-
For more information on how to avoid pop-up ads and still support SkiTalk click HERE.
Teaching with CARV
- Thread starter Erik Timmerman
- Start date
I wouldn't be too sure about that... it knows what percentage of your weight is on the outside ski vs the inside, where the pressure is inside your boot (toe vs heel), it has an accelerometer... ok, I see you've found that on the webpage just now
Speculation:
Inertial measurement unit generates location and orientation of the device over time. You can then plot the path traveled on the snow as well as fully define the orientation of the unit. Then you could look at the angle of the unit along that trajectory (ie the heading vector should be tangent to the path in a pure carve). If the angle of the unit relative to the path changes over time, skidding happened. This assumes the sensor unit is well fixed to the boot or ski.
Edit: you'd probably need some filtering to remove slope angle/gravity influence. I think it is possible to determine if turns are skidded. The algorithm failing for undulating terrain would make sense.
Here is a screenshot from my Carv app, which confirms it knows the exact path I'm skiing:
Another screenshot where it shows me each turn I made and it's grade for one specific run:
Another screenshot with just a small sample of the data it's collecting:
Another screenshot where it shows me each turn I made and it's grade for one specific run:
Another screenshot with just a small sample of the data it's collecting:
Here are four important characteristics of a turn and my guess how Carv could estimate the performance.
Start the turn before the fall line: This should be fairly easy since Carv has both pressure sensors and accelerometers. If the forces are high early in the turn this should give a higher score
Balance on the outside at will: Also pretty easy to determine by Carv, but this is still a tricky one because there are many cases where it is totally fine to have inside pressure, hence the "at will" part of the statement. How can carve know if it was intentional or if it was a balance problem?
End the turn facing down: This one is probably quite hard to estimate by measuring at the foot
Accelerate up or down, always: Easy to measure if the turns have high forces, but what about slow speed turns on mellow terrain, probably not so relevant then. Also what was the intention, and furthermore did the snow conditions require a lot of acceleration or was it hero snow which allowed a more static CoM height?
Start the turn before the fall line: This should be fairly easy since Carv has both pressure sensors and accelerometers. If the forces are high early in the turn this should give a higher score
Balance on the outside at will: Also pretty easy to determine by Carv, but this is still a tricky one because there are many cases where it is totally fine to have inside pressure, hence the "at will" part of the statement. How can carve know if it was intentional or if it was a balance problem?
End the turn facing down: This one is probably quite hard to estimate by measuring at the foot
Accelerate up or down, always: Easy to measure if the turns have high forces, but what about slow speed turns on mellow terrain, probably not so relevant then. Also what was the intention, and furthermore did the snow conditions require a lot of acceleration or was it hero snow which allowed a more static CoM height?
Given the wide variation in what people consider "better" skiing, this is a topic I'll only touch on lightly.
At the moment, I can safely say that Carv definitely tells me what I'm doing poorly, and also seems to accurately tell me outside/inside pressure throughout the turn. Carv's edge angles are also consistent with video evidence and, frankly, feel.
But as with my earlier post, you can be doing something right according to Carv but miss its relative contribution to your overall ability.
To me, that's just one example among an infinite number where better Carv numbers can become uncorrelated with whatever you're trying to improve. Into those gaps step instructors.
But if you know what you need to work on (and how to work on it) and it happens to correlate well with a Carv metric, then having it monitor your progress is incredibly appreciated.
I mean, the name kinda gives it away ....
I have wondered why Carv didn't call it "Carv IQ", which would've made so much more sense to me and be more accurate to its intent.
But perhaps they've socked away enough run data to hope to use Ski IQ for other situations.
Last edited:
The same way NOAA make weather predictions from sparse, highly non-granular, data: by creating mathematical models which correlate well with historical outcomes.
Only Carv attempts to gather this type of data for skiing, but just off-hand: while gross foot pressure alone may not mean much, that combined with the distribution of forces (say LTE or BTE), accurate accelerometer numbers (centripetal forces throughout the turn), and position on the slope over time could be sufficient for determining edge angle by regressing those covariates against a set of edge angles determined by reviewing video of expert skiers carving.
The real question is what diversity of ski techniques have contributed to Carv data models. Over on the unmentionable forum of a certain technique, its founder stated that Carv did not seem to correlate well with his approach, so right off the bat Carv may not be able to help train using that method. (I feel this was a missed opportunity).
But really, we’re only just scratching the surface of utilizing live data of skier dynamics.
I think I would’ve enjoyed being an early Carv adopter who could’ve played with that level of data ….
Last edited:
Yeah, unless a Carv run was started and stopped in the range where you're really focused on carving well, 100% will drills screw up your scores (watering down by averaging).
Interestingly, while my greatest weakness remains moving forward at the start of the turn, I had been working hard on leveling to move my COM over my outside ski (assuming that's what your student was working on). Those pressure metrics are almost maxed out if I'm "carving well" — but while my video shows me apparently carving, any skier worth their salt will immediately cringe at the back-seatedness.
One problem at a time. As a old-school noodler skier, I was overjoyed just to get separated and feeling my outside ski edge, though it has to be said that was accomplished before I got Carv (though Carv verified it).
The real test now is if Carv can help me gain control over my fore-aft balance ....
I have a similar challenge with fore-aft balance! Looking forward to trying to develop it...
I got a little color on Carv's fore:aft ratio and how it relates to actually being forward/back throughout the turn.
Carv says I have poor-to-middling forward balance at the start of the turn. Consequently, I thought that the Fore:Aft monitor would help me achieve better, well, forward positioning.
Turns out the Carv Fore:Aft monitor just tells you how centered you are (e.g. 45-60%), not whether you are forward or backward at the right point in the turn.
I found this out when I reached out to Carv's customer support because I was confused why my "Start of Turn" metric was so terrible (25-50, rather than >60) despite my Fore:Aft monitor numbers being very good (keeping it between 55-60).
Carv first sent me this chart of the "ideal" forward/back pressure when plotted against the turn:
You see how the "Intermediate" skier only moves forward well after the middle of the turn? (Likely because they're ramming into their skid and thus being forced forward).
Carv then sent me a similar chart of my own balance performance on the same disappointing run I was asking about — and it was wonderfully illuminating:
Using the Fore:Aft ratio monitor, I "learned" to maximize my numbers in the middle of the turn, in my case by projecting my rib case down the fall line (scary!). You can clearly see from my chart that I was "succeeding" at doing just that: max forward pressures in the middle third of the turn.
THIS WAS WRONG. I can achieve forward balance just fine, but I do it in the wrong part of the turn.
(I'm unsure why there's that sudden dip in forward balance in the middle of the turn. Perhaps someone here can hypothesize about that).
So while Carv correctly identifies the problem, it does not yet provide either the necessary charts or the monitor to help me improve being forward at the start of the turn. Even the suggested (and otherwise superb) video from Tom Gellie really only addresses the end-of-turn backseat positioning.
That was disappointing, particularly as I spent much of my only two days on the slopes (without my kids!) trying to improve just this one aspect of my skiing, only to find out Carv never provided me the tools for success.
Last edited:
Having said that: Carv support has clearly been very forthcoming with more in-depth interpretations of the data, and has outright stated that while the developers consider adding them as default visualizations to the app I can always contact support to have them re-plot the data on any given run.
So just bear in mind: Carv is still very much a product in development, with tech support clearly enthusiastic to help customers understand the numbers.
But (to bring it back to this thread's topic): maybe consider booking that lesson if Carv says you're still not doing it right!
Last edited:
can you trust the edge angles in the app? it suggest I was doing 70 something degrees.
When I was teaching I got a consistantly pretty low score, when I was just showing off some random "show off turns" it rewarded me, but they just looked cool, wasnt very good turns. (artificially high edge angles)
When I was teaching I got a consistantly pretty low score, when I was just showing off some random "show off turns" it rewarded me, but they just looked cool, wasnt very good turns. (artificially high edge angles)
When I was practicing two-footed releases the edge angles were low (10-20°).
I also got high edge angles with otherwise bad form.
From this I actually came the conclusion that Carv edge angle measurements are at least consistent, and why the composite Ski:IQ number better reflects a well-carved turn.
Regardless, I’ve been thinking of ways to double-check Carv edge angle numbers by filming my skis right at ground level with something like a GoPro or 360º camera.
The consistency of the Carb edge angle number means I haven’t been all that motivated though.
You need to study Bob Barnes' Infinity move. Notice that forward pressure in the boot is an effect and not a cause.
https://www.skitalk.com/threads/the-infinity-move.277/
Notice how far behind the CoM the skis are, both down the hill and along their path (the skis take a longer path). That is what's giving the skier forward pressure in the boot.
The jaggy pressure spikes are a consequence of your trying to do with muscle action what BB's ideal skier does with body position. You pressure, you rebound off the cuff, you re-engage. Perfectly normal, but as you already noticed, it isn't what's being attempted to be coached.
Perfect, thank you very much.
I suppose this speaks to an advantage of Carv: it provides yet another avenue for a learner to convey objective information to even a remote instructor. That's pretty powerful.
Carv could be a useful toy for a very knowledgeable, self aware skier. We've already ran into issues of less experienced skiers really not understanding what to do with the information. One was pretty crazy.... wanted the toe of the boot 7mm higher than the heel, and insisted the boot be planed accordingly. Hmmmmm.
Therein lies Carv’s dilemma: if they reveal too much of the (very noisy) data they run the risk of burying themselves in the criticism of confused customers.
Yet by restricting access to data, they run the risk of losing more data-aware customers for whom the curated visuals are insufficient — like me.
I personally think they should consider building an API that allows users to create their own metrics and share them, like the way the iPhone exploded in popularity after anyone could write their own apps. That would really let us explore the true potential of Carv.
In the meantime, the saving grace is how eager Carv appears to be to engage with users. That’s the key to any product success, and I’m glad that’s a focus for them regardless of whether I keep my Carv or not.
Last edited:
There is so much you can get from CARV for the price of one lesson from a pro, IMO. CARVs are not magic, they're a tool like anything else. Most of us have had some good lessons in our day, where something clicked from something a good instructor said (or, as I think is often the case, "when the student is ready, the teacher appears." ) Most of us have also had a lesson where there were no breakthroughs, and maybe even a not very good lesson, or one that we just weren't able to take much away from. They still cost $150-300 regardless of whether you got anything from them
As said above, CARVs are a tool for the curious student who's also learning from other places. They are a guide for things to try to improve one's skiing, some instant feedback on whether the changes you've tried making are making your runs better or worse (note, you have to be curious enough to figure out whether you're in some local maximum or not). As almost everything as an adult, improvement is a self-directed journey. CARV invites you to try some things and see what they do for you, and gives you some feedback on whether that new behavior gets you closer or farther from an idealized carved turn. It's part of a balanced diet.
Sponsor
Members online
- Living Proof
- DoryBreaux
- Matt G
- SpeedyKevin
- ctsnowplow
- Jerez
- John Fritz
- BLiP
- Philpug
- KingGrump
- MOMENTinTime
- Johnny V.
- Nomax
- BeiJiHu
- dusty
- Crank
- Mendieta
- mikes781
- Brian Finch
- Wade
- ev_steve
- salvatore
- Dean
- dbostedo
- GregK
- tromano
- vociferoussky
- Dislogical
- wooglin
- David Edwards
- Chris V.
- Cheizz
- Neurodrive
- Doug Briggs
- coldski
- Plowester
- Mazama
- Gapak
- scrubadub
- Tricia
- BullMcCabe
- DanoT
- Jack skis
- dan ross
- raytseng
- 4aprice
- tomahawkins
- David Chaus
- angstroms
- deadheadskier