Zwift’s August 2022 update included the welcome addition of a Legsnapper KQOM segment to the Innsbruck map. As any Zwift racer will tell you, the ‘Snapper is a crucial climb that has broken the will and legs of many, so it only seems fitting to make it an official segment.
As keepers of the official set of Zwift Strava segments, we’ve had several inquiries from observant riders who have noticed that the new in-game KQOM doesn’t perfectly match the Legsnapper segment we created years ago when Innsbruck launched.
To be precise: the in-game segment starts a bit later than our Strava segment, and ends a bit earlier.
To be honest, we weren’t at all surprised to learn the segments don’t match perfectly. While Zwift had reached out to us for details on the start and finish of our Strava segment so they could make the in-game segment match, our original segment was simply our best guess as to where the official climb should stop and start, and we figured Zwift might see things slightly differently.
When the update dropped, we went in and recorded a ride so we could make a new Strava segment matching the in-game KQOM. And that’s when we ran into a roadblock.
Minimum Segment Length on Strava
When we clicked the button to create the new Legsnapper segment we were greeted with an error message: “Please make your segment longer. It’s currently too short to be accurately timed.”
For the past few years, Strava’s minimum segment length has been 300 meters. This was a big change from the early days of Zwift when the limit was 100 meters, which explains why short segments exist for older sprints like the 180-meter Watopia Reverse.
We were surprised to see this error pop up, since the Legsnapper segment is well over 300 meters long (430 meters according to Zwift’s leaderboard).
So we did a bit of digging, and it turns out Strava recently increased their minimum segment length to 500 meters. Which is a bad move, in our opinion. Here’s why…
Reason #1: This is Zwift
In Strava’s support post about Very Short Segments they explain that one key factor forcing the need for a minimum segment length is GPS accuracy – specifically, the sampling rates of GPS devices. “Some devices only record a GPS point every 5 to 10 seconds,” Strava says, so having a segment that only takes 30 seconds to complete could lead to lots of inaccuracies.
Fair enough. But this is Zwift. Strava knows the difference between a virtual ride and an outdoor one. Virtual platforms record GPS coordinates very consistently (Zwift does it every second) and can therefore accurately and consistently measure times on shorter segments.
Reason #2: Roads Go Uphill
Strava’s simple length limit is a bit of a bodge if you think about it, because it doesn’t bring gradient into the question.
If the real issue is how long it takes to complete a segment (and that is the real issue), Strava needs to do more than look at segment length. They need to look at segment gradient and compute an estimate of how long it would take a rider to complete the segment. That 500-meter downhill section? Maybe it’s too short for a segment. That 300-meter turn-to-turn section of the Alpe d’Huez climb? Probably plenty long.
Incidentally, this is why we were able to create turn-to-turn segments for most of the Alpe du Zwift, but not for the shortest segments (bends 16-17, 18-19, and 19-20). We all know those segments take significant time to complete because they are ridden at slow speeds. But Strava doesn’t care.
Mountain bikers are particularly up in arms about the new 500-meter limit, because as any MTBer will tell you, a 500-meter uphill segment is no joke. Especially on difficult terrain!
Reason #3: It’s Just Us
Zwift Insider’s Strava account is the only one allowed to create public segments for Zwift activities. Strava doesn’t need to let the entire world create shorter segments. But if they trust us enough to maintain the official segments for Zwift, they should be willing to flag our account and let us create shorter segments.
We’ve repeatedly reached out to our contacts at Strava, though, and have heard no reply.
Wrapping It Up
While we understand Strava’s reasons for limiting the creation of short segments, any implementation of segment length limits that does not take gradient into account is simplistic and over restrictive.
Additionally, the key reason for those limits (GPS accuracy) simply does not apply in the virtual land of Zwift!
We would really like to see Strava roll back the virtual segment minimum to 100 meters so we can create segments for Zwift’s new Legsnapper KQOM as well as the various sprints and Alpe sections. What do you say, Strava?
(A link to this article has been posted on Strava’s support forum. Please visit, comment, and Kudos the post to help get Strava’s attention.)
Questions or Comments?
Post below!
Strava doesn’t do engineering anymore. They just slap on band-aids. Yes, it’d be trivial to have the “short segment” detection work on the average time to complete the segment – but that’d take work. Easier to just do a simple length calculation and wash your hands. Similar to how they “fixed” the virtual segment issue (if there even was one) by DELETING EVERYTHING and permanently locking out everyone from creating segments. Similar to how they “fixed” the Flybys issue (if there even was one) by disabling it for everyone and letting it die a silent death. Similar to how they… Read more »
I sure wish they were listening. Would love to see Strava segments for all the segments on Zwift (even the super short sprints, but definitely the AdZ segments) because Zwift doesn’t show the old records and I’d love to have them somewhere to compare to old results.
your 100% Right, strava got it wrong here. i hope they can work it out. Love King Bath salts.
One way to fix this is if Zwift had their own overview of segments in the companion app or on Zwift power somehow that would work the same way that Strava does it. (This would mean lifetime PRs to be visible as well)
Yep – go for zwift segments – and order with those zwifters you follow
Zwift does remember sprint times – you just need to expand this?
Regards
Huw
I think Zwift is actually moving toward this, with their upcoming Holographic Replays and also the route-based workouts feature that is being worked on.
If Zwift does this well then there’s not really a need for a Strava version of the same data in my opinion. Would be great to see!
Whatever you do, please don’t edit the old segment or delete the old segment, either way we lose all our old efforts. Thanks.
Eric I ripped off all the safeties hung and accelerated into the downhill in a planned effort hanging beyond what was previously possible for me during part of an A race as a B. I may or may not have partly died and shifted into a different metaverse during those seconds I can’t know. A coach I used to have believes everyone’s 100% is the same effort/pain level I disagree. I have an opinion about not somehow deleting that effort as counting. In fact, lots of decisions had to be made (pens to banner for instance) on segments that count… Read more »
Adding insult to injury is the fact that interpolation is not hard to do at all. If you have one recorded data point just before the segment start, and another one just after the segment start, it’s simple geometry to interpolate quite an accurate segment start time, enough to give even sub-second precision to segment times. But Strava has shown zero interest in implementing this, over more than a decade.
“the Legsnapper segment we created years ago when Innsbruck launched”. According to https://zwiftinsider.com/verified-zwift-strava-segments/ the Legsnapper segment was created 9/14/2020. And according to https://zwiftinsider.com/innsbruck-launch/ Innsbruck launched in 2018.
If Strava are so reliable on accurate data, why dont they monitor segment times more closely. Recently found a QOM that recorded an average speed of 87 Mph to take the crown. This had been un-noticed for 2 years. Obviously the person was in a motor vehicle not cycling. It seems that Strava have lost the ability to accurately monitor segments at all. 😳
How is this a surprise at all? You can’t possibly expect Strava to keep an eye on every segment on their system… that’s an impossible job. I always wondered why anyone would bother making a segments for km 14-15 on the Ventoux for example… who gives a damn? You could make a case for two smaller segments from bottom to Chalet Reynard and from Reynard to the top but no people just make one big hot mess of all sorts of random bs segments, and then you expect Strava to actually monitor every single one of these backwards segments no… Read more »
490 mts of road that is at a gradient near 10% could easily be used as an everest, for which it would be useful to have a segment for.
Strava is lazy when the issue is accuracy. The GPS coordinates and times may be exact but the interval between recorded points depends on the device, sometimes much more than 1 second. Strava fails to account for the effect of large intervals. The problem is most apparent when you see a KOM on a segment in which the average speed exceeds the maximum speed. What was that again? The average speed exceeds the maximum speed. Strava computes the average speed using the time between two points divided by the nominal segment length. However the start time may be well after… Read more »
Best of luck with this. Strava shows as much interest in actually taking feedback from their forums as I do passing on ice cream…
One minor correction: there can be differences when different games include the same climb. This isn’t generally a factor in the fictional Zwift worlds, but with real world courses, like Innsbruck, there are Magic Roads on RGT and likely video-based routes on platforms like Rouvy. So precision/accuracy is technically an issue, but 500 meters is way too long. 100 meters would be more than enough.
Ahh ok. I KOM’d the segment a couple of times in recent weeks but Strava didn’t recognise it as a KOM and didn’t even update my PR so maybe this is why…?
Admittedly with no evidence, just life experience, I’d like to suggest that money changing hands has something to do with this. Maybe Strava is holding out and using this as a bargaining chip with Zwift. We’d never know. I’m no engineer but it seems like it would be so easy and obvious for Strava to include all the Alpe du Zwift segments. Personally, think the current limitation makes Strava look petty.
Voted for the strava ticket now, and published in my local cycling forum to.