No announcement yet.

Tracked Swimming Parameters

  • Time
  • Show
Clear All
new posts

  • Tracked Swimming Parameters


    In my opinion, swimming mode needs more love regarding parameters.

    At the moment we do have things like time, strokecount, lapcount/distance, swolf.

    SWOLF is a rather useless parameter, and I never encountered a situation where I would have needed to take a look at it. Also it's rather simple to calculate, you don't need your watch to calculate it.

    What actually would be useful parameters for in-water Training Feedback would be:

    Minute Frequency and Distance per stroke, because those 2 truely matter to every swim coach and should matter to every athlet. Those 2 Parameters do say a lot about the quality of your Stroke, and you can't calculate them while resting when doing a sprint set.

    To have accurate data you should also track the time/velocity from the wall to the first movement of the arm, since then you also could calculate your submerged distance. This would work for all strokes except breaststroke. But if you do a proper underwater pull, even this could be calculateable.

    Indoor Swimming mode, if you export you data to tcx, is too close to running data. I understand why you do this, and if you do use hr data it makes sense but you're missing out on valuable swimming data, because should also track the times for everey single stroke.

    Tcx export does have it's issues too. On the one hand you do get speed data with 16(!) decimal places, on the other hand you do export wrong laptime. Not only do you truncate Lap time to whole seconds, instead of getting at least 2 decimal places, but you also export wrong data. I have swam:

    25m breaststroke, 7 strokes, 19.0 seconds in tcx export
    25m breaststroke, 7 strokes, 20.4 seconds in garmin connect

    Would be nice if you could fix that.
    Last edited by neymo_aut; 01-25-2018, 06:04 AM.

  • #2
    As you can see the calculation with considering of submerged on the right. Normally a coach does this with a stopwatch, but if you train alone it would be very useful to have this information on your watch.


    • #3
      Originally posted by neymo_aut View Post
      To have accurate data you should also track the time/velocity from the wall to the first movement of the arm, since then you also could calculate your submerged distance.
      The way swimming detection works, tracking underwater time would be inaccurate and velocity would be impossible, because the watch doesn't know the distance covered from wall to first stroke. And if you were thinking about calculating velocity using the accelerometer, then I believe it is not nearly accurate enough to give a meaningful result.

      Current: nuvi 255, eTrex 30, Forerunner 310XT, Forerunner 735XT, HRM-Tri, HRM-Swim, SDM4, Virb Elite, tempe
      Retired: eTrex Legend, Swim


      • #4
        It's a watch so accuracy will never be 100% satisfying(just think of calorie calculation), therefor you'd have to attach the IMUs inbetween shoulder blades or at your hip. But with a proper push off the wall and mesured velocity and time from velocity peak to it's next velocity peak (acceleration of the arm) you could calculate the distance.

        Even if it's not 100% accurate it would still be good enough to have at least some aproximate data for Lx and Mf calculation.

        You could also set a default parameter for your emerging points, but this would have to be one for each stroke. I know that my average ermerging points for freestyle and butterfly are 9 Meters and for Backstroke and Breaststroke 11 Meters for slow pace and 15 m for fast pace.


        • #5
          As far as I am convinced that the built-in accelerometer can detect a velocity peak, I seriously doubt if it knows the absolute velocity at any given moment. The simple sensor inside the watch is simply not precise enough to derive velocity from acceleration with any degree of satisfactory accuracy. And without it distance between strokes or distance from push-off to first stroke cannot be calculated.
          Current: nuvi 255, eTrex 30, Forerunner 310XT, Forerunner 735XT, HRM-Tri, HRM-Swim, SDM4, Virb Elite, tempe
          Retired: eTrex Legend, Swim


          • #6
            Agreed. While there are accelerometry-based devices that have now become reasonably accurate for land-based tracking of movement (position/velocity), they are typically used for straight motions such as walking. In theory a similar device could detect velocity and to a lesser extent, distance of wall push off (it may be difficult to determine the transition from glide to stroke immediately), but the cyclical motions of swimming and between stroke and individual variations would make it difficult to track other parameters reliably. So it is not just a feature issue, but also a hardware one. This technology could potentially emerge if 1) the right engineers and programers are on it, 2) there is constant input from a swimmer/coach, and 3) there is a sufficient market to drive that. Although swim tracking on these expensive devices is nifty, you don't see competitive swimmers with them often, and the technology has not evolved much from the Garmin Swim (circa 2012). Even updating the software to fit its capabilities (e.g., set parameters of what time range constitutes a length to reduce miscounting and the ability to edit length count errors as they happen, eliminate the need to hit the rest button) by doing #2 above would be an awesome start.


            • #7
              Well, here I am, your swim coach/athlete who could give constant input.

              I'm also a software engineer, working on my own Coaching Software, already capeable of reading/writing lenex properly and providing competitional analyses. Next big thing would be using .fit files for unifying training documentation and do some reporting on training progressions. Since this is whole lot of sh*t work, when done by hand/excel.

              The reason why you don't see competitive swimmers with those devices is, that normally there's a coach at the pool, tracking parameters like strokes, minute frequencies, ... But if you don't have a coach all the time or you're a coach with a big group, it's simple impossible to track data for all athletes. So we loose all this precious data.

              And yes, you're right. Since the garmin swim, there hasn't been any progress on the swim tracking software, except for the hr tracking.

              So spread the love for some improvements! Hell, it's about time!


              • #8
                It is true coaches provide most of the function of a watch, and swimmers don't need a lap counter or timer in most instances since they are adept at using the pace clock. They also know how many strokes they take. At the moment the cost of these watches is prohibitive for many teams (especially age group and HS swimming). I expect that to change eventually. I haven't rained with a team for decades, and use the watch mainly as backup, sometimes to look at splits in GC, and most important to me, to track open water distance when I get to swim outside (the only time I would miss not having this watch). In the mean time someone with your skill set and intuition of what is needed could write a killer IQ app to utilize the hardware's simplistic functional capacity. Even something like Peter's Pacer for running would be useful for swimming to know at a glance if at or below pace. And for those who swim in crowded places and want accurate lap counts, the ability to input rules for when a lap is a lap rather than slowing down or speeding because of other swimmers (e.g., minimum and maximum lap time). But the extra features you seek are a dream we should hold onto.


                • #9
                  Ahh! The Connect IQ API, never thought of that so far. You're absolutely right. I'll se what I can do and will keep you up to date here. Thank you!


                  • #10
                    Please do -- it would be great to maximize functionality for swimmers.


                    • #11
                      So let me give you a brief update. Eclipse is set up for the API and I'm now able to speak monkey C. Also I did some research with the acceleration unit, and let me tell you that you have no idea how accurate it is working. I still have to do some field testing and recording data, but I think, everything I've planed to do will be possible. Testing will happen in March. I'll keep you updated.


                      • #12
                        Awesome. Thanks for continuing the quest!


                        • #13
                          So these are the first accelerometer charts. Values already in m/s². Still a lot of work to do. At the moment I have just written a WatchApp, which is able to record accelerometer data. What you can see in the charts, is a push off with some seconds of freestyle swimming.


                          • #14
                            Hey neymo_aut, any progress since your last update on this subject?