Stride length calculation on treadmill way off after upgrade from 2.50.0.0 to 2.70.0.0

Did a treadmill workout with 2.70 and my HRM Run last week and stride length was ok, I just checked.

Thanks for the input. Is the curve points balanced or is it varying much?

Same - i had not connected it being worse after the update but that is what I have seen. Prior to recent update, I was seeing it be off by about a half mile on treadmill runs which was pretty good I thought. Now, I'm breaking world records running a 3:15 mile pace when in reality my pace is 10:45. Occasionally it is too slow by a couple of minutes but usually it says my pace gets faster and faster. I get audio pace updates every few minutes and each update is faster.

I reviewed my cadence plot and it is similar - it's all over the place. My cadence / stride length in the past is extremely flat.

--------------------------------------------------

2.7 software treadmill run -- https://connect.garmin.com/modern/activity/3971782938

edit: now public

edit: here's the plot from runalyze --- pace starts way off and remains so -- treadmill pace was 10:42 from start to finish of the run

----------------------------------------------------

2.5 software treadmill run -- https://connect.garmin.com/modern/activity/3926759394

also, I reviewed a few GPS runs since update - cadence is all over the place on those as well. It is just not as obvious unless you're looking at your graph because distance is coming off GPS rather than the calibrated stride length

I have done the "calibrate & save" for each treadmill run - does not seem to improve the distance. I think the issue is with how cadence is counted. the distance is a calculation based off the avg stride length it seems.

2.5 software GPS run  -- cadence relatively flat -- https://connect.garmin.com/modern/activity/3935238187

It is ok. Forgot to mention: I was also using my Garmin footpod as always on treadmill.

Doing a treadmill run with and without footpod can't really be compared.  Both are trying to do way more than just count steps, but the accelerometers in the footpod are likely to be much more accurate than those in the watch on your wrist.  You'd expect within 2% accuracy with the footpod most of the time out of the box.  The wrist really needs to learn your specific movements and is a much more complicated algorithm. 

Just to be clear, you are setting your watch on treadmill mode but running outside?  No idea how it's supposed to do with this as the movement of the arms will be a bit different. It's designed to learn from your outside runs for running on the treadmill, but not sure if it's supposed to work the other way as well.

In general, you need to do a number of outdoor runs (with GPS) for the watch to learn your motions in order to get better accuracy on the treadmill.  I wonder if this update reset that data.  These watches do a lot more than just counting your steps and multiplying that by your stride length.  As you change pace, your stride length is going to change.  The watch will use the acceleromter data from your arm swing to try figure out speed as well.  There will be some combination of these two computations it uses.

The treadmill run outside was just a test to see what was going to happen with the stride length. I knew it is of no scientific value at all. Was just curious on what it would record. If it had worked the way i thought i should, I should have a steady pace and a stride length around one meter. I ended up with an average stride length of two meters, about twice my regular stride length.

I have had several runs outside, to have the watch learn from my runs. But I think you are on to something. with your thought about the update resetting something in the watch. I have tried to do at least 5 calibration runs on 2.70, and it has not got worse, but it also has not been any better. The stride length recording is not working right now.

Perhaps the watch is missing some data point, and gets the next one, and do calculations based on that?