Yesterday, Zwift made the unprecedented decision to pull Gravel Mountain’s new Red Rock Loop as the week’s featured Pas Racing series route, swapping it out for Watopia’s Jungle Circuit.
They’ve also changed the plan for stage 4 of the series, swapping out Red Rock Loop Reverse for Watopia’s Serpentine 8.
It’s a surprising move from Zwift, who have been teasing this week’s Gravel Mountain launch for weeks. But was it the right move?
The Bugs
On Monday morning, April 6, I walked upstairs to my office and fired up the ol’ YouTubes so I could watch the very first Gravel Mountain races and see what riders thought. I was excited about the launch – I had already seen the new route, having “pre-ridden” it using my bot on a pre-release build of the game so I could create a Strava segment for the new route.
Here’s my bot riding the course at 4 w/kg:
Watching the bot on the route, I had a few thoughts:
- The map is gorgeous. The rocks and scenery are epic and beautiful.
- The road is still too wide – and this is something I said about Gravel Mountain years ago. One of the cool things about gravel riding is that feel of going fast on narrow roads. Portions of Gravel Mountain feel much too wide, and this makes the road feel more boring and less speedy.
- It felt fast, because Zwift was pulling some sort of trickery to make the course move past more quickly than normal, while keeping your actual speed realistic. It didn’t feel overly fast to me – but I also wasn’t experiencing the increased speed of racing in a pack.
Overall, my takeaway after bot-riding the route was that racers would enjoy the experience, and that it might bring something new to the world of Zwift – the fast-paced gravel crit.
So I was genuinely surprised at what I saw in the livestreams of the first gravel races. Here are the two streams I was watching, from Bike Bonk Biff and Road to A:
The first comments of these two streamers are pretty telling:
“Woah! That was a fast start! What? What was that? What the hell is this? That’s surely a bug, right? Or…?” (~Road to A)
“Oh! OK. This is actually… why is it so fast? What? What is happening right now?” (~Bike Bonk Biff)
Watch the streams, and you’ll quickly see what I saw:
- Too fast: Speeds were much higher in the pack than on my solo ride (for example, Road to A’s speed was 10 kph higher at the end of the initial s-curves than my bot).
- The bouncing: While the rider’s own avatar didn’t seem to bounce on screen, other riders’ avatars intermittently were, and the effect was not good. Not good at all.
- Erratic movements: Riders were shifting back and forth left to right too quickly, and sometimes flying up through the middle of the pack unrealistically fast.
- Riding through rocks: riders were moving through rocks and bushes on the edge of the road in places.
- No rider placings: The game stopped showing what place the rider was in (on the right-hand rider list) near the start of the first lap.
- Negative feedback: Feedback via in-game messaging was instantly negative. I’m sure Zwift has all those messages stored on a server somewhere, and if they were to make a wordcloud out of them all, the biggest phrase would probably be “wtf”.
Zwift’s forum quickly lit up, with one rider starting a topic that has received over 100 replies in 48 hours.
To top it all off, riders reported not getting credit for finishing the stage, including the promised 1000 XP bonus.
Zwift Pulls the Plug
Just after 9am my time on Tuesday, April 7, 24 hours after the first Red Rock Loop events were held, Zwift’s event guru James Bailey posted this on the forum topic linked above:
Based on early feedback to the faster experience on Gravel Mountain, we’ve decided to make route changes for Stages 1 and 4. There will be a more detailed update coming shortly.
A few hours later, Zwift’s VP of Product Mark Cote wrote this:
Hey Zwifters,
As James mentioned, we’ve changed out the routes for PAS stages from Gravel Mtn as we address the issues some experienced. Crediting for Stage 1 completion is being handled as well and all should see credit for completion by EOD today.
W/r to Gravel Mountain: The ‘issues’ were especially notable for high powered, non-steering paired users on specific refresh rate devices…ultimately this led to a few people having a very bad experience with camera swaying and some on-screen jitters. The rest of us had the desired (sped up gravel) experience we were shooting for. We did test this on several occasions internally but missed testing in the high-speed, large packs that were apparent during the first events. Yep, obvious in hindsight, but here we are and we’ve adjusted the events.
Gravel Mtn WILL be back soon and tuned via additional testing. There’s something to the feeling of speed that we’re all quite excited about and we will dial this in. We definitely want to be trying new, fun things out but we missed the mark on the quality that we strive to deliver.
Thanks as always for the commentary and reporting here. It’s super appreciated and valued by our team.
Mark Cote
What About the GC?
At the time of this post, Zwift hasn’t determined how to handle the time-based GC for the Pas Normal Racing series. It wouldn’t make sense to mix finishing times for two different courses, but it’s also not ideal to force riders to race again on the Jungle Circuit the same week when they already raced Red Rock Loop.
My take: Zwift should just not count stage 1 in the overall GC of the series. That seems like it would be easy to implement on their end, and racers would understand having the stage neutralized due to a mid-week course swap.
I’ll update this post once Zwift determines how they’re handling the GC.
My Thoughts
First, I do think Zwift made the right call in pulling the plug on Red Rock Loop this week. The race experience was just too buggy.
It all seems so avoidable, though, right? Why wasn’t this caught in early tests?
Hindsight is 20/20, of course. But surely this could have all been avoided by Zwift testing the race course with a large enough pack of riders at realistic race pace. I’m sure Zwift’s QA department is taking notes in this regard.
I think most racers will agree when I say that I do hope they quickly fix the bugs, and launch a much-improved Gravel Mountain soon. I like the idea of a fast-paced gravel crit course, and I think Zwift can make it happen. I do, however, share the thoughts/concerns of Aoi Niigaki, who posted this on Zwift’s forum:
If Zwift is after a sped up experience then that can be achieved through good course design. One of the things I noticed with the Gravel Mountain stage is how incredibly wide the road was. Narrower roads, tighter corners, shorter straights would all improve the perception of speed without the need to artificially boost it.
It’s a dangerous road you are going down Zwift (which started with the accelerated subway exits in New York). If people get used to double speed races then riding at normal speeds will start to seem slow and boring. Power creep is a thing in games and needs to be carefully managed less you stuff it up completely.
Aoi_Niigaki
I’m not convinced Zwift needs to speed up the visuals in order for gravel racing to be fun. And I definitely know that narrow roads with more turns/shorter straights can make a course feel fast and engaging without leading to “power creep”. I’d like to see Zwift explore this, particularly for virtual gravel and MTB racing.
That said, I’m not completely against Zwift’s idea of artificially boosting the perceived speed (let’s call it “visual speed”) of a route for a particular gamified race mode. This kind of experimentation is important, and I like that Zwift is willing to give it a try. It definitely needs to be dialed back though, particularly in the faster race categories.
Your Thoughts
Did you race Red Rock Loop? How did it go? And do you think Zwift made the right call in pulling it? Share your thoughts below…
wpDiscuz
