Strava has made some welcome changes involving Zwift segments in recent days. And while most of the comments on this post will probably be from Zwifters complaining about the company’s recent API TOS changes, I still want to take the time to share this totally unrelated news, since it’s something I’ve been asking for for years!
Dive into the details below…
New Verified Segments
Earlier this month, Strava announced their new verified Segments. They say verified segments “help highlight the iconic stretches of road and trail where the community competes. Strava inspects segments from our segments database, identifies which are the most representative of the road or trail, and verifies them. This feature aims to reduce confusion and improve the accuracy of the world’s most popular segments.”
Real outdoor segments will benefit more from verification changes than Zwift segments because popular outdoor roads tend to have lots of segments that are similar but with slightly different endpoints, confusing names, etc. Zwift segments, on the other hand, are mostly only created by us here at Zwift Insider, so you don’t have to deal with the segment duplication issue.
These newly verified segments are basically the most iconic segments on Zwift. Strava asked us to send them the top 10 segments to be verified, so I worked with some folks at Zwift HQ to select the top 10 based on segment popularity (stars, goals, completions) and other factors.
Without further ado, here’s the list of Zwift’s first verified Strava segments, with my comments:
- Alpe du Zwift: by far the most starred Zwift segment, plus it’s the game’s most iconic climb. No question it’s gotta be #1.
- Volcano KOM: surprisingly the 2nd most starred segment in game!
- Epic KOM: 3rd most starred.
- Mont Ventoux KOM: 4th most starred.
- Box Hill KOM: 5th most starred.
- Hilly KOM: 8th most starred, and where it all began on Watopia. Seems fitting.
- Tempus Fugit: second-most starred full route on Zwift, and most attempts of any route in Zwift in last 365 days.
- Fuego Flats Forward Sprint: we figured it would be cool to have one sprint segment in the list. This is the most starred and most attempted segment in the last 365 days.
- Volcano Circuit: 7th most efforts in the last 365 days, and most starred of any route on Zwift!
- PRL Full: the last segment was the toughest to pick, but we thought it would be cool to verify the longest route on Zwift on the list.
You’ll see a checked badge next to the name of any verified segment you complete will show at the top of your list of PRs when viewing ride results.
Read more about verified segments on this Strava support page >
New Short Segments
If you’ve followed this site for long you may have noticed me complaining about Strava requiring segments to be at least a certain length. When Zwift first began, that minimum length was quite small – somewhere between 100-200 meters. Then a few years later, Strava bumped it up to 300 meters, meaning we couldn’t create a Zwift segment for a 290-meter sprint.
Then sometime in 2022, Strava increased the minimum again, to 500 meters! That’s when I got really frustrated and wrote this Zwift Insider post as well as this post on Strava’s community hub. Why was I frustrated? Because there are a lot of important short segments needed on Zwift, and minimum segment lengths don’t make sense on virtual platforms where GPS data is always accurate.
Then a few weeks ago, I found myself on a call with a couple of folks from Strava and Zwift. I asked if there was some way they could give Zwift Insider’s Strava account special permission to create shorter segments, and a few days later… they did!
I quickly got to work, creating all those missing short segments that had piled up over the years. Here’s are the short segments I’ve created since then:
- Alpe du Zwift Bend 20 to 19
- Alpe du Zwift Bend 19 to 18
- Alpe du Zwift Bend 17 to 16
- Innsbruck Sprint
- Innsbruck Sprint Reverse
- Leg Snapper KOM
- Stoneway Sprint
- Stoneway Sprint Reverse
- Acropolis Sprint
- Acropolis Sprint Reverse
- Sasquatch Sprint
- Sasquatch Sprint Reverse
- Woodland Sprint Reverse
- Jarvis Sprint
- Jarvis Sprint Reverse
- Tower Sprint Reverse
- Castle Park Sprint Reverse
- Castle Park Sprint
- Boardwalk Sprint
- Boardwalk Sprint Reverse
- Tidepool Sprint
- Tidepool Sprint Reverse
- Shisa Sprint
- Shisa Sprint Reverse
- Country Sprint
- Country Sprint Reverse
- Village Sprint
- Village Sprint Reverse
- Yorkshire Sprint
- NY Sprint
- NY Sprint Reverse
- Lutece Express Sprint
- Champs-Élysées Sprint
- Ballon Sprint
- Marina Sprint Reverse
- Broad St. Sprint Reverse
- Monument Ave Sprint Reverse
- Breakaway Brae Reverse
- The Clyde Kicker
- Champion’s Sprint
- Bell Lap Prime Sprint
- Downtown Dolphin Prime Sprint
What’s Next?
There are still a few things I’d love to see Strava do to improve their experience for Zwifters. Here’s my list, in order of priority:
- Automatically Remove Ridiculous Efforts from the Leaderboards: this is a tough nut to crack, but it would also provide the most value for Strava and Zwift users. Segment leaderboards currently contain a lot of spurious efforts, and these need to be removed. Read more >
- Overlay Zwift Map for Zwift Activities: ten years after Zwift launched and thousands of people were riding on some island in the middle of the Pacific, the situation hasn’t changed on Strava. Our maps still show the real-world islands, and not Zwift’s maps that could easily be overlayed on top. Read more >
- Enable Zwift Segment Backfilling: if you create a new IRL Strava segment, their system will “backfill” the segment with any matching rider activities from the past. But this doesn’t happen on Zwift. When I create a new Strava segment, only Zwift rides that happen on that same or later days will match (with some occasional and odd exceptions). Strava says this is because it’s too big of a load on their systems to backfill with so many Zwift activities on a small number of roads.
- Disallow Segment Flagging: users can currently flag Zwift segments as hazardous, which is just silly. Virtual segments shouldn’t be flaggable as hazardous. Read more >
Most of these were shared with Strava contacts on recent calls, and while I can’t share where exactly Strava is at with each issue, I’m hoping we’ll see some movement on them soon!
Questions or Comments?
Share below!
And did you ask them about the API changes? Just kidding, thanks for your work for the community!
Your #2 “Overlay Zwift Map for Zwift Activities:” is by far my most wanted update. The sooner we can stop cycling in tbe sea and see the real Zwift maps on Strava, the better !
Zwiftmap for Strava is a Chrome extension that does some of what you’re after – for some reason it doesn’t work in the main Strava feed, but will for individual activities: https://chromewebstore.google.com/detail/zwiftmap-for-strava/eiaekjobfimlijhpggbbkhoihlchdhdl
Great news about the short segments!
Agreed, this is great news!
Is it just me, or are all the elevation profile views on those verified segments now broken? The alp is downhill, volcano climb is blank etc etc
Yes, it looks like it’s broken. It seems to be the real-world terrain on those coordinates.
LOL, carefully “verified”. Oops!
Yeah, I’ve reported it to Strava. We’ll see what they say…
Hi Eric! Any news about the buggy Alpe du Zwift segment? I reported it as well in the support forum a few days ago. No answer so far.
One more wanted update: Automatic matching of new segments on old rides. Strava does that for real-life rides, doesn’t it?
Correct, they do not. But I just added it to my list, because hey… it’s almost Christmas!
That’s probably one of the only wish-list items Strava has an “okay” excuse for. That *is* a lot of data to back fill. But it should still be possible.
They already match every new outdoor ride against every one of the existing millions of outdoor segments. Matching Zwift rides against a comparative handful of Zwift segments pales in comparison. Yes, it’s a lot of data, but computers are pretty good at processing lots of data.
In some ways I agree, Jonathan. But the problem with Zwift roads is the number of activities on them.
With IRL, you can quickly filter out the vast majority of efforts when backfilling, because there are relatively few efforts on any given IRL segment. Because not many people live/ride in any one particular IRL place.
Zwift broke all that, because Zwift’s virtual roads get way more traffic than any IRL roads. So when Strava goes to backfill, they have to look closely at WAY more activities than they would with any IRL segment. It’s that level of matching that they say is too resource-intensive.
That’s not to say Strava couldn’t make it happen. They could. But backfilling one Zwift segment does, on average, take a lot more CPU cycles than backfilling one IRL segment.
I don’t disagree. But it’s a one-time hit, on just a very small number of segments. Given the unique way Zwift segments are created, it hardly ever happens, so they would only need to do it very sporadically (and mostly for new segments on new roads — the present case of creating new segments on long-existing roads is kind of a sui generis situation).
And I don’t disagree. 🙂 I’ve tried to pitch that thinking to Strava in the past, but so far no go…
My guess is, their systems aren’t easily tunable, or lack limiters, and processing a Strava segment perhaps overwhelms things…?
I know it’s probably too esoteric of a fix and may have some quirky technical impediments but one solution might be to only kick off a backfill for individual users who actually ride the segment AFTER it gets added…would probably depend on how the whole shebang is indexed under the hood, but at least in theory they would not have to chew through the ride portfolios of every user who ever existed, just the ones who display a current interest in the segment (by riding it). Yeah, it means if someone crushed it 5 years ago and then quit Zwifting they’ll never show up, but given that it’s all imaginary anyway, I’m going to claim that doesn’t really matter – and individual users who want to see how they stack up against their own history will be able to.
Good news!
Is there a way to match old rides on these new segments ?
Another aspect about this which is tricky, with the Zwift major pack dynamic changes over the years, the speeds have slowed quite a bit, making it effectively impossible to challenge many of the older KOM/QOM times for a lot of courses.
Along with all the changes they’ve make to surface rolling resistance…
Thank you. Finally someone posts about how the pack dynamic changes have rendered KOMs and loop times obsolete. This is my number one gripe with Zwift. Coming from a running / racing background, they really missed the boat with this. No one wants to know you went up the segment 30w more with a slower time recorded. Loop times and KOM were an important motivating factor for me and some of my friends. Really love the platform but you really hit the nail on the head. To challenge some of my times is impossible.
One update i would like to see implemented into Zwift is the ability to pause a long ride, London PRL Full for example, and resume where the route was paused.
For many it is probably not possible to complete that route in 1 day, and not be able to leave the PC/Laptop or other device powered on to avoid turning off Zwift.
However, to be able to do that route over 2 consecutive days, while able to power down Zwift would be useful.
Congratulations on a most excellent troll. I was thinking who’s the lazy $#!+ then just started laughing,,, Might have been better with 3 consecutive days though…
This has absolutely nothing to do with being lazy. Not every Zwifter has the time to do such a long ride in one day. Family, work or any number of other commitments can prevent someone from riding for any more than say 3 or 4 hours in 1 day.
And as for my post being considered a troll post…it’s far from it.
Then i would suggest riding shorter routes.
OK. More than one mindset on Zwift. Some not at all like mine. 😁
If I have to stop and come back the next day, I personally don’t consider that completing that route. To me, that’s the essence of the challenge of a long-distance ride.
I’ve done IRL and Zwift centuries. If nothing else, I don’t want to contaminate my Eddington number. https://en.wikipedia.org/wiki/Arthur_Eddington#Eddington_number_for_cycling
Perhaps my mindset has been brainwashed by (Strava shared) events that have timed starts and finishes, For example, I did a 10K turkey Trot run today. There was an option to do it virtually. But it won’t be accepted to do 6K of it on Thursday and 4K another day.
(There was also a 6K Turkey Trot option for those who didn’t want to do the 10K. Plus a 6K walking registration for those who did not want to run it.)
Ah – that explains my confusion at getting a couple of PBs last night on a couple of the southern coast road sprints even though i hadn’t really sprinted them! I hadn’t actually realised they weren’t segmented already – they always feel a bit longer than a true sprint!
Would love to see a segment for the Mayan mountainside KOM and Itza KOM combined since so many routes have these back to back.
That’s a good call. I’ll do it!
Wow, thank you so much.
New pack dynamics are off the charts….so much harder to stay with the bots…riding for years on zwift…why do they have to keep making drastic changes rendering loop times and draftable koms obsolete ? Any thoughts Eric ?
Thank you, Eric! To your point 1 IRL strava clocked me over 90 mph pulling my daughter in her trailer.
Hello! Thank you very much for your work. It was so nice to see all these new segments! However there are errors for the segments “Alpe du Zwift Bend 17 to 16” and “Alpe du Zwift Bend 20 to 19” which are triggered during the descent. Could you please correct this? Otherwise thanks again for everything.
Hmmm… you sure about that? The segments are correct – you can see the elevation profile has them traveling uphill. If they were accidentally created on the descent that wouldn’t be the case.
Yes look
Weird. No idea why Strava would do that.
It’s not how they were created, it’s the radius that strava use to match the start and end points. It’s needed outside because of GPS drift, people’s paths don’t snap to the road. You do see the same thing on tight switchbacks in the real world.
Thanks for your work on this, glad they finally got the short segments working (for you at least, the rest of us are still out in the cold for IRL rides, but that’s a battle for another day).
I’d propose a fourth item for your list:
4) Automatically match new segments against existing rides. This happens automatically for real-world rides & segments, but it’s very annoying that it doesn’t work for Zwift rides. I know you can hack it with the crop-ride workaround you’ve mentioned previously, but realistically nobody is going to go back and do that for hundreds or thousands of Zwift rides.
Thanks for advocating for the Zwift community, Eric! I noticed the new sprint segment results in some races last week and wondered what had changed.
I’d like to see Strava reduce the radius for segment matching on virtual rides – there’s no gps drift in Zwift, they can be much more precise than the real world.
And maybe add a decimal point to the time, these short sprint segments will end up having hundreds of equal KOM holders if it’s to the nearest second.
Or the best solution of actually working together and getting the segment times directly from zwift.
Top work that man, thanks…
@Eric Schlange
Alp sections not working.
I think what worked in the past is to have two short turns together.
Ex. 20-18 and 18-16
Otherwise, the path tends to blend together, resulting in incorrect readings.
Tell Strava they should allow VeloViewer to analyse Zwift related segments, as well as all IRL segments.
It all sounds good, but, until this “Automatically Remove Ridiculous Efforts from the Leaderboards” is sorted, Strava segments for Zwift are basically useless.
I disagree, because most of what I use Strava+Zwift for is tracking my own PRs on various segments.
When you’ve got hundreds of thousands of people competing on a segment, I don’t really see the point of paying attention to where I’m at on the leaderboard. Maybe if I was a top elite racer that would be different…
Fair and good point. I suppose it’s just the leaderboards on Strava for Zwift segments that are useless. The segments themselves are still useful.
Even if I’m nowhere near the top (I am 4th on one, though!!), it’s nice to see where you fall in the scheme of things. And it’s a shame that the leaderboards for the segments we all know and begrudgingly love are messed up beyond meaning.
I like to track where I am as a percent, like top 10%, top 5%, this is useful info. It doesn’t really matter that there are fake efforts at the top if you look at it this way.
That’s how I use it as well. Sort of wish there were a way to set the default to “show my results” instead of having it be “all time results” and I have to scroll down to “my results” every time I open a segment. All I want to do is see how much more slowly I did today’s effort than my best. Don’t care how fast elite riders are doing it. Maybe it’d be different with an IRL segment that only a few people do (and that I compete with others for). On Zwift, that’s not something that is going to happen.
Super nice work Eric. Link to Monument Ave Sprint Reverse is broken.
Ah – it was marked as private. Made it public so now the link should work for ya.
I did not. It’s in the list above! https://www.strava.com/segments/38147800
good work!
Can we get Strava to not rotate some of the worlds by 90*? Makuri and France come to mine especially. They are rotated in Strava compared to how they look on the Zwift overhead as well as on ZwiftGPS.
I just joined Zwift and recently did epic + radiotower and was really surprised there was no such segment there was however epic kom, radio tower a smaller bit of epic and an unverified epic+radiotower+descent
Thanks very much for the (re)new)ed short segments. That’s very helpful.
Another thing at least the verify Alpe du Zwift segment perhaps other segments too show incorrect elevation gain and gradient (223m and -0.8% for AdZ)
This is now fixed.
What good does starring a Zwift Strava segment do? Will they actually somehow pop up on your Garmin while going? I can’t see how that would happen given you’re not outside / no satellite…
Thanks for your work, Eric. Backfilling would be so good. Work your magic, please!!!
Great news, especially on the shorter segments – if you don’t (keep on) ask(ing) you don’t get! 👍
Can you please consider adding a descent segment from the RADIO TOWER because this is so often decisive with the kickers on the downhill and yet there is only a descent from the Epic and that doesn’t register anyway when coming down from the Radio Tower in the forward direction as it’s past the banner.
A public version of my private stand in here https://www.strava.com/segments/32530991
would be great.
Thank you for all of that work, Eric. Next time a group of segments are made could you add Temple KOM? Again than you for all your strava effort.
Temple KOM segments were made back when Makuri Islands was first released!
I rode up it three times yesterday. It isn’t there. Only Sea to Tree. Weird.
Can you share a link to your ride on Strava here? I’ll check it out if so.
I just checked on my computer. Temple KOM is now a hidden effort. I hadn’t checked before. It doesn’t show up in the app. My error. Thank you for following up.