Haha! Our fields aren’t the best, but no cliffs nearby.
You are the coach, but I do not think your citing of the finals at the Cal State Meet is really inconsistent with my statement. “Pretty darned quick” is not the equivalent of “unheard of” or “beyond the reach of the best.” In an average dual meet, how many of the varsity sprinters finish in under 11 sec? Now line an average middle-ager up against those kids…
One reason I cited the marathoner is that I’ve run them myself. One year a decade or so ago, after watching the Chicago marathon I went out for a 26 mile bike ride, seeing how hard I needed to ride to beat the winners. Yeah, I woulda beaten them - not by much. But it required a concentrated effort. 13 MPH for 2 hrs on a hybrid bike on multi-use trails is not exactly creeping along.
I was a mediocre HS track runner - hurdles and middle distance. I long ago stopped even calculating how much my max 100 would be dusted by an elite quarter-miler. Heck - I was never a speed demon, so I doubt I could even have kept up with a 4-minute miler for 100 yds!
I had a computer on my first grownup bike and remember struggling to get the peak to 30 mph and it still took what passes for hills around here to do it. That was a mountain bike with knobby tires, not made for speed, but I was a pretty fit 18 yo. 20 mph was sustainable flatland but quite a workout.
I’ve use a skiing app like mentioned upthread. It says I’ve hit 40 mph which is, for this skier, too fast for comfort.
Consensus clear that the number is not real, I wonder if the movements of Ultimate Frisbee are such that speed is going be inaccurate more easily?
My WAG is that the sudden rapid changes, running full force one way then trying to rapidly stop and dive back the other direction, even briefly, confuses the accelerometer based algorithm.
Even heart rate can be funky. I’ve got a forearm based heart rate monitor device and every so often the first few minutes are reported as way above my heart rate max while I am just warming up. I’m guessing it’s counting both lubs and dubs by mistake at that point then gets it right as I pick up speed. No I do not want to update my heart rate max thank you app!
In general the rule is that the default explanation for a weird number is error. Just not always.
There’s no way it’s using accelerometers to calculate speed. I’m sure it’s only using GPS.
I went to a random accelerometer page and looked at their support page.
At high speeds, your running form might not be as natural, and your arm movements may become irregular, resulting in inaccurate speed and distance readings.
As a rule, any activity that interferes with your natural arm swing (e.g. pushing a stroller or walking a dog) will have an adverse effect on the accuracy.
This tracks (heh) with what I said above. Lots of varied arm movements in ultimate, both at high speeds and while relatively stationary. If the tracker relies on the swings of your arm, it’s going to go nuts during ultimate.
Some Garmin watches feature an onboard accelerometer that can provide speed and distance indoors or when a GPS signal is lost . These devices are pre-calibrated with a table of default values based on the averages of many runners and the calibration is refined every time the watch is used outdoors with GPS.
From here. So, the accelerometers would only be used for indoor activities.
ETA: I have a Garmin watch.
Devices with GPS use that to estimate speed, but most fitness trackers don’t have GPS. They measure steps over a given time period along with other data (like the swings of your arms) and make an estimate.
OK, my fitness tracker has GPS (it can connect to three different systems, US, Euro, and Russian, but I only have it connecting to US GPS).
ETA: I see where some of the questions about this has come from. I should have mentioned GPS earlier, I guess. I agree that if I’m running around like a lunatic with just a step tracker, the results would be garbage.
Biking speed sensors often use magnetic sensors relative to the earth’s field.
I don’t know that this device uses the accelerometer but often they will vary what sensor based on listed activities. On my app GPS gets turn on for running but is off for many others as a matter of programmed in.
Yes, there’s no way my watch was using the accelerometers. It even gives me a nice map at the end showing all the places on the field I ran around on (I should probably spend more time deep in the end zones if I want to catch more goals).
Garmin says their GPS accuracy is within 10 feet assuming good satellite coverage. If it’s measuring over 1 second, those are pretty big error bars.
Say you cover 25 feet in 1 second. That’s 17 mph, which is probably pushing it for a mid-50s guy but theoretically possible for a quick burst. But your watch might think you’ve covered anywhere from 15 to 35 feet, or 10 to 24 mph.
Add in poor satellite coverage with clouds, and you’ll easily see some ridiculous speeds occasionally.
Part of me really wants a watch like that, but the rest of me doesn’t want hard evidence of all the lazy poaching I do.
Haha! Poaching is an important part of the game!
I’ll try it with three satellite systems next time. Signal should have been fine – wide open field.
That is probably it. I just looked at my last run. I averaged a pace of 9:58 per mile. I’m not fast. That average number I believe. But it lists me going as fast as 7:22 per mile at some point. And nah. I never sprinted at more than what I could maintain for at least four or five minutes and even all out I was never that fast and certainly not now.
The sampling errors even out.
My first thought was “No way!”
Then I looked up what an athlete doing a quite respectable 10-second 100-yard dash is doing: 20.5mph
World record 100-yard dash for men is 9.1 sec, or 22.5mph
Of course, you are thinking more along the lines of burst speed, which is apparently much more possible than I had thought.
To get an idea of how feasible it would be for an ordinary Joe to do this, next time you are on a treadmill, firewall it and see how the highest speed feels to your legs.
For me, the 10mph max speed of treadmills at my gym feels pretty close to my limit, and I can’t imagine doubling that pace, and I have been running for 20 years.
Yeah, I’m not starting from a standing position, I’m running close to flat out, and then really flat out for another 10 or 15 yards. (This is not to say that my watch was accurate, but comparisons to 100 yard dashes don’t make much sense – standing start and much longer distance)
I know OP is talking about VERY short spurts. But when our kids were middle school, I was around 40, and in pretty good shape. Not running, but doing a ton of martial arts and lifting. We took the to a track and did some sprints/quarters. MAN, if you are not used to it, 100 yds is A LONG WAY to sprint all out! 
And my last long slow, averaging 11:49 minute per mile pace for a bit over 12 mile run, has me peaking at 7:42. 100% nope.
So, just for fun (whee!), I went for a run around a track, four laps. I averaged an 8:40 mile, and didn’t see any spikes in my speed – started faster, slowed down to just below my average for the middle, and a little faster at the end, with no spikes along the way. The GPS tracking seemed to be a little wobbly for the first two laps and then really locked in for the second two.
And, I looked at my detailed tracking from my Ultimate games, and the high peaks were when and where I remember them – running for the disc in the end zone one time, chasing down a defender on the other side of the field the following week.
Anyway, the watch is directionally correct – showing spikes when I’m really taking off, steady pace when I’m running a steady pace. Based on most of the responses here, I guess 20 mph is pure fantasy, although this random cite is less pessimistic than most people here.