto correction / 6,490 / 6,497
MB with all elevation corrections disabled / 6,567 / 6,573 (check mark cleared in auto correction and grayed out in MB gravity so that the page shows "corrections disabled)

[Mr. Bloom]

I just went looking for it again. Up at the top right of your Activity Dashboard there's an icon called "Activity Options." Click that. Then for a Garmin Edge with barometric altimeter, check "Auto-Correct GPS Signal Errors" and uncheck "Use MB Gravity Elevation Correction."

I can't attest to the accuracy of that vs. SportTracks correction vs. Training Center, though.

I made the change too and this brought the climbing down to what I had last year on my Hilly Hundred route.

But, this is still 15% higher than the GeoBike Profile provided by the club.

HOWEVER, when I compare the MB elevation profile to the GeoBike Profile, the peaks and valleys track with precise accuracy.

So, this seems to indicate to me that the elevation profile is accurate, but the cumulative count is off in some way.

Does this make sense?

[OakLeaf]

How did the club obtain their data?

[TrekTheKaty]

I have a Garmin Forerunner 305. I've never paid much attention to the altitude until my century. I got a 6800 ft altitude change, and saw others as low as 5800 ft altitude change on the same ride. I have the autocorrect on because it is recommended. When I turned on elevation correction, it lowered it to 6300 ft.

I'm also curious what other do about this. How do I know what's accurate?

[SadieKate]

Mr. Silver, your results absolutely make sense. It's all on the algorithms, and how frequently the device is recording and a number of other whizbang mathematical stuff that confuses me. You'll see the same problem when comparing to many websites that are just using GPS data. So read onward . . . .

cunninghamair, barometric altimeters must be used to get readings at all accurate. The Forerunner 305 does not have one. Do some searches on this forum and over on bikejournal for more discussion.

I've used Polar and Ciclo which also have barometric altimeters with very consistent results with my Edge 305 run through GTC and now SportTracks. So, I'm comfortable using those numbers. Accuracy is in the eye of the beholder, it's pretty unanimous that a barometric altimeter is required to get close. Just remember that weather patterns can cause some inconsistency, but generally not as much as GPS-only data.

[OakLeaf]

The Forerunners do not have the barometric altimeter, so for those, MotionBased recommends that you turn on the Gravity correction.

We went for a hike in the woods this afternoon. I was hoping the slower speeds of hiking might bring all the results closer - not so. I brought my Edge 705 with me (partly for the barometric altimeter, partly because of the receiver that's much more sensitive than my Forerunner 301). Even so, with the tree cover, the reception was poor. So everything's going to be off - but FWIW, the data displayed in the unit called it 5.1 miles with 727 feet of elevation gain. GTC calls it 5.1 miles with 863' of climbing (and 859' of descent to the start point, so pretty close on the net anyhow). SportTracks calls it 5.98 miles with 1053' climbing (but only 956' descent!). MB, with autocorrect on and Gravity off, calls it 5.94 miles and +1076/-1050.

As far as speed smoothing, GTC had a max speed of 4.5 mph (probably about right); ST says max speed of 19.1 mph! :D (I hate to turn the smoothing up because it honestly seems pretty accurate on the bike - wish it would let you set smoothing differently for different sports). MB didn't smooth that spike out too well either - it gives me a max speed of 17.2.

What's the most accurate? :confused: Who knows - only thing I know for sure is that the max speeds on ST and MB are in need of some serious smoothing. :p

EDIT: I realized that I still had my 705 turned to smart recording. Probably for a hike, especially under tree cover, it would've been a lot better to set it for 1 second recording. Doh.

[Mr. Bloom]

How did the club obtain their data?

I'm presuming GeoBike is a different vendor, but it does raise the question of whether we're comparing equally errant sources:confused::confused:

[Veronica]

I've seen rides where riders are asked to submit their data from Polars, GPS units, whatever... I assume the ride organizer looks at all the numbers and makes some determination.

Did you see this for the Hilly Hundred?

http://www.hillyhundred.org/PDF_Files/Hills.pdf

Those aren't hills; most of those are rollers. :p

Remember max incline could be from one spot on the hill, not the entire hill. It would be like saying our Mount Diablo is a ten mile 18% climb, just because the last tenth of a mile is 18%.

Veronica

[SadieKate]

EDIT: I realized that I still had my 705 turned to smart recording. Probably for a hike, especially under tree cover, it would've been a lot better to set it for 1 second recording. Doh.Doh! :D

Anytime I have satellite issues, all bets are off for anything to interpret the data.

[SadieKate]

Did you see this for the Hilly Hundred?

http://www.hillyhundred.org/PDF_Files/Hills.pdf

Those aren't hills; most of those are rollers. :pHey, Ms Teacher. Did you do the math? How do you use your momentum to "roll" over a grade that is well over three times the length of a football field. Nothing is shorter than .2 miles or 1,056 ft or 352 yards or 3.52 football fields.:p

[Veronica]

Hey, Ms Teacher. Did you do the math? How do you use your momentum to "roll" over a grade that is well over three times the length of a football field. Nothing is shorter than .2 miles or 1,056 ft or 352 yards or 3.52 football fields.:p


No, I did not do the math. :rolleyes:

It's all in you definition of hill. If it takes me just a few minutes to get over it, it's not a hill. My 27 mile ride with 1,280 feet of climbing gets labeled rollers, not hills because none of the hills take very long to get over. Not even the ones that require me to go into my small ring.

Veronica

[OakLeaf]

okay - we've wrangled over what's a "hill" before (I think probably NOBODY except for Alpinerabbit would claim they ride "mountains" - don't we have anyone in the Rockies?)

But now the gauntlet's laid down.

V., come over here and ride the Columbus Fall Challenge next year (provided they hold it - they've had organization trouble the last couple of years). I'll train and do it with you. Similar type of terrain to the Hilly, but about 1.5x more climbing per mile, and 110+/- miles each day. That's a ride I was afraid to do when I was racing and I'm still a little afraid now that I'm braver but not nearly as strong. But I challenge you to do that ride next year and THEN say they're not hills.

[Veronica]

Sorry, I can't travel to Columbus Ohio in the middle of the school year.

You could come do DMD with me in April.

http://www.caltriplecrown.com/images/DMProfile.jpg

I'm hoping to finish this time. In 2006 I got to the top of Mt. Hamilton and my lower back quit. It's a lot stronger now.

In this profile - Diablo and Hamilton are mountains, all the rest are hills. Of course others opinions will vary. I do label this ride steep at BJ because of the nature of the hills.

Veronica

[OakLeaf]

:eek: April? 105-mile-a-day, nice and flat TOSRV is too much for May for those who don't winter in the South!

Girl, I never said you aren't strong!!! Don't ever think I said that! I don't doubt you could do CFC in your sleep.

Just don't dis our terrain or the riders here. Like my one riding buddy who did the Cinderella Century this year on a rented, poorly fitting mountain bike (the last bike she could get on short notice). And passed a whole lot of people while she was at it - she tells a story about how her knobby tires were making so much noise, people would yell "Car Back!" before she passed them. :p All of her training was here in the heartland, except for Mountain Mama, a little bit south of us - looks to be at least as many feet/mile as DMD, if half the distance - lots of people from our area did Mt. Mama, with no training on what you would consider "hills," and they finished it just fine (including one guy who broke his wrist early on and finished the ride anyway, but he's just nuts :rolleyes:).

[Veronica]

Those aren't hills; most of those are rollers. :p

Veronica


Do people not notice Smiley Faces? I wasn't dissin anyone's terrain.

It's very different from where I typically ride. I'm always surprised at how high the reported average speeds are for these rides in the Midwest with the amount of climbing that they have. Since it's not the sort of stuff I ride, I'm trying to reconcile the numbers in my brain. So here's my current thinkin'...

On a climb like Diablo, I'll average between 7 and 8 miles an hour for ten miles. The descent doesn't allow me to make up much speed because it's so twisty. I'll average around twenty, maybe twenty-two if I'm lucky and feeling brave, on the descent.

In that link I posted there's only 6 miles of climbing on Saturday's route. I assume that same 7 - 8 mile an hour climbing pace, although because the climbs are shorter, a rider could climb faster. I also assume that the downhills are less twisty. I haven't been there, so I could be wrong :D But if that is true, a rider would get more speed on the downhills and maybe average closer to 30 mph. That would all lead to a much higher average speed than I get on a typical ride with 3,500 feet of climbing.

Like I said it's very different from where I ride.

Veronica

[Veronica]

Funny... I turned off my computer and walked away and was thinking about my post....

That link does only list 6 miles of climbing. But the profile says 3500 feet of climbing. So either that link does NOT list all the climbing or those 6 miles are WICKED steep. :eek:

Now think about this, if you hadn't factored in the 3,500 feet of climbing...

Six miles of climbing spread out in a in a 48 mile ride, with the longest climb being 1.3 miles, wouldn't you think the ride was kind of easy...

Maybe you would, maybe you wouldn't - all depends on who you are and where you're at. :) Bottom line - Ride Your Own Ride and have FUN!


Veronica

[SadieKate]

V, you know this: the roads as you travel farther east are built more and more vertically up the fall line and not the contour line as they do out here. Those river bluffs in the midwest are wicked steep and, as you know, it is hard to get into your climbing zone on "rollers" and those are long enough to actually force you into climbing mode as opposed to "momentum" mode.

This is what a friend has said about riding this stuff:
The "mere rollers" are constant, but after 100k of continuous, short 6-7% climb/descent cycles most people get a little rubbery in the leg.

Yeah, it is different than here in the wild, wild West. :p

Oakleaf, I don't live in Switzerland, but I've got a 23 mile climb nearly right from my door. The climb up Mt Diablo (on V's profile) rises the same elevation but in half the distance. I'd call both mountains. But I'm not arguing what is a hill or mountain, I'd just saying you can't "roll" over something the length of 3 and a half football fields in length. Your "mo" won't last that long. Though, wouldn't perpetual mo on a bike be wunnerful?

[maillotpois]

I'd agree on DMD Hamilton and diablo are the only mountains - but it that just a nomenclature issue? :rolleyes: It is Mt. Hamilton and mt. Diablo after all, so maybe it's just what we're used to.

I love the rollers, even the long ones. And I love our sustained climbs. I'm not sure how well I'd do in the eastern part with the steeper shorter stuff. It would be hard to get in a rhythm with that.

And from what I've seen and ridden of the Alps/Pyrenees, they're more like our Sierra roads with reasonable to high gradients; however, they pitch you up between the switchback instead of in the switchback like we do here, so you can often get more climbing overall. It was weird to have the actual switchback be the flat part where you can recover because it's so often not the case here.

Back to the original issue - and I'm not sure if anyone noted this yet, but weather can affect the barometric corrections. I spent a morning climbing the geysers and I and the other 5 people with Garmins all showed we were steadily descending. It was VERY odd, but the weather was changing dramatically from rainy/cold to warm.

[jobob]

When I start & end my ride at the same place, usually the total ascent/descent values (and the start/finish altitude) agree within a few percent, but once in a while they're very different. I assume that's because of weather changes over the course of my ride.

[TrekTheKaty]

Thanks, Veronica. This was very educational for us newbies! (Going to search for "hill":) )

[maillotpois]

When I start & end my ride at the same place, usually the total ascent/descent values (and the start/finish altitude) agree within a few percent, but once in a while they're very different. I assume that's because of weather changes over the course of my ride.

I get that a LOT if I start from home. I concluded that it's because it is hard to get a good satellite lock from my house in all the trees, and even though it "says" it has the satellites I think it really takes some time for them to lock on. (My car Garmin won't even try to figure out where we are til we've got all the way down to town.)

Your reason makes a lot of sense as well. I'm sure there are a lot of variables.

[SadieKate]

And from what I've seen and ridden of the Alps/Pyrenees, they're more like our Sierra roads with reasonable to high gradients; however, they pitch you up between the switchback instead of in the switchback like we do here, so you can often get more climbing overall. It was weird to have the actual switchback be the flat part where you can recover because it's so often not the case here.Next time you're here and the trail is cleared, we'll ride the Flagline Tie singletrack. At the very end when your heart rate is highest the trail suddenly goes up the fall line and then the switchbacks and the trail between keep going up and up . . . and over roots and rocks. But then you get another 1000 ft above the South Fork drop. Yum.:D


Back to the original issue - and I'm not sure if anyone noted this yet, but weather can affect the barometric corrections. I spent a morning climbing the geysers and I and the other 5 people with Garmins all showed we were steadily descending. It was VERY odd, but the weather was changing dramatically from rainy/cold to warm.Yup. But you've reported the weirdest case I've ever heard. Usually, it is just greater discrepancy between the climb/descent totals.
Just remember that weather patterns can cause some inconsistency, but generally not as much as GPS-only data.

[maillotpois]

Next time you're here and the trail is cleared, we'll ride the Flagline Tie singletrack. At the very end when your heart rate is highest the trail suddenly goes up the fall line and then the switchbacks and the trail between keep going up and up . . . and over roots and rocks. But then you get another 1000 ft above the South Fork drop. Yum.:D


Is it worse than Green Gulch? No thank you... :p

[SadieKate]

Much better traction and smoother over all. Of course, you top out at 7,000 ft but the grade has shallowed out by then. It is so worth it.

[SouthernBelle]

I couldn't figure out for the longest time why there would be a 2-3 feet difference in ascent and descent on some of my rides. I finally realized it was a matter of where in my driveway I hit the start and stop button. :p Slight climb out of my drive.

[OakLeaf]

Oh, none of them are THAT accurate. If you're within 2-3 METERS you're doing good. Usually I'm within a 10' range when I ride from the local Y and start my computer in their parking lot (pancake flat).

[jobob]

Agreed!

[kfergos]

Geologically a hill is (usually) anything less than 1,000 from base to and a mountain is anything more than 1,000 feet from base to top. The problem is perception -- short, steep, climbs can be killer compared to those long, long inclines.

I like my Garmin Edge 605 -- it says I can ride up a 45% grade! :eek: For those of us sans barometric measurement devices, what is the most reliable way to find the elevation profile of a route? Is mapmyride pretty accurate?

[SadieKate]

Oh, your elevation profile should be fine without a barometic altimeter. What goes astray is the cumulative climb totals as in "I climbed 3,212 ft in 100 miles." A GPS should give you a decent recording that the base of the hill was at X feet and the top was at x feet of elevation. But if the road has any undulations on the way up (which just about all do), the total climbing will be off.

Make sense?

[OakLeaf]

And another Doh! from my commute on Thursday.

I didn't reset the computer between legs. In the hour and a half I was at the gym, apparently the barometric pressure must've dropped a bit. So my whole return trip was 200 feet "higher" than the outbound leg.

Because it was all one trip, none of the correction algorithms (TC, ST or MB autocorrect) was "smart" enough to correct the elevation data throughout the trip. When I apply MB Gravity, that one is that smart - so my elevation profile is a nice mirror image as it should be for an out-and-back - but it gives me a total elevation gain that's like 40% greater than the autocorrect alone. :confused: Hmph.