View data on Google sheets here >
The data above is for lap times ridden in isolation (no draft/veering effects) at constant wattage on various Zwift courses.
Hints:
- Use the “Showdown” filter in the Google sheet if you just want to see which setups are the fastest
- Use filters and column sorting on to narrow your dataset as desired
- To see how differences in weight, watts, and height effect lap times, first filter by “32mm carbon” wheels on the Richmond data, then short by the parameter you want to test
Strava Segments Used
When these tests were run we hadn’t yet created the “Zwift Insider verified” Strava segments. Here are the segments used for the above times:
Richmond
- Full Lap: Richmond 2015 UCI Worlds Lap
- Flat: Zwift TT Flatout (3rd Str)
- Libby KOM: Richmond KOM
- 23rd St KOM: 23rd Street Climb
- Downhill 1: Downtown Downhill Stretch
- Downhill 2: High Speed Straightaway
Watopia Figure 8
- Full Lap: Watopia – Figure 8
- Flat 1: ODZ Down Under
- Flat 2: Ocean Road TT reverse
- Reverse Climb: Zwift Island Climb (Reverse)
- Forward Climb: Zwift Island Climb
- Reverse Downhill: A decent descent
- Forward Downhill: Sqwiggly Wiggly
great work – very comprehensive.
I would love to read your summary thoughts
Summary thoughts can be found here: https://zwiftinsider.com/tests/
This is awesome work. Thanks for taking the time to do it and gifting to us!
Interesting……And to think I was just choosing my bike/wheel set-up based on what I liked the look of.
Eric – in the latest update v1.0.9778 the Buffalo has weight increased. Weight now 5140gm and CdA of 0 (as opposed to 3000gm and -0.005). Which in itself is a bit of an odd one. As the original weight far exceeds the other bikes stripped weights (TT 450gm), which suggests then that its the CdA that has the biggest effect and not the weight. It would be great to rerun the Buffalo test with the new cDA to see the difference.
Testing it now! Thanks for the updated numbers, Clive.
By any chance do you know where the Specialized S-Works Tarmac falls on the speed scale?
Sure–you can find it in the numbers above (check out the “Showdown” tabs). Basically it’s not terribly fast–Zwift Aero is better. It’s between the Trek Emonda and the Zwift Carbon. It climbs decently, but isn’t very aero so it gets dropped on the flats.
Amazing data! What about the Zwift TT with 50mm carbon wheels?
I didn’t do a lot of test laps with the TT, because it’s obviously the faster bike. You can look at the different wheel speeds (for Richmond laps, choose 225, 183cm height, Zwift Carbon for bike) and extrapolate how they would effect laptime for each bike. Difference between 32mm and 50mm carbon wheels is about 5 seconds over 1 lap on Richmond.
Funny that the Tron bike is such a hard thing to get, yet it is slo much slower than the Cervelo or even the zwift aero (its slower than many bikes actually). You’d think it’d be the fastest bike out there because it is so hard to achieve it.
I think it’s just because the Tron bike’s wheels are its cool feature, but it’s upgrading the wheels that makes you really fast. The Tron bike itself is really fast (only 2 seconds behind the fastest, the Cervelo) but that’s without wheel upgrades.
You can’t have your cake and eat it too!
Is it possible the Tron bike has had a wheel upgrade? I just did 5 laps on Watopia (straight laps, not figure 8) with my Kickr set in ERG mode at 200 watts. Not quite as scientific as you, but should still have given me a performance idea. The only thing I changed was bikes. Lap1 – Tron, Lap2 – Tron, Lap3 – S5/808’s, Lap4 – ZwiftTT/808’s, Lap5 – Tron. The first two laps with the tron bike were 20 seconds faster than the S5 and marginally faster than the TT bike, both with 808’s. So I did a 5th lap switching back to the Tron bike just to verify. And it was still faster.
So could it be the lower wattage of 200 that made the difference, or has the Tron bike actually gotten an upgrade? Also wondering, but haven’t verified, if the tron bike might use whatever wheels you were using last. They look pretty deep dish to me.
Would love to see the difference between TT and road bike on the new mountain extension. Both for the KoM section, but also for the whole lap. Theoretically the weight savings on the climb should mean faster climbs for a road bike, but that doesn’t seem to be reflected on the shorter climbs of Richmond and regular Watopia. If there is a time saving on the climb it would interesting to see if it’s wiped out by the aero advantage of the TT bike on the descent.
I agree. Another vote for mountain extension testing please!
Really interesting data.
What makes no sense is that the times are THAT identical at 400w and 500w. That shouldn’t be the case, especially the climb times. The 100kg rider should, for instance, be dropping at least 15-20 seconds to the rider up the 21st street climb following Libby Hill, and at least 30-40 seconds up Libby using real world model and assuming .4M @ 6%.
So the real interesting question here is why aren’t these difference being seen in the tests? Something has to be VERY off how Zwift is calculating speed, either in general or certainly over 400w, in order to obtain results like yours.
Might have to test this for myself with various weight settings.
What does it take to get zip 808’s??
Hit Level 10!
Just curious how does the Specialized Allez compare to the Zwift Carbon, and likewise, the 50mm vs the 32mm Wheels (I do see that the 50mm made it into the chart a couple of times).
Thanks
Allez and Zwift Carbon are basically identical in speed (see Google sheets data).
50mm wheels knock about 5s off of a lap of Richmond compared to 32mm.
Big question I (and many others I know from their comments) have on group rides with wattage targets: what’s the w/kg adjustment to the target wattage for riders of varying weight on flat sections to maintain similar speed?
I used your data for the Richmond “Zwift TT Flatout” section using most common height (6 ft), bike and wheels (zwift carbon) to hold those variables stable. I’ll try to email a shot of the full data/results/graph but for instance for 25mph (common speed for these rides) 100kg (340 or 3.4w/kg), 75kg (300 or 4w/kg), and 50kg (250 or 5w/kg). For 23mph, the values are 100kg (285 or 2.9 w/kg), 75kg (240 or 3.2 w/kg), and 50kg (195 or 3.9 w/kg). (Note: I did have to visually interpolate 3 missing values on the chart to create a smooth line).
What does this mean? Assuming the ride leader is calling out w/kg for a 75kg rider, on the flats, a 50kg rider would bump the watt/kg by 20-25% (e.g., 3->3.7w/kg) and a 100w/kg rider would decrease them by 10-15% (e.g., 3->2.6 w/kg), or we can guestimate our place in that range for other weights. The absolute wattages listed appear high perhaps due to most being shorter and using aero frame/wheels, and most importantly for drafting in the pack. As you found once wattages get to 400 (8w/kg, 5.3w/kg, and 4w/kg respectively) the figures collapse, but those are far above most wattage-based rides. However perhaps we can all make sense of crazy numbers for our weight by remembering a VERY rough rule of thumb with lighter riders adding ~10% and heavier riders subtracting ~20% from the ride target w/kg on flat sections.
Since w/kg matters more than absolute wattage uphill (though as you found not entirely), the higher the rode tilts upward the more lighter riders should drop their w/kg towards the target while heavier riders increase their w/kg, both approaching but not quite reaching the target ride wattage.
Hey Eric, any idea if this data is still accurate since all the updates the past year? I noticed that in the menu to select your bike (the star system) Zwift identifies the S5 and the Madone as being equal. Did Zwift actually modify them to be equal or is that just a rough guide to hello people choose a bike?
It still is, Jeremy. If you click the link at the top of this post for the Google sheets, you can see I’ve added a couple tabs (sheets) for recent tests. I ran these just to confirm that times are still accurate, and they are!
Thanks for doing all these test I find the data very interesting. Am I going craze or was there data for different rider weight as well?
Wondering how the new Meilenstein Lightweight fit into the mix of best all-around/climbing wheels on Zwift.
Is that what is meant when “lightweight” is put in the wheels column on the Alpe tab? Can we get a writeup on those somewhere?
Been doing test runs up the Alpe all week… hoping to have a post up today about what setup gets you up the mountain fastest.
Eric – have you found any aero differences with helmets / gloves other pieces of kit? Just wondering if it’s been tested.
I know it’s in progress still, but should the P5 read 34:41, rather than 35:41 for the full lap? Thanks for your work!
Yes! Good catch.
Possible to add filter and sorting buttons in the 2019 tab?
Thanks for all of the excellent information, both on this topic, and everything else that is available on Zwift Insider! With all of these new bike frames, I’m having trouble keeping track of which ones are road bikes, and which ones are TT bikes. Or, more specifically, which ones can draft, and which ones can’t. It would be great if that could be added as another column in the spreadsheet at some point, in order to help identify the fastest TT bike, and the fastest draft-eligible bike. The “price” and “Level Req” columns from https://zwiftinsider.com/drop-shop-release/ would also be cool to have in the speed test spreadsheet… but if just one could be added, I’d love to see a Draft/NoDraft column. 🙂 Thanks again for all of the time that you dedicate to this web site!