Incident Detection (all devices that use it!) NEEDS TO USE ITS OWN INTELLIGENCE to stop FALSE DETECTIONS!

So, since I wrote the first post, I've had at least 3 more times that I can think of that the FALSE Incident was "detected" and in each case, my wife was notified, I never even knew there was a "detection" until she had tried multiple times to contact me and finally got through. Meantime, my ride was ENDED on the Fenix 6 Sapphire. That's why I run my Garmin EDGE 830 at the same time (which I had for almost 2 years before getting the Fenix 6) so that I have a backup device. The EDGE did NOT DETECT ANY PROBLEM and continued normal recording, so I got credit for miles, time, training data, etc that would have been blown had I not had both devices running. And there was NO REASON for the incidents other than stopping my bike to add or remove extra clothing. There was NOTHING that a wrist-watch should have sensed as an "incident." And especially if the EDGE device mounted right on the bike didn't detect anything. 

I have no reasonable choice but to turn OFF the "feature" on the FENIX because it has caused so many of these incidents that put my wife to worry, forced me eventually to stop a ride when she was able to call me, and caused lost data from the detection point forward, amounting to MANY miles that should have been recorded on the device. 

As I pointed out above, it should be VERY EASY for the device to "perceive" that there is NOT an incident, or at the least that the user of the device is doing fine and continuing the activity, and should be able to make notifications for any help that MIGHT be needed. INSTEAD, the ride ENDS, maybe segment bests are ruined, etc etc. 

In my opinion, the software of the device is badly designed and clearly needs to be developed. The INTENT of the feature is an excellent one. The EXECUTION of it is poor, especially with the host of actual data that the device is monitoring that could enable very logical conclusions to counter the false detections and thus dismiss the alert right away.

DouthatBiker said:

As I pointed out above, it should be VERY EASY for the device to "perceive" that there is NOT an incident, or at the least that the user of the device is doing fine and continuing the activity

How would this be easy let alone very easy?  What if one of your sudden stops was you actually hitting a tree?  You may be concussed and immediately start moving without having proper control over your senses.  The watch sees movement and then "thinks" you're fine, let's continue?  Year's ago my wife was bike riding and struck on her head by a side mirror from a hit and run driver.  Witnesses say she immediately got up and started walking around aimlessly. Had there been no witnesses and the technology was available back then it would have been invaluable. 

I can come up with rare circumstances too... Not minimizing or ignoring what happened to your wife, but that is a SUPER RARE scenario. So maybe someone grabs a person and kidnaps them, etc... 

But come on. Those are NOT the intended purpose of the "incident detection." And the VAST, INNUMERABLE number of times that users get "detections" are NOT INCIDENTS AT ALL! That's the point of the complaints, and why the "feature" becomes basically useless! 

Furthermore, the DATA need not be discontinued AT ALL. The RIDE/RUN/whatever ought to continue regardless of the POSSIBILITY, (slim that it is) that there really is an incident. The person that is getting NOTIFIED can then still TRACK THE person online and know whereabouts, etc. 

As to would this be easy to detect, I still say absolutely. The watch tracks ALL KINDS of info about what you are doing, EVEN WHEN YOU ARE NOT doing a specific activity that you are recording. It knows when you are actively walking, or running, or cycling, etc and puts them into your day's activity without you even intervening to start a recording. (As long as you allow that activity tracking, which I do; and it isn't flawless, it thinks I'm cycling when I mow my grass, I guess because it senses speed as faster than walk, and yet my arms are not swinging but more still, on the mower's handlebars.) The point is, there is great potential for logical analysis that happens all the time anyway, with NO changes to the already existing software. I contend it should be very easy for Garmin to UTILIZE the data that is being recorded to REDUCE FALSE DETECTIONS. 

They ARE a problem and they DO cause a huge number of folks to just not use the feature because it is simply a hassle far too often. Happened again yesterday on my road ride, absolutely nothing happened other than me stopping to add some clothing for more warmth before a long descent. No sudden anything; I simply added clothes from my pack. My EDGE on the bike didn't have any issues or show a problem. The watch most likely vibrated to alert me but it does every mile for my lap reading anyway, and as it is under a sleeve, I don't always look at it. So halfway down the descent, she has texted me, and then is calling me. The EDGE shows the incoming call, I have to stop the bike and dig out my phone to answer, and it's yet ANOTHER of these false detections. Only a week since the last one, exactly the same scenario, (and the rest of my rides between had been indoors, when it's turned off anyway) not a good track record. I finally totally disabled it for road rides too.

Interestingly enough, I've been recommended by Garmin tech support persons to turn it off on MOUNTAIN BIKE rides because it goes off on those just because I've rolled over a log, (SMALL one, mind you, maybe 12" tall at most) or because I descend aggressively on trails I know, so there WILL be sudden changes in direction, etc. Have had many "detections" for nothing more than those reasons, and the watch COULD have also then recognized that I was continuing on in exactly the same manner as what led up to it's "detection." 

I turned it off on MTB rides because of that, but an MTB ride is a MORE LIKELY SCENARIO for a genuine incident! And now I've turned it off on the FENIX for ROAD RIDES TOO. Why? Because it is NOT SMART in applying a detection.

The point is, the watch takes almost NOTHING to conclude you had an incident, yet there is a VOLUME of data that continues to come in to the watch that tells it OTHERWISE but is ignored. THAT is my point. And GARMIN NEEDS TO FIX THIS FEATURE to make it far more valuable and actually USEABLE for the vast number of us that end up deciding it is a worthless feature that creates hassles on a regular basis.

DouthatBiker said:

I can come up with rare circumstances too.

Of course it's a rare situation.  How many people are getting injured every time they do an activity?  Am I missing an injury people get often enough that requires emergency attention that isn't considered rare?

DouthatBiker said:

And the VAST, INNUMERABLE number of times that users get "detections" are NOT INCIDENTS AT ALL

Well, I'd like to the official numbers on that but I've only had two and both times I bit on the trails riding my bike.  I can't provide you an answer on why your watch seems to be so sensitive.  I've rode at a high rate of giddy up and had to hit the brakes hard with no trigger. 

DouthatBiker said:

My EDGE on the bike didn't have any issues or show a problem.

Maybe it's your watch specifically?

I had this today. Two false alerts. One I caught, the other I couldn't as I was speeding down a hill. Having to stop to try to prevent a message to my wife was actually quite dangerous in itself. I also lost the second half of my ride. Feature now disabled :(