1. 
   GravityGuy
   

   new member
   Joined: Apr '09
   Posts: 1

   Hi rlerdorf,
   It seems you've got a good understanding of how the sensor works and what the raw data looks like, although, we should clarify that the raw data you see through the API has already been processed, and is not strictly raw sensor output.

   Would you mind giving us all a brief description of how we can maximize the accuracy of our run distances. It seems from my experience, any large variation in pace will cause problems. I've been doing a lot of walking lately in order to work up to running. If I begin to jog, the pace display goes to zero, and the distance is not accumulated. If I return to a walking pace, it fluctuates wildly and then settles down to a proper pace value.

   Another runner on the Nike forum told us about her half marathon run with walking periods. The calibration was done earlier with a 5k run. She got a wildly inaccurate distance for the half marathon. I think is was caused by her walking some of it.

   Could you explain this from your own experience and from your knowledge of the sensor. Nike is not very helpful in this regard. A good description of the sensor can be found at http://electronics.howstuffworks.com/nike-ipod1.htm.

   You say that the raw data you see is distance values every 10 seconds. Nike has obviously processed the actual pressure data from each footfall and turned this into a distance. It seems to me that they didn't account for workouts that have a variable pace such as wind sprints, hill climbs, or run/walk alternating periods. It also seems to me that the insole pocket is the best way to maintain consist and proper operation of the sensor with each footfall.

   I'd really love to see the actual raw sensor data.

   GravityGuy

   Posted 3 years ago #      
2. 
   rlerdorf
   

   senior admin
   Joined: Nov '08
   Posts: 258

   There is separate calibration for running and walking. I would assume that if you stuck to either running or walking in the same workout, it would be more accurate. Switching back and forth in the same workout probably doesn't work too well. But I don't really know, I tend to always pause my workout when I walk. I only measure my runs. And for me it has been very accurate over the years.

   From a technological standpoint you have to realize that this is a $25 souped up pedometer. It is limited how accurate it can get. I think the fact that they got it this accurate is pretty amazing based on how simple it is. Throw wildly varying parameters at it, and it is bound to be tripped up.

   Posted 3 years ago #      
3. 
   morey000
   

   new member
   Joined: Aug '09
   Posts: 4

   I've calibrated my sensor on a 1 mi measured course. I ran on a marked, level course which has markings on the pavement every 0.25 mile. on the outbound for 2-3 miles it was dead on. Accurate to within 0.01 mi at each marking. Then, on my return trip running on the same flat course at the same speed- it was a good 0.1 to 0.2 miles off each mile. So at times it's very accurate, at other times off 10% to 20%. haven't figured out why.

   Recently ran a race on a measured 3.5mi course. Sensor came up at 3.2 miles. Once again- not sure why.
   Today, ran 6+ miles. First 2 mi came up perfect, hit the measured 2 mi mark right on- but by the end of the run I was off about 0.2 to 0.3 miles after 6 miles. Granted, not bad for a $29 pedometer.

   I've heard that the Garmin Forerunner 405 gets the distance pretty good, but the pace varies wildly.

   Gotta' admit. It's a great little gadget for such a cheap price.

   Posted 2 years ago #      
4. 
   wiremanart
   

   new member
   Joined: Jan '09
   Posts: 1

   I have found the sensor to be about 97% accurate if you run at the pace you calibrated. Deviating from that pace going faster will result in skewed performance and an increase in miles, however if you go slower, the same skewing applies where the pace drops significantly lower than it probably should so really even that kind of evens itself out over time. I'm not an expert or anything, but I've ran nearly 5,000 miles so take my input for whatever it's worth. I reiterate what rlerdorf and morey000 said, for a 20$ piece of equipment it's as close as I think you can get with accuracy. Plus factor they are not custom made, they are made for millions of people, and it's that much more amazing.

   I don't have a forerunner as my budget doesn't allow for it. The thing you have to worry about with those is charging and the software, but moreso trees, buildings, and weather affecting the satellite synchronization. (I've done my research ..)

   Posted 2 years ago #      
5. 
   nolli
   

   new member
   Joined: Aug '09
   Posts: 2

   For what the device is and at the price, I am surprised how accurate it is.
   I recently bought a Sportband V2 and ran 10.5KM the night I purchased it. Without calibration is gave me 9.5KM, so a little over 10% off wasn't too bad.

   On the same 10.55KM route I get results anywhere from 10.44KM-10.58KM which I think is a very acceptable margin for the distance.
   Definitely about keeping to the pace you calibrated it at. When I run a route faster than when I calibrated, it does often give me a slightly shorter distance but overall it is great for £40 with the sensor.

   I don't think I have ever walked, or calibrated a walk so can't really help you out there.

   Posted 2 years ago #      
6. 
   4U2C
   

   member
   Joined: Mar '09
   Posts: 24

   For me it depends on the terrain. Sidewalks it's usually within .1m. Sensor always is over, not under, unless I run on a treadmill, then it is under. When I used to walk with the sensor it seemed dead on. I once calibrated a run via a track, then went for a run after calibrating, and it said I ran 4min miles... lol so I went back to default and have left it at that. I wrote a vb script to offset my pace with the actual measured distance with http://www.gmap-pedometer.com/ . that way I can accurately get the cals burned, distance, and pace.

   Posted 2 years ago #      
7. 
   D-A-D-D-Y
   

   new member
   Joined: May '09
   Posts: 1

   I haven't calibrated mine and it is fairly accurate when I run an 8 min/mile pace. If I run faster, it thinks I haven't gone as far as I have, if I run slower, it says I've gone further than I have. For instance, I just ran 3 miles in 23:01, but it thinks I went 2.92 miles. So it is kinda like I get penalized for going faster! Luckily I typically run close to 8 min/mile, so the default calibration works for me.

   Posted 2 years ago #      
8. 
   fingerz1969
   

   new member
   Joined: Oct '09
   Posts: 1

   hi - new member - love the site!
   i have found that it varies when going up/down hill quite a bit, i assume that the angle of the foot is different when hitting the floor. other that that i never calibrated and a 10k run came up at 9.82 (i think). loses accuracy when battery is running out (on my 2nd one)

   Posted 2 years ago #      
9. 
   ATJ777
   

   senior member
   Joined: Sep '09
   Posts: 95

   I would think that while running uphill your stride would be shorter which would mean more strikes per set distance so it would over estimate the distance. Running downhill would cause a longer stride, less strikes per distance and under estimate the distance.

   Posted 2 years ago #      
10. 
    morey000
    

    new member
    Joined: Aug '09
    Posts: 4

    I read an article about how the sensor works. About 40 years ago, somebody figured out that your pace is roughly inversely proportional to the ground contact time of your foot. So, by measuring how long your foot stays on the ground, you can get a pretty good estimate of your pace. It's a bit different for each person- which can be adjusted via the calibration. It's very accurate if you run at your cal pace on level consistent terrain. Change your pace and it gets a little off.

    I have a similar experience as DADDY above. When I run faster, it underestimates my distance (and pace). When I run slower, it overestimates.

    Posted 2 years ago #      
11. 
    zoranbo
    

    new member
    Joined: Feb '09
    Posts: 2

    For what it's worth - I ran a training run on a half-marathon course (21.1 km) with my calibrated sensor showing 21.2, so it was pretty good. However, during the race on the same course, where I chased my personal best and intentionally lengthen the stride, the sensor showed 20.2km. It seems that the frequency of your steps in a certain period of time affects its accuracy.

    Posted 2 years ago #      

Reply

You must log in to post.

RSS feed for this topic