Top

Accuracy of Suunto Ambit3 Peak at Okay/60-second GPS pings (200-hr)

Following my long-term review of the Suunto Ambit3 Peak, friend and fellow guide Alan Dixon sent me an email with the title, “Suunto 200 hr mode not cutting it,” and shared that on a recent 8.2-mile hike his Ambit3 shortchanged him by 1.5 miles (18 percent). He had set the GPS accuracy of the Ambit at “Okay” and was comparing it to the distance measured with his iPhone using the GaiaGPS app.

Alan’s experience is consistent with that of Ethan, a reader who recently commented that his Ambit3 Peak was also 20 percent short when set to “Okay,” versus the listed mileage in the guidebook and on trail signs. (Importantly, he found it to be 75 percent short when the auto-pause was on.)

I’ve never intentionally and precisely tested the accuracy of the Ambit3 Peak in the “Okay” setting, but the observations made by Alan and Ethan sound intuitively correct: the recorded distance will be short of the actual distance. The exact margin of error will vary; 20 percent is probably a safe expectation.

In this post I will explain the cause of this inaccuracy and offer guidance on best uses of the Ambit.

Context

The GPS accuracy of the Suunto Ambit3 Peak (and all other Suunto GPS watches, too, including all Ambit2, Ambit3, and Spartan models) can be set to:

  • “Best,” or 1-second pings;
  • “Good,” or 5-second pings; or,
  • “Okay,” or 60-second pings.

It takes energy to acquire a GPS location, so unfortunately the watch’s accuracy is inverse to its battery life. The Ambit3 Peak is rated to 20, 30, and 200 hours when set to 1-, 5- and 60-second pings, respectively. I pushed the Ambit2 and Ambit3 Peak to the brink on multiple occasions (notably, during a half-dozen 100-mile races) and have found that these battery life estimates are reliable.

Note that the “GPS accuracy” setting is different than the “Recording interval,” which can be set a 1- or 10-second intervals. Ambit does not just record its GPS location — it also collects data from its accelerometer (e.g. cadance), barometer (e.g. pressure and altitude), thermometer, and any accessory devices like the Smart Sensor HR strap or Stryd power meter. The 1-second interval will generate the most accurate data set, but it will fill the internal memory faster. On longer outings with infrequent opportunities to download data, use the 10-second interval.

Most Garmin GPS watches have a similar battery-saving mode, known as Ultratrac. I’m less clear about how it works, but it will also extend battery life by a multiple. For example, the Garmin Fenix 5 is rated “up to 24 hours” in its most accurate setting, but “up to 60 hours” in Ultratrac. With older models, Garmin has been criticized for optimistic battery life estimates; I’m unaware of the consensus about the Fenix 5 models.

Accuracy of Ambit3 Peak at Okay/60-second GPS pings

Why did Ethan and Alan get shortchanged by 20 percent when using the Okay/60-second GPS interval?

One reason is that the Ambit fails to record every meander, such as turns and switchbacks. To illustrate this, last night I walked in a zig-zag pattern down the nearby frontage road, once with the Ambit at the Best/1-second interval and a second time with it at Okay/60-second. The recorded tracks are shown below. Notice how the red line (1-second interval) records my zig-zags perfectly, while the blue line (60-second interval) missed all of them.

The red line is my recorded track at 1-second GPS pings. The blue line is my recorded track at 60-seconds. Notice how all of the meanders are missed when recording at the 60-second interval.

On straight trails, the accuracy of the Ambit at 60-second intervals is probably quite good. But on trails with lots of tight turns and meanders, its accuracy will decrease.

On straight trails, including trails with long straight switchbacks such as here in Rocky Mountain National Park, the Ambit does well. Notice how closely the recorded track (blue line) follows the mapped trail.

In addition to missed meanders, I think there is a second explanation for the inaccuracy of the Ambit at the Okay/60-second interval: its GPS location data does not seem as accurate. For example, notice in the zigzag test that the blue line shows a path through the grass, even though I was on the road. And in the image below, notice the dramatic zigzag across Fifth Lake even though I remained on its west side.

I have a few theories about what’s happening here. It’s possible that less battery power is being sent to the GPS chip, resulting in less accurate location data. Or, with GPS pings far and few between, the Ambit struggles to remain connected with multiple GPS satellites, thus compromising its accuracy. If you have a better understanding of GPS and GPS watch technology, please share.

When set to 60-second pings, the Ambit occasionally records an erratic GPS position.

Accuracy of the Ambit3 Peak at Best and Good GPS settings

When using its Best or Good setting, I have been remarkably impressed with the accuracy of the Ambit3 Peak (and its predecessor, the Ambit2), especially in consideration of its $300 price, 2.50 oz weight, and battery life.

With the Ambit in Best/1-second mode, I have run three certified marathon courses: Colorado, Boston, and Houston. The Ambit recorded distances of 26.44, 26.39, and 26.51, respectively, versus the official marathon distance of 26.22 miles, or about 1 percent less. However, this difference can be partially explained by how I ran the course (e.g. swinging wide around tight corners, and not running in a perfectly straight line between apexes), not the accuracy of the watch.

With the Ambit in Good/5-second mode, I have run multiple 100-mile ultra marathons. Throughout these races, the Ambit has displayed reliable pace, heart rate, and vertical change data. The recorded distances seem fairly accurate, too. At Ultra Trail du Mont Blanc (UTMB) last year, my Ambit3 Peak recorded a total distance of 105.18 miles, versus the official 104.76 miles, or an overshot of just 0.4 percent. It recorded 31,821 of vertical gain, versus the official 32,398, or 1.7 percent less.

In other ultras, the difference between the official distance and my recorded distance has been greater, like 107.12 miles at Run Rabbit Run 100, which officially is 102.9 miles. That’s a difference of 4.1 percent. However, I don’t necessarily trust the official distances for ultra courses — they are not professionally measured or certified. (UTMB is the biggest ultra in the world, and I bet they have measured their course with more accuracy than most.)

Guidance

Based on my experience with the Ambit, in terms of its accuracy, here are my usage guidelines:

Use Best/1-second mode if:

  • Your activity will be completed in less than 20 hours; or,
  • You need the greatest level of accuracy.
  • Examples: marathons, 50-mile ultras or fast 100k/100M courses, day-hikes, trail mapping

Use Good/5-second mode if:

  • Your activity may exceed 20 hours but will finish within 30 hours; or,
  • You need reasonably accurate data.
  • Examples: 100k or 100M races, 1- or 2-night backpacking trips

Use Okay/60-second mode if:

  • Your activity will exceed 30 hours; or,
  • The data needs to be only generally accurate.
  • Examples: 3-day backpacking trips or longer

If the Ambit is not offering the level of accuracy you need for the duration that you need it, remember that it can quickly and easily be recharged in the field (or during a long stop in an aid station). It has a 500 mAh battery, so a portable charger like the Anker PowerCore+ mini could recharge it about four or five times (after accounting for the energy lost from current conversion).

Guidebook writing

When performing field research for my guidebooks, I use the Okay/60-second setting. It:

  • Produces a functional GPX track of where I went;
  • Displays roughly accurate average pace data, especially over longer distances;
  • Records accurately my net vertical change, which is usually the limiting factor each day, not horizontal mileage; and,
  • Gives me a generally accurate mileage read-out for the day and between landmarks.

This GPX track confirmed what I observed in the field: that the mapped location of the Gourd Lake Trail is not correct. However, it is not sufficiently accurate for public sharing — it clearly missed a few switchbacks and got some wonky location pings.

However, the mileage data and GPX track are not suitable for public consumption. The listed mileages in my guides are measured in Caltopo later on. And I intentionally do not include a GPX track of the routes — if you need that level of assistance, you probably shouldn’t be on it. If I wanted to skip the Caltopo step and/or share a very accurate GPX track, I would need to use the 1- or 5-second setting.

Have questions about the accuracy of the Ambit? What’s been your experience with it at Okay/60-second GPS pings? Leave a comment.


Disclosure. This website is supported mostly through affiliate marketing, whereby for referral traffic I receive a small commission from select vendors, at no cost to the reader. This post contains affiliate links. Thanks for your support.

26 Responses to Accuracy of Suunto Ambit3 Peak at Okay/60-second GPS pings (200-hr)

  1. Hunter G Hall February 24, 2018 at 5:10 pm #

    Great post, thank you.

    I recently got the Ambit Peak 3 and did the same comparison with Gaia, which is my go-to GPS solution.

    In addition to the distance issues, I remember altitude issues when compared with Gaia as well, which you and Alan were kind enough to explain to me.

    I wish it had a vibrate alert though…

  2. Amir Erez February 24, 2018 at 5:19 pm #

    My Ambit 2 sometimes gives me wonky data on 10 sec interval recording. Have you done any experiments adjusting the interval recording rate? On a 3-8 day long backpack what interval recording rate do you use? Thanks!

    Amir. Boulder, CO

    • Andrew Skurka February 24, 2018 at 5:48 pm #

      What do you mean by “wonky data”? What I have noticed is that Strava does not interpret it correctly, so a 24-minute mile get translated into a 4-minute mile. But Movescount knows how to handle it, no problems there.

      When I use the 60-second ping, I always set the recording interval to 10 seconds. I don’t know for how long I can run the Ambit in these settings without running out of memory. Last summer on a 9-day trip I ran the Ambit2 for a little over 99 hours without downloading the data, and had not received any warning messages about nearing the memory limit.

  3. dgray February 24, 2018 at 6:40 pm #

    Can you change the accuracy setting in the middle of recording a track? So, for example, can it be set it to “Best” for a twisty section of the route and switch to “OK” for a long straight 10 mile road walk to maximize the balance between accuracy and battery life?

    • Andrew Skurka February 24, 2018 at 11:30 pm #

      Yes, by using the multi-sport feature, but instead of transitioning from swim to bike, you’re going from twisty trails (1- or 5-sec intervals) to straight roads (60-sec intervals). You would just need to create two Sport Modes. They could be identical except for the GPS accuracy interval.

      To switch between sports, press and hold the upper-left lap button.

  4. ankit February 24, 2018 at 8:29 pm #

    How does this compare to the Fenix 5 Ultratrac? I think the general consensus is that the Ultratrac is worse, but I am unable to find a head-to-head comparison with the Ambit3 @ Okay.

    • Andrew Skurka February 24, 2018 at 11:28 pm #

      Uncertain. Someone would need to wear both watches simultaneously. If Garmin would like to send me a watch I’d be happy to.

  5. Bob S. February 24, 2018 at 11:12 pm #

    Andrew, I believe newer Garmin GPS watches including the fenix 5 uses smart data recording (the default setting) and not Ultratrack. When data recording is set to smart it records key points when direction, pace, or heart rate changes.

    The other setting is every second.

    • Andrew Skurka February 24, 2018 at 11:27 pm #

      Per the Garmin website, the Fenix 5 series have the Ultratrac option. I don’t see any mention of a “smart data recording” setting. Or am I misunderstanding some of their hyperbole?

      • Bob S. February 25, 2018 at 4:23 am #

        Andrew, I dug into this a little more. Here is the link to the fenix 5 “smart data recording” feature. The setting is under system settings –>> data recording. I guess the claim is when data recording is set to “smart” it records key points when direction, pace, or heart rate changes. It doesn’t turn off the GPS power; it just makes a smaller file to save space on the watch so you don’t gain any significant time on the battery”

        https://www8.garmin.com/manuals/webhelp/fenix5/EN-US/GUID-1500E73F-F386-49AF-A542-25D4B1655A08.html

        My newest Garmin did not include the UltaTrac feature so a made a bad assumption. The fenix 5 manual was a bit murky but I fund the UltraTrack info in the back of the PDF version. – Thanks

  6. Rick February 25, 2018 at 12:43 pm #

    I tested the Ambit3 Peak against Gaia and my IPhone 6S when I got the watch, on hikes in the mountains. I was very disappointed in the 60 sec recordings – they were off by about 30%. Pretty worthless for date use from my perspective. The 5 sec recordings were low by about 2 1/2%. The 1 second was pretty close to the Gaia. I decided to use the 1 sec recording when I hike/backpack and take a power source/solar charger to get more precise data. I would love to have 200 hrs of battery life for recording time, but not at that lack of accuracy. interestingly, my Garmin inReach Explorer + is pretty accurate at 60 sec recordings compared to Gaia and my iPhone – so not sure what is different with the technology between the inReach and Suunto and why inReach can get accurate for hiking at 60 sec where Suunto can’t, but maybe the technologies are different.

    • Andrew Skurka February 25, 2018 at 1:31 pm #

      Curious: What are you doing with the data that you need such precision?

      • Rick February 25, 2018 at 7:14 pm #

        For one when comparing to some of the data books like for the CT/CDT for distance between water, trails, specific places, etc. Or on some marked trails that say x miles to x. Second for my ego – if I do a 100 mile high route and it only says I did 70 miles I feel like a wimp :). In all honesty it’s kind of a thing where when I get tuned to data from hikes where I did x miles and x vertical – knowing the data is not right just bugs me……

        • Andrew Skurka February 26, 2018 at 9:18 am #

          Got it.

          BTW, I think the vertical readings are probably accurate. The watch is recording its altitude every 1- or 10-seconds (depending on the specified recording interval) and does not need GPS to find its altitude, since it has an internal barometer.

  7. Brady February 26, 2018 at 12:36 pm #

    I can’t speak to the specific processing involved in the Suunto, but typically your accuracy increases with the length of continuous time recording with a set of satellites. Using the ‘OK’ setting seems to shut off the GPS connection between point recordings which likely reduces the ability to further process the data using any sort of continuous time correlation.

    Perhaps a more likely issue is the number of satellites used to record each point. When turning on and off to record a point, there is likely a minimum number of satellites needed to save a point (probably 4). While using 6 satellites would result in a more accurate reading, if the focus is on saving battery, the software may collect the location once the minimum number of satellites are available. More frequent measurements where the GPS never turns off would result in a greater probability of being connected to more satellites for each measurement resulting in greater accuracy.

    Again, I come at this not from specific knowledge of Suunto/Garmin software, but just general GPS processing.

  8. Michael March 29, 2018 at 4:48 am #

    Sorrry for my english.
    My Ambit3 Peak seems to only use the “ok” setting instead of the “good” setting. I set up a sports mode with “good” GPS accuracy but just got data every 1 Minute. When i looked up the sports mode on my smartphone it still was set to “good”. What did I wrong?

    • Andrew Skurka March 29, 2018 at 7:17 pm #

      It sounds like you might need to sync your watch, using the Movescount app or computer cable with MovesLink.

      Changes in Movescount don’t relay to the watch until a sync.

  9. Joe April 1, 2018 at 8:52 pm #

    This is a very low-tech question from a man who wears a $30 Timex on his wrist.

    I run about 7 miles a day on average. I try to always run at least 6, and almost never run 10, although I often aspire to.

    So…I’m a 66 year old slug. So be it.

    Don’t need a lot of the stuff y’all are discussing, but it would be cool to know actual distance, and elevation gain and loss, I’m thinking.

    And I would love always knowing my current elevation, whether I’m running or not.

    Don’t need splits or maps or laps or apps.

    Don’t need to save data.

    Given my semi-Neanderthal nature and needs, any thoughts on a good watch for me?

    Thanks in advance.

    Joe

    • Andrew Skurka April 2, 2018 at 3:35 pm #

      Not sure there’s a device that will give you distance and vertical without recording it. I would probably still point you to an Ambit, but a cheaper or discontinued model. Occasionally you’ll have to download the data in order to free up the memory, but you won’t have to look at it.

  10. Chris Robinson April 2, 2018 at 7:26 pm #

    I bought a used Ambit3 Peak a few weeks ago. It seems to work perfectly but last week I ran an 8k and it added an additional quarter-mile for EACH MILE. I thought it must be set to battery save mode but Movescount says I’m at “best” for Running mode–the activity I selected. The race is run through downtown Chicago and changes directions every few blocks, but I assume at 20%+/- either I’ve been had or I am missing something. Any ideas? Thanks.

    P.S. One confession: I was checking the watch every mile and it was 20% all miles except the last one where, in all of the commotion, I forgot to stop the activity. And didn’t until after an Uber ride home. Is it possible that it re-calculated mileage after I stopped?

    • Andrew Skurka April 2, 2018 at 7:47 pm #

      That seems way off. In comparison, I’ve run three marked marathon courses with the Ambit2 and Ambit3 Peak, and each time it’s only been off by .01 miles per mile, or about a quarter-mile for an entire race.

      I wonder if the tall buildings have something to do with it. Look closely at the track — a few zig-zags could really mess with the distance. Most GPS units will struggle to get an accurate reading in that kind of environment.

      Also, did you ensure that you a GPS fix before starting? If you didn’t get that initial fix, this could also explain it — the watch is constantly trying to “find” you, which is hard when you’re moving at 10 mph among tall buildings.

      Not sure to what platform you are downloading the data afterwards, but Strava has a “crop” tool so that you can remove either end of an activity (in your case, the end).

  11. Christopher Robinson April 2, 2018 at 8:18 pm #

    Thanks, Andrew.

    We definitely were running in the shadows of a lot of tall buildings. And I also didn’t get a GPS fix before starting.

    I’ll try again on a more open trail.

    Really hope that’s it!

    Thanks again.

    • Andrew Skurka April 2, 2018 at 9:04 pm #

      The lack of GPS fix probably explains more than the tall buildings.

      If the fix is not happening quickly, you probably need to sync you watch with the app or desktop software so that the satellite locations get updated.

  12. Brett Tucker May 11, 2018 at 6:44 pm #

    One thing I’ve noticed about the Ambit3, whether at Good or Best GPS accuracy setting, is that the watch doesn’t necessarily record for posterity a gpx trackpoint at the stated interval (5s or 1s). In fact, it seldom does. I’ve confirmed this in mapping software, based on the interval at which the trackpoints show up on the map (in satellite view) relative to the estimated speed with which I traveled between any two points. In one case, walking in my neighborhood, with the watch in Best setting, on a straight section of road, two adjacent trackpoints are over 200 feet apart, a distance I would have covered in perhaps *30 or 40* seconds walking at 3 mph. I assume this to mean the watch is actively discarding points that it determines are more or less on a straight-line trajectory. This would be merely curious, except that whatever algorithm the watch uses sometimes leads to a discarding of points on curved sections of the route that it otherwise should be picking up on as relevant to total distance traveled and track accuracy. Compared against Gaia on an iPhone, that same straight section of neighborhood trail produces a waypoint about every 70 feet rather than the Ambit’s 200 feet. Consequently, Gaia also records a greater distance traveled – granted, only a hundredth of a mile over a one mile distance, but it does add up over a full day of hiking. Deal killer? Absolutely not, especially since I’d much rather run the watch over the course of a day of hiking than Gaia. It’s just strange that the Ambit doesn’t quite seem to live up to expectations in this regard or produce a result consistent with the way its GPS functionality is advertised.

    • Andrew Skurka May 14, 2018 at 9:54 am #

      I have thought this, too. However, you might want to try downloading the file in a different format. I think that some download options and platforms dumb-down the file.

      Recently I wanted to crop an activity that I recorded with the Ambit. I used https://www.fitfiletools.com, and it seemed to give me 1-second resolution. But when I view this same activity in Movescount or Strava, or when I download a GPX of the file and upload it to, say, CalTopo, the resolution is not the same. I don’t know where it’s being dumbed-down, but clearly it is.

      • Brett Tucker May 25, 2018 at 4:57 pm #

        Thanks, Andy. I haven’t found the file format to matter (although it could in certain instances), but in my case it was being dumbed-down during import to CalTopo. In 1-second tracking mode, the watch produces more track points than are needed to draw the track with reasonable precision, and so it seems CalTopo (and probably other programs) renders the file with fewer points, based around the shape of the track, perhaps to improve performance on the platform.

        A theory I have, and perhaps others could confirm or deny this, is that GPS watches like the Ambit3, which have built-in accelerometers, use this technology to improve tracking accuracy. For instance, in 1-second tracking, the watch doesn’t seem to produce a track point precisely every second, but more in a regular, even manner over a given distance traveled. Whether my pace increases or decreases, the track point interval remains fairly consistent, maybe every 10 or 15 feet of forward progress between points. When I stop moving with the GPS recording, the watch mostly stops adding points to the track until I resume forward progress. This is a great feature for anyone using the Ambit in hopes of producing a serviceable track line for use in mapping software, since it avoids the need to do much post-editing or “cleaning” the track of random “squiggles” that plague many standalone GPS units (and some apps like Gaia) when movement slows or stops, particularly in heavy tree cover and other challenging environments.

Leave a Reply