For the Polar Grit X as outdoor-focused watch, one of the major (new) features is the Hill Splitter function, meant to track and help analyze uphill and downhill running.

Let’s have a closer look!

As Polar explains it:

Hill Splitter™ tracks your performance on the uphill and downhill sections of your training session. You get detailed data about the elevation profile of your session, and you’ll be able to compare hill stats between different sessions.

Hill Splitter™ automatically detects all uphills and downhills by using your speed, distance and altitude data. It shows you detailed insights into your performance, such as heart rate, distance, speed, ascent and descent, for every detected hill on your route.

Polar

Nearby, I definitely have nothing higher than a hill. There is a track that makes for decent ups and downs, though, a bit like a stadium stairs climb, so let’s see!

I will also want to have a look at this function in the mountains, but these small elevation changes make it all the more interesting to see when uphills and downhills are detected!

The Altitude Profile

Here is the altitude profile for this run, doing five laps on the same course:

Easily visible that there are two longer ascents (at the beginning and in the end) with two extra hills in between, then the “long” and relatively constant downhill back to the bottom.

What does Hill Splitter make of it?

Hill Splitter During a Run

While running, the Polar Grit X uses one dedicated screen of the training views for Hill Splitter.

This screen shows the distance and ascent/descent on this uphill / downhill section (or flat section), the current pace, and the count of uphills/downhills. (Flat sections are also shown as such at the bottom, but not counted.)

Polar says that it takes an altitude difference of 10 meters (or 15 meters in sports such as skiing and ski touring) for the watch to count something as a hill; they do not mention over what distance.

Hill Splitter is said to adjust to the current training session and count small ascents/descents on flat runs, only larger ones on activities with greater altitude differences.

On this course, there are some short and low ups-and-downs in between where the Grit X only “sees” a flat course, but it counts 20 meters of ascent over 140 meters (after such a “flat” section) as an uphill already.

This becomes clearer – or at least easier to see – from the display in Polar Flow on the web.

The “All Hills” View in Polar Flow (Web)

Training session analyses for Grit X recordings are basically just like those for a recording with a Vantage – except for the “All Hills” data into which it can be split, alongside “Laps,” “Automatic Laps” or a “Split into Laps”.

To make things clearer, I have zoomed into one of the laps (the first one) and marked the different hill sections as Hill Splitter segments them:

Hilly Flat Splitter

This makes it clearer that the up-and-down of the two hills in between the distinct uphills is shown by the altimeter – it has enough ascent/descent to be recognized by that.

Thus, it is shown in the altitude graph as two “hills” – duh.

But, it does not go far enough in distance and ascent/descent to trigger the uphill/downhill count of Hill Splitter.

At 300 meters in distance, it is counted as long – and distinctly flat enough – as opposed to the ascent before and after (of 360 m of distance and 30 m ascent, and 140 m and 20 m ascent, respectively) – to separate these two (as) uphills, however.

Similarly, the 12 to 40 m (depending on which split one looks at) at the top are enough to be separated out as a “flat” section in between the last uphill and the downhill.

Constant Downhill

The downhill section – as visible from the altitude profile – is so constant that each downhill was counted as one such section.

Thus, each lap here got counted as two uphills (with flats before, in between – or they wouldn’t be separated – and after) and one downhill.

Split Hill Distances

Interesting also – I find, especially on such a short track – what data is recorded per lap and how that compares.

One, in distances:

Distances per “phase” are somewhere in the same ballpark, but quite a spread.

• Uphill 1 was measured, on average, at 357 m, with an average deviation of 14.7 m. Distance measured ranges between min 330 m and max 379 m.
• Flat 1: 307 m, deviation 8 m.
  Min 294 m, max 322 m.
• Uphill 2: 152 m, deviation 12.8 m.
  Min 135 m, max 169 m.
• Flat 2: 29.6 m, deviation 11.7 m.
  Min 12, max 41 m.
• Downhill: 680 m, deviation 26.6 m.
  Min 648, max 747 m (though that max was on the last lap and counted the 100 m back to my start/end point from the point where I turned around in between laps before).

Well… This could use some statistical analysis over more data.

For example, the “Flat 2” is the section with the two “hills” that are shown by the altimeter, but not enough for Hill Splitter to separate them.

Meaning, it’s a short section with a strange (up-and-down-and-up) course; it’s understandable for that to result in a large deviation in measured distance (and a 3D distance could change things again – and get confused).

Split Hill Ascent/Descent

Hill Splitter also looks at the ascent or descent of uphills and downhills, of course…

This is a nice example for how much more sensitive a barometric altimeter is, compared to GPS for distances (let alone altitude, as far as I know).

Averages and average deviation here are hardly worth mentioning – or all the more:

• Uphill 1: 29.8 m ascent avg, deviation 0.7 m
• Uphill 2: 22.2 m ascent avg, deviation 0.6 m
• Downhill: 52.4 descent avg, deviation 0.7 m

Comparing ascents and descents per lap and overall finds the same result, a difference of 3 m at most in the individual values, ascents and descents that are only 1-2 m different…

Below, I’ve also pasted together the altitude profile view from Polar Flow, zoomed into the different laps (as visible from the slider).

Things to Check and Note

One big question about Hill Splitter will be if there is much use to it outside of training situations.

Things might well end up looking very different in the mountains, over various times – but, for example, because weather changes could mess up the (barometric) altitude recording.

I wonder also how sensible it will be for a mountain trail run to get a separation into uphills and downhills.

Still, from this hill repeat, I’d say that I like this feature more than I thought I would – for novelty and a bit of insight.

For analysis, it is good – and this is where Polar’s strong suit lies.

Between Polar Flow web and app, the potential for analysis and the help their system(s) give with phased workouts, training plans, and all that, there is a lot to like, look at, and make good use of.

For a mountain ultramarathon, in conjunction with navigation, it does not yet look/sound like Polar will offer anything much. There, the other well-known sports/outdoor watch brands have their advantages.

But, let’s see what is offered and what, perhaps, will come.

Oh, if you are interested in a comparison between the data recorded on this run from the Grit X versus the Vantage V, you can find that here.