Zwift made some significant changes to frame performance in their February 2022 update, and we’ve been running lots of tests ever since.
Today we’ve finally posted the updated numbers to the following pages:
Groupset Bug Fixed
The “overweight” groupset bug affecting bikes using the new Dura-Ace 9200 and SRAM Red eTap groupsets was fixed in February’s update. That means the following frames now climb significantly faster:
- Cervelo S5 2020
- Chapter 2 TOA
- Cube Litening C:68X
- Canyon Aeroad 2021
- Liv Langma SL Advanced Disc
- Moots Vamoots
- Dogma F
- Uranium Nuclear
All of these frames are now 19-21s faster on our Alpe climb test, and 1.5-3s faster on flat ground.
Oddly, the Specialized Tarmac SL7 was not fixed, even though it was (is) affected by the groupset bug. In fact, it actually got 5s slower up the Alpe! We assume this is an oversight from Zwift that will be remedied in the next update.
Other Changes
Other changes were included in this update, although Zwift isn’t telling us what specifically changed. Our theory is that someone at Zwift updated some of the data which combines frames with groupsets and weights, when they cleaned things up to fix the groupset weight bug.
Here’s what we’ve noticed though:
- Flat (aero) test results for other frames have changed – the Specialized Venge S-Works sped up by 1.5s, and the Felt AR sped up by 2s. The Tron is also consistently 1s faster.
- The Felt FR was made lighter, as we suspected it would after our test results showed it climbed worse than the (supposedly heavier) Felt AR. It’s a better climber now, but still not a standout performer.
- The Lauf True Grit, already a slow frame by all standards, slowed down even more, losing another 7.5s on our flat test (not no change on the climb).
Significant News
Looking at everything that’s changed in terms of frame performance in the latest update, there are three significant things worth mentioning on the road bike side:
- The Chapter2 TOA is now the best all-around frame, by a significant margin. In fact, we would say it’s the first real “Tron beater” in game, when paired with the ENVE 7.8 wheels. See Tron vs Top Performers chart for more >
- There are now 4 top frames which turn in matching flat performance: the Specialized Venge S-Works, Felt AR, Cervelo S5 2020, and Uranium Nuclear. They don’t match on climb tests (separated by 4s), but they do match precisely on the flats. The Uranium Nuclear, especially, is a great buy for lower-level Zwifters since it’s only level 20 locked.
- The Canyon Aeroad 2021 is back to being the top all-arounder for beginner Zwifters (level 10+).
Questions or Comments?
We still have a few pages to update, including the Drop Shop Buying Guide, Fastest Frames by Level, and various Fastest Frames posts. Until then, if you want current data we recommend checking out the Frame Charts and Tron vs Top Performers.
Got questions or comments? Post below!
Thanks !
What happened to the Scott Addict? It is not even on the chart anymore, in December 21 it was the top all around performer?
The Chapter2 TOA now tops the Addict.
3 sec faster on the flats and 1 sec faster on the climb!
https://zwiftinsider.com/charts-frames/
Eric, I know some clubs host Jr. Events. If we are going to grow future users and future cycling greats, can Zwift perhaps host Junior and Collegiate series Event series.
Not something I’m in charge of… you’d have to ask Zwift!
It’s interesting to observe that i have no problem with patch changes in for example League of Legends where bi-weekly changes drop certain champions from “pick or ban” to absolute dog-tier again and again, but checking Zwiftinsider and seeing that “frame x” is now the best every few months is a bit exhausting and i wonder why.
Also shoutout to the beauty that is the Dogma F, it was so close to the top yet so far and now there is another frame better than it 🙁
Question: If I’m doing a solo ride, maybe with a structured workout, is it faster to use a TT frame or a draftable frame, benefiting from whatever incidental draft might happen?
Hey Bernie – this depends on the density of riders around you. If there are enough of them, a drafting bike will be faster. But if you’re mostly alone, a TT bike is faster. Hard to really say where the “line” is, but I’ll say, for example, that on Tempus Fugit it’s generally faster to ride a drafting bike.
Interesting, thanks!
If OP is on a workout, drafting doesn’t take effect does it? Also I don’t think it should matter as watts will be the target, right?
You still draft in Workout mode: https://zwiftinsider.com/drafting-in-workout-mode/
Thanks as always for the effort involved getting these new tests out. Just updated my cheat sheet with fastest setups I have available and added several items to the wish list for future purchases!
what level do you unlock and how much does the Chapter2 TOA
426,000 Drops, level 18
Call my paranoid, but I’m gonna wait a bit before buying the Chapter2 TOA just in case zwift realizes a mistake and changing the performance.
Hmmm…interesting point. I just hit level 18 and I was going to immediately go in and buy it, but I’m still at the point in Zwift where drops are at a premium, so maybe I’ll exercise the same caution as you.
Me too. I bought the Dogma F and after did not have enough money to buy the disc wheels only to find out a few days later that it was not the best all-rounder anymore
Since I have the Tron, I’m going to wait until I level up enough to get the ENVE 7.8 wheels because that’s the only time I’d use it. I had planned on buying the Scott Addict last week, but I’m glad I didn’t waste my drops. I bought the Cervelo S5 2020 when it was the fastest on flats. I almost bought the S-works Venge when they nerfed the Cervelo, but now that the Cervelo is back, I probably won’t. Maybe, once I get up to level 42 (37 now) to get the DT Swiss wheels, I’ll get the Venge… Read more »
Hi Eric In lieu of testing bikes on all courses, how about a 60kg (because zwifters are mostly in the lowest skewed weight of adults!), 70kg and 85kg rider’s results on the flat and Alpe course with a set watts for a select frame so we could mentally plot the change of frame over frame and wheel changes? Depending on the results we could see the curve or straight line of each flat vs vertical. Then each person could suss their strength vs need for frames. Outliers would be like me – fat and no sprint – so it doesn’t… Read more »
It would be very nice of Eric if he would also test each bike with 3 different rider weights, 3 different rider heights and 3 different rider power outputs. I also would like to see a less binary test (not just pure flat and not just pure climbing) – something like “Big Foot Hills” or “Legends and Lava”. But then Eric would at least need to do 10x the testing.
There’s nothing stopping people from doing it themselves…
ive been doing it for bologna. But zwift keeps changing the bookends of aero and light weight equipment.
Hi Jason,
This might help you: https://zwiftinsider.com/rider-weight-speed/
I’ll look into doing a post that sort of brings all the variables together (weight, height, power, and bike frame choice) to show what sort of time gaps you might have between different bike options at your given power numbers + body size.
It’s not an easy answer – lots of variables. But it might help give you a better idea of the actual time gaps you’ll encounter.
Tron bike not really special anymore
It’s still your best option until level 35. That’s pretty special considering I got mine at 23 for zero drops. I guess the TOA is easy to get though at 18 and similar (for now anyway).
🙁
Still the only one that glows!
Awesome! Thanks so much for continuing to do this!
Didn;t they recently fix the groupset issue with the Cervelo S5 2020? If so, I would have thought that it would be faster than the Specialized Venge S-Works.
Yes, mentioned that above. They made the Cervelo match its old time – but the Venge also sped up to match it, as did the Felt AR and Uranium Nuclear.
Question, are the TT frame speed test results still accurate?
Haven’t retested them, except for the fastest (Canyon Speedmax) which is still the same. No new groupsets on those, so hopefully nothing changed!
I guess the thing to remember is that it’s the *combination* of TOA + Enve 7.8 that makes the best allrounder. That means you have to wait until level 39, and until then Tron is the way to go…
Even then… going by ZI’s tests, the TOA + 7.8 is 8 seconds faster on a (steep, hour-long) climb and 8 seconds slower on the flat. I think the Tron’s aero advantage would be more beneficial in most situations, so I’m still picking the Tron when I want an all-arounder.
Good point. Of course, the DS Swiss ARC 62 wheels are pretty good all-arounders too, and available at a low level unlock.
The best choice for you depends on lots of variables, of course. Are you contesting a sprint? Afraid of getting dropped on the hill? Etc…
With many Zwifters getting dropped on the climbs, going with something like the TOA plus lighter wheels may help them hang in there longer than an aero setup.
Another part of the reason I’m leaning toward the Tron is I’m (maybe?) on the lighter side at 72 kg and shrinking. Sometimes I find myself walking away from the group on climbs but have a bit of trouble keeping up on the flats, so I’ll benefit more from the aero advantage. The TOA + 7.8 would make more sense for a bigger rider with more raw watts who does well on the flats but could use some help on the climbs.
So it’s looking there are three frames that still have errors: the Felt FR CdA (used to be 50th percentile on the flat, now 95th), the Tarmac SL7 weight (used to be 10th percentile climbing, now 40th) and the Trek Speed Concept weight (currently the slowest climbing TT frame).
Yeah, those do seem like odd outliers. Let’s see if Zwift decides to change them!
Four. Also the Lauf True Grit!
What happened to the post about the methodology for these tests?
Not sure what post you’re referring to… I sort of explain the methodology on lots of different posts. 300W steady, 75kg rider, no drafting, on Tempus Fugit and up the Alpe.
Maybe they are asking about how a test is executed. How to do steady 300w? How to test without interfering from other riders? How to buy all bikes and wheels? How to not register test XP and stats?
The answers are: ANT+ simulator. Once logged into to zwift the computer the testing is done on is disconnected from the internet so no other riders show up. He has a test account that he’s used for tons and tons of rides, it has plenty of drops. The stats and XP register but only to his test account so it doesn’t matter, he only uses the test account for running tests, not competing against others.
Would the Chapter 2 plus disc wheels perhaps combine the best of both worlds?
It’s a great option if you’re looking for an aero setup that climbs better than the Venge. You lose 2.5s of aero performance but gain 11s climbing. A solid trade on a course with any significant climbs!
Do you know whether frame and wheels have any effect on drafting in a pack (with the obvious exception of TT bikes) or is this only influenced by rider height and weight?
Not sure. Too small to measure, I think.
If they DO, it would be super-minimal, the difference between one bike’s CdA and another.
Thanks, I wasn’t sure if the difference would be the same riding solo as in a bunch, but I should imagine it would be incredibly difficult to measure, giving the movement within any pack.
Would this therefore mean that, if in a group ride/event, weight should be prioritised over aero as the bunch/drag would be there anyway regardless of how aero the bike is?
I have sort of been leaning that way. Unless it’s a race where you’re competing for points in sprint primes or the bunch sprint to the finish or you get dropped and want to try to get back to your pack or are dropped and want to stay ahead of the chasers or…but group rides, probably.
Thanks Eric for the numbers! 🙂 Curious how the SL7-issue turns out. Last spring, when I began zwifting, it was the only frame (or one of two with the SuperSix EVO, but the only sub-level 30) with 7 stars including a 4 star rating on weight. What a waste of drops. In the meantime, its weight rating was lowered to 2, now it’s up to 3 but still messy timings.
Well now everyone and their mothers are going to be on CH2 TOAs! I purchased the Scott Addict RC early on and was in disbelief that the TOA with such a low level unlock (18) and drops (426k) could outperform the Addict by such a margin – I decided to wait a while before purchasing a TOA to see was it a bug that made it so good!? Seems not to be the case now, but could it be a bug?
Zwift should offer refunds for the frames and wheels they change.
Is it okay for S5-2020 to freeze in the garage? But is that true? This data is inadequate for downhill speeds. For example, Volcano Climb. It is not the top of the mountain that determines the result. It is a flat finish from the downhill. What is the fastest there ??
I have tron bike but I’m only level 24 and have bought the toa but dont know what wheels to pair it with anyone got any suggestions?
DT swiss
I love that you do these tests! One thing I have always wondered, however, is how much a pack/race dynamics impact “best” equipment option? For example, I have an event on Lutscher next week. Field will likely be small (~12-15). I suspect it will blow apart on the climb. Maybe a lead group of ~3-4. I could ride the Tron bike given it’s balance of aero/weight, but wonder if I should opt for Aethos/Alpinist wheels since the climbs will likely be where the race is decided. BTW, I’m just level 34 so aero wheel choices are limited. Any thoughts, Eric,… Read more »
I would go with Aethos plus Aero wheels like Zipp808. On the flats you can hold the pack, which can blow up on the KOM. That is the Most decisive time. But you have to come down the hill as well, so I would not go with lightest wheels. Plus: the difference in climb performance is not that high as the difference in aero performance.
Thanks for the reply. At first, what you said seemed to make sense to me. However, in looking at the data, I think it suggests otherwise. The 808’s are the most aero wheels available to me. If I pair them with the Aethos, the data points are 3051 and 2932. Now, if I pair those same wheels with the Chapter2, the data points are 3038 and 2936. So, I would only lose 4 seconds on an hour climb and I’d gain 13 secs on the flats. Now, if I stick with the Tron, it’s 3028 on the flats, so is… Read more »
24 second climb on an hour long ride? In that case your time to the finish line is determined completely by the pack you are with. Are you able to stay with the pack you want or are you getting dropped? If you are getting dropped then pick a bike that is most advantageous in whatever place you are getting dropped. If you aren’t getting dropped then you just need to save your legs as much as possible for the sprint at the end that determines your place. 10 seconds save/lost on an hour corresponds to 0.8 watts at a… Read more »
Hmmm… when I read this, I might ask myself if “marginal gains” is just a method of the bike industry to take our money out of our pockets… 🤔
Eddie, how about TOA & DT Swiss 62? Data points 3045.5/2926.5. Faster than Aethos & 808s on flats AND climbs!
Still, if you’re struggling to hold the pack on flats go with Tron. Depends on your strenght/weakness.
I’m planning to pair the Chapter2 TOA with the DT Swiss ARC 62 wheels since they top the all-arounder wheel chart (https://zwiftinsider.com/charts-wheels/). I’m not high level enough for any disc wheels. Thoughts?
I was wondering, if one can simulate the timings for any route with only those two timings for flats and climbs. For example, if you just add the both timings you would simulate a 46,8km, 1.068hm course. That mix is 50% of the flat timing and 25% of the climb timing. By scaling with one factor, one could simulate a route of any length with 2.3% elevation. Other example, if you mix 75% of the flat timing and 25% of the climb timing, that would be a course with 1%elevation, 40%flats+60%climbs would add up to a 3%elevation route. I am… Read more »
so… I’ve been testing out the TOA and reached the conclusion that as a ‘lightweight’ rider such as myself (let’s say sub-70kg), the TOA is too lightweight and not aero enough for use as an ‘all-rounder’. I find myself getting shot out the back on the slightest descents. I’m going back to an aero combo that climbs well such as the venge/454s (haven’t got the 7.8s yet). Descents are my downfall ergo I need heavier and more aero equipment. Even with all these charts and stats, individual rider weight and stature is still a major factor in in-game performance and… Read more »
Am I correct in thinking these updates aren’t (yet) reflected in this spreadsheet? https://griffinevo.shinyapps.io/ZwiftGear/?fbclid=IwAR0N3BV_VlqnicB0NsGcLjNSYghVC8Za1XQdOelgGl5xJKuCwhh1v4AHjoU
Wow, I purchased the Nuclear a while back based on its price (180,000 or in that area) and it’s speed on the flats with the hope that one day the bug might be fixed. I didn’t check the drop shop but assuming its the same, it’s probably the best performance for least amount of drops. I just used it for the opening stage of the Wattopia tour and it’s the first time I used it for an entire track. I even set one climbing PR. Feels good to make a good investment every now and then.
Thanks for all your work on this – is the data available? I’ve really appreciated the previous versions.
Which is now the faster frame, the Chapter2 TOA or the Canyon Aeroad 2021? Thanks