The odd-yet-fittingly-named WhatYumeziWe’reLost? Makuri Islands route was recently unveiled for the Makuri Madness ZRacing series, to be raced this week for the first time.
It’s a 17.5km point-to-point route that begins with a short climb, then takes riders over more rolling terrain before finishing on the infamous dirt Temple KOM:
With the route including some dirt sections, riders are asking: does it make sense to swap to a gravel bike for all or part of this race?
The first stretch of dirt on this route comes around 10.7km into the race. Riders begin the Temple KOM, then turn right onto the cutoff road which takes us to the other side of the Temple KOM for a descent back to the Castle area. While this is a dirt section where gravel bikes will be a bit faster than road bikes, it doesn’t make sense to swap from road to gravel when you hit the dirt at 10.7km because you’ll be switching back to tarmac around the 12.9km mark, for a paved stretch just under 2km long.
But then we arrive at the next dirt stretch: the full Temple KOM. The dirt begins around 14.8km into the route, and continues to the finish line (with the exception of some wooden bridge sections near the end). The finish line is the KOM banner in the giant mystical tree atop the Temple KOM:
Does it make sense to swap from a road to a gravel rig for this final dirt climb? The short answer is: yes, if you have the right controller setup. But let’s unpack this, because there are a lot of considerations that go into the bike swap decision.
Key Bike Swap Considerations
Time Savings
This is the first question I always ask when considering a bike swap. All other things being equal, how much faster can you complete the dirt section with a gravel bike vs road bike?
I did a quick test with our standard bot rider (183cm, 75kg, at 300W). Here are the Temple KOM times for a top-spec un-upgraded road bike setup vs a top-spec un-upgraded gravel setup:
- Road Bike (Specialized S-Works Tarmac SL9 with Princeton CarbonWorks Wake 6560 wheels): 6:54.34
- Gravel Bike (Cannondale SuperX LAB71 with Roval Terra Aero CLX wheels): 6:29.47
That’s a difference of almost 25 seconds – slightly more than the time gap I’ve measured previously, thanks to the new Cannondale frame and Roval wheels being faster than the older gravel setups.
Route Layout
It’s crucial that this route ends atop the Temple KOM. If it kept going and ended on a paved section later, we would need to consider switching back to a road setup, since we would be at a major disadvantage sprinting against roadies on our gravel bike. But the route ending on the dirt tips the scales firmly in the direction of the bike swap.
Brakes (Sort of) Required
Are you able to brake in Zwift? Doing so quickly today requires the older Zwift Play controllers, or the controllers on the Zwift Ride.
You can also brake via the pairing screen, but to do so you have to bring up the pairing screen and go to 0 watts. So the fastest and simplest way to stop is via manual braking.
Pro tip: Set “Braking Sensitivity” to 100% (see above), as this will bring you to a stop 2-3 seconds faster than the default 50% setting.
Time Cost and Fast Swaps
Of course, bike swapping has a cost, because you have to stop moving before you can swap. Since the race ends atop the KOM, there’s no need to swap back to the road bike. So the major cost of your bike swap is the time lost while stopping at the bottom of the Temple KOM, swapping, then getting moving again. (Of course, there’s also the cost of losing the draft. I discuss that below.)
Doing it quickly takes a bit of practice. But if you have braking controllers (see above), you can go from full speed racing on your road bike to fully swapped and moving ahead on your fastest gravel bike in less than 5 seconds. In fact, I did it today, on my way to an unimpressive 8th-place finish:
Equipping the Best Wheels
One thing that helps make this swap more effective is last week’s game update that saves wheelsets to frames. Now you can equip your ideal gravel setup beforehand, then race on your road setup. When it comes time to swap, just click that gravel frame, and it’ll have the wheels you last used attached.
Where To Swap
You don’t want to swap on the flat and fast start to the KOM, because your effort isn’t high there, and you’re benefiting from the pack draft at those higher speeds of 40+kph.
Instead, wait to initiate your bike swap until the group slows down around the big right-hand turn where the road pitches up to 5%. This allows you to derive maximum benefit from the pack draft at higher speeds, initiate a faster bike swap, and get back in touch with the pack sooner since they won’t be so far away once you get going again.
Losing the Draft
Assuming you’re in the front pack when you initiate your bike swap, you can plan on most or all of the pack going up the road without you as you stop to change bikes. (In today’s race, I think I was the only rider who stopped to swap!) This means you’ll lose the draft benefit of the pack until you catch them again.
This is an added “cost” to bike swapping, but there are two things worth mentioning that reduce this cost:
- You’re on a climb traveling at 15-20kph for most of the remainder of the route, so you’re not saving a lot of power in the draft.
- As mentioned above under “Where To Swap”, timing your swap smartly lets you stay in the draft when it’s most helpful and get back in touch with pack riders more quickly once you get moving again.
Bike Upgrade Status
This isn’t actually a key consideration, but I know people will ask, so I’m including it!
Many riders have a fully-upgraded road bike, and a partly-upgraded or un-upgraded gravel bike. Does it still make sense to swap in this case?
Short answer: yes! As shown in our Performance Improvements and Power Savings at Each Stage charts, a fully-upgraded road bike gains ~37 seconds on the un-upgraded version across an hour of climbing the Alpe.
Using some approximate napkin math, we can say that the Temple KOM is ~6 minutes long, or 1/10th of an hour, which means the fully-upgraded bike would be 37/10=3.7 seconds faster up the Temple KOM than its un-upgraded version.
So if you have a fully upgraded road bike and an un-upgraded gravel bike (using the frames and wheelsets I used in my tests above), the gravel bike will still outclimb the road bike by over 21 seconds.
Conclusion
Given all the considerations above, I’d say bike swapping on WhatYumeziWe’reLost? is a no-brainer, assuming you have access to quick menu navigation (best done via Mac/PC keyboard shortcuts).
Your Thoughts
Have you tried a bike swap on this route? Do you agree or disagree with my conclusions above? Chime in below!
You lost me at “Assuming you’re in the front pack when you initiate your bike swap”
bwahhahahha! Ride On! 👍
Ha ha! Well, if you’re not in the pack when the last Temple KOM comes, it’s even smarter to swap, because there’s no draft to lose…
How much worse is using a gravel bike for the whole race?
Much worse. Basically, you’ll probably be dropped from the pack, and start the final dirt climb minutes behind the front riders. You’d be better off swapping at the bottom of the Temple KOM, even with a clumsy AppleTV setup, than riding a gravel bike the whole time!
I did this race in the A-cat entirely with the fully upgraded cannondale gravel bike and I won! It was a small group, 15 riders and I knew they were not going to go full gaz on the first 2 paved climbs, so I stayed at the back, maximum draft, pushed sightly higher watts than the others, but nothing excessive. At the top of the 2nd climb (which is gravel) I could move to the front easily. On the final climb I accelerated 1km before the finish, driopped all but one and won the sprint to the line
I swapped today exactly where you suggested, however I rode an upgraded Project 74, which made a hard start. It was worth it and I made 4 spots up on the final climb despite giving up the draft for a moment!
I am a B in the Advanced pen, so i was definitely not i the front group.
The older Zwift Play controllers are king for this type of bike swap. Being able to brake makes it worthwhile.
Update: it worked like a dream. Stuck in the 2nd group of my race I quickly overtook them all and also some of the stragglers from the front group.
Eric, I used chatgpt for this route. Gave it my rider power profile. My bike unlocks. Etc. It said a smaller field stay with the road setup, but more riders / more draft, go gravel the whole route. It didn’t think the swap (as a B rider) was worth it. The time lost to brake/swap/start/no draft. I ended up going gravel the whole route and WON. It helped I hit my in year best 30 min power, but still I stayed near the front, the group slowed on the flat. And once we hit the gravel I was all in. The biggest disadvantage to road or a swap is if you are in a group where another rider on a gravel bike made the front group and conserved enough – they could dust you. Yeah you can swap and climb the whole climb a bit faster, but will you catch the other rider? Probably not. This isn’t a TT. Its a Zwift race. All or nothing 🙂
Interesting! I’d guess you had a slow group of B’s, or you’re a very strong B. Or both. What was your finishing time?
I totally disagree with ChatGPT, clearly. But it’s been drawing flawed conclusions from my data for years at this point… 🤪
That is a good point about the danger of swapping if there’s someone in the front pack already on a gravel bike. You would want to work to drop them early for sure.
Yeah it’s hit or miss. And it agrees with you too easily when prompting. Finish time was 31:15
For reference, my race finish this morning (in the video above) was 29:44… so about 90 seconds faster. And I was in 8th place, 22 seconds behind the winner.
I wonder if you could have held on to win in my race?! So many questions.
Anyway – nice work.
Man I did this race this morning and asked Chat last second, rode the gravel bike under its recommendation, and got the EXACT SAME finish time lol. Finished third in group B but could have probably won if I was paying closer attention to when the front two took off. Interesting stuff
According to zwifterbikes, for 65kg me, on the Temple KoM, my fully upgraded Specialised Tarmac SL8 with Wake 6560 wheels is 5 seconds faster than my un upgraded Gravel Cervelo Aspero with Terra CLX wheels. So a bike swap is definitely not worth it. Am I missing something?
Sounds like ZwifterBikes is missing something. 🤷♂️
“Sounds like ZwifterBikes is missing something.”
Wow… really, Eric? I don’t presume to judge who is right in this situation, but simply assuming that the other party must be missing something, especially when you are applying such flawed methods, is anything but good scientific practice.
First of all, you have incomplete information from Mike because he did not mention the wattage or rider height he used for the comparison on ZwifterBikes. At 65 kg, he is also a completely different rider from your standard rider bot. Your standard rider bot is not representative of the average male cyclist, at least if the minute-power, FTP and weight data on Intervals.icu (currently covering 86 624 athletes) is used as a reference. You are extrapolating from a somewhat “elitist” cyclist profile and assuming that the results apply to the broader cycling population. Furthermore, in what you call your “napkin math,” you compare Temple KOM with AdZ. Dirt versus tarmac? A 3.6% versus an 8.5% average gradient? With all these differences in rider and segment conditions, you can hardly expect to extrapolate the results that easily.
I don’t want to defend ZwifterBikes because I have no way of knowing whether their calculations are correct. But at least they are attempting to provide bike setup comparisons tailored to a specific rider.
A proper answer from you would have been: “I don’t know. The results might be different for your particular setup, but I haven’t tested that specific scenario.” That would have been far more appropriate than simply assuming that you did everything correctly and that the other party made mistakes.
It is remarkable how so few words can erode my trust in your work. I genuinely appreciate the work you have done so far, which makes this careless response all the more disappointing.
Zwifterbikes has always (at least until spring 2026) had trouble simulating Temple KOM. It claimed Tron was faster than a gravel bike both up and down that hill, and it still did half a year ago, while we know the gravel bikes are significantly faster (and it has been for years).
Thus, for every route that includes Temple KOM, I know I can’t trust Zwifterbikes.
(Disclaimer: They may have improved it this summer, of course. I haven’t tried it since about March, and I believe I sent them an email suggesting an update. I also did that a couple of years ago…)
+ follow up: I just tested, and Zwifterbikes still claims that Tron, SL8 og Aethos will all be faster than any gravel bike on routes like Kappa Quest or Spirit Forest (with lots of dirt roads).
I have seen multiple times, in small and big fields, that you need a gravel bike (not a bike swap, just start on that bike…) to do well in those routes.
Thank you for this information, anderfo! As I wrote, I don’t want to defend ZwifterBikes, since I have too little experience with it.
It’s also not about the result or whether a bike swap is beneficial on this route. Judging from the other comments here, many people seem to feel, based on their own experience, that it is.
It’s more about the attitude in Eric’s reply that got under my skin, especially considering all the flaws in his study that I explained earlier and that make the results apply to only a small subset of riders. When an author feels the need to respond to a genuine question, one can only hope for a better answer than simply “They are wrong” in harsher words. This is particularly true when the author’s own study provides no basis for assuming that its results also apply to the person who asked the question.
And I’m quite sure there are rider and bike setup combinations for which the difference between an upgraded road bike and an unupgraded gravel bike is not as clear-cut.
The good thing about the bike-swap topic is that Temple KOM is relatively short, meaning everyone can easily test for themselves whether the swap is beneficial with their available setup if they want to be sure.
Chris, what do you mean by “flawed methods”? Eric measured how long it takes to climb the Temple KOM for a bot 183 cm, 75 kg at 300 W. No difference in weight (especially just 10 kg), height, or power could possibly lead to such a difference in results.
Anyway ZwifterBikes shows that a rider with the same parameters on the same gravel bike and wheels would climb 20 seconds slower. Yes, he definetely has right to claim that ZwifterBikes is missing something.
Are you sure you have Gravel bikes ticked in your options on the side as opposed to just Road bikes?
Looks like you also need to set the level high, add lots of drops and then it picks up gravel bikes?
Saying “I don’t know” would have been lying or inaccurate though. Because I know ZwifterBikes is giving the wrong answer in this case.
That said, I certainly I don’t want to be prideful or have blind spots in my thinking, and I appreciate you voicing that I came across that way.
I WILL admit to being a bit annoyed with Zwifters regularly asking me why my conclusions differ from ZwifterBikes, when ZwifterBikes is, in the end, using my public data for their (sometimes flawed) calcs.
Good grief…
ZwifterBikes has several issues. A wonderful site, but a site that needs to be taken with a grain of salt. For instance, ZwifterBikes has the Dicut 65 wheels faster than the Hardron Ultimate 650’s on Mayan San Remo route. Per Insider, the 650’s are faster on the flats and climbs (beyond the margin of error), so how can the Dicut 65’s be faster. This is just one example as there are several above.
Keeping up with all that is Zwift is a lot of work and Zwifterbikes is a hobby and obviously isn’t keeping up with everything. Lots of new bikes, wheels, and routes are still not added.
Just a quick update from my side:
ZwifterBikes has now been updated to fix the issues mentioned here — including Temple KOM’s dirt‑surface behaviour, fixed‑wheel frames, and the June 2026 reference‑speed change.
The model now distinguishes pre‑ and post‑June ZwiftInsider test data and adjusts older components automatically so they compare fairly against newly tested ones.
ZwifterBikes relies heavily on the public test data ZwiftInsider publishes, and I really appreciate the work that goes into those tests. Keeping up with all the changes in Zwift has been a challenge, so I have built a new automated pipeline that should make it much easier to stay up‑to‑date going forward.
Thanks for all the feedback — it genuinely helps improve the tool.
Great work, Thor!
I was about 10 seconds off the main group at the start of the last dirt section, had one rider behind me by like 6 seconds. I was passed in my bike swap by the rider behind, but ended picking up 4 or 5 spots on the climb to finish 4th and only 17 seconds behind the rider that won on a road bike. I’d do the bike swap on the route almost regardless, but I’d also do the swap as soon as you hit the dirt.
I did the bike swap in the race today before seeing your post. I DID swap at the first dirt section different from your recommendation. At the time of my swap, I was in the peloton and I got dropped during the swap. After the swap, I was able to make up most of my gap during the remaining gravel section, but didn’t quite catch the peloton before the tarmac. However, on the tarmac section, the peloton eased up a bit from the effort on the gravel section enabling me to catch the peloton again prior to the turn back onto the gravel. As a result, that put me both in the peloton and on my gravel bike for the final climb. I think that is an optimal position to be in, so I would recommend considering the bike swap at the first gravel section.
Interesting! Your strategy adds risk due to the tarmac section and the possibility of being dropped, but as you say, the group is probably chilling there.
And it adds major reward as you can start that final climb in the pack, which is quite ideal. Hmmm…
I did the same thing today to got too keen and swapped the first time up on the bend burnt a match catching the group (they are going slower vs the final so doable IMO) sat in for the same watts down the hill then was in the 1:1 spot for the final climb with draft and no swap to do….
Two of us tried the “switch the first time up the hill” method yesterday in a C race and it didn’t go well for either one of us. We were not quite able to rejoin our group before the tarmac, which left us burning matches just to stay in reaching distance before the final hill. In the end, the other guy was able to pass a couple of the slower people in our group before the finish line. I was not. We both finished about 5-6 positions below our predicted finish according to VELO score. I’ll try the race again today or tomorrow and switch at the start of the final hill and compare how I do.
so if you don’t have braking controllers ( I just have zwift hub with virtual gears ), then there’s no quick way to brake? So if I was to swap, I’ll just have to let my avatar slow down, which would take a while longer than the 5 seconds. Trying to work out if it’s still worth it.
Try bringing up the pairing screen at 0 watts (see below). Keyboard shortcuts are your friend!
Luckily I’d read Mike’s post before I raced just before. I did the swap where you suggested, brought up the pairing screen but then you still have to quickly go out and back to menu, then swap? Anyway, I think it was successful. I was already a bit behind a strong group of four who made more time on me the whole way up. But, I was with another guy, as soon as I was riding again he was 7 seconds ahead, I caught him easy and beat him by over 40 seconds, at not much higher power output than him.
Incidentally, it looked like one of the guys in the front group did it too, and made up 10 seconds to finish second.
So all in all, I feel like it was worth it. Even just for the novelty factor of swapping bikes in a race.
I don’t know why you think the pairing screen brake doesn’t work anymore but it absolutely does, they only disabled the brake if you keep applying power. And yes I just tested it to confirm.
Ah, that’s a good callout! I’d forgotten they changed it (quite a while ago) from how they had first set it, where you had to unpair your trainer. Updated the text above.
As Mike said in another comment, the change was the pairing screen now requires you to stop pedalling before it will brake.
I did simple test of braking with the controllers vs the pairing screen. On the flats, riding at 35kph, the Play controllers took around 3 seconds to stop. Using the pairing screen took around 5 seconds. So, controller braking seems to be the best method. No idea if other variables impact this, but I would guess not?
The change made to auto braking was that it won’t brake if you’re applying power, not if you have a device connected as a power source. Stop pedaling, hit A, come to a rapid stop. I’ve never stopped using the pairing screen brake and yes I tested it again just now. Coasting downhill at 40kph, hit A, come to a rapid stop.
Double posted because my first comment didn’t show up for an hour and a half…oh well now everyone gets to read it twice
I did a swap last night. Took 7 seconds. I was in a group of around 10. Soon caught them, at slightly less power than they were putting out and beat all but 2 to the finish. I definitely think it’s worth doing. You can see the difference in speed to watts as you ride.
I did the bike swap last night, and did it as we hit the dirt for the first time. I was just getting dropped off the front of the pack so no real draft to lose. Plus, because I managed to get with a few riders for the tarmac bit before the full KOM I was able to tuck into their draft, use my Draft PU, and not have to over exert to stay with them. I didn’t progress well due to dead legs, but Gravel clearly seems a good choice for those who don’t have dead legs, and with the chance for draft on the tarmac I think it does make sense to switch on the first time into the dirt
I find it interesting as an intellectual exercise, but think bike swaps in races should not be a thing. Decide on a bike and race it from start to finish, just like you would outdoors.
That poses the question: would a “no bike swap” rule be enforceable in a Zwift race?
Except that I’ve seen pro riders swap bikes in IRL events, so I’m happy to swap…but wish Zwift would automate it with a “team car” like animation that also fixes the time lost for anyone who does it
But only in time trials, in regular races pros are only allowed to change bikes in case of a mechanical. However, not everybody does care… it’s unusual to swap for a different bike in a road race though.
On top of that, irl they can’t swap mid pack, so the time to swap and regain the group is much longer than on Zwift…
Lol, yeah, sudden yanking on the brakes and jumping off for a switch mid pack would be rather problematic 😂
Let me be the first to accuse me of hypocrisy – I executed the bike swap in the Makuri Madness race and it worked beautifully. I reeled in a number of riders who had just dropped me. 🤭🫣
I use AutoHotKeys to do bike swaps where you code in the keystroke sequence ahead of time and set an activation sequence (I use zzz) I have done one there in a race in the past. The change takes just over 1 second so the only penalty is braking and accelerating, so that sport is perfect since braking is quicker due to lower speed.
nice, does that mean going down in the list of bikes one by one, or can you jump to say a bike 2 rows down the list by an assigned number? how would the sequence look?
in the garage, the bikes are always in the order they were last used, so the top one is the currently used one, the one below is the the one used previously etc.
when the script starts, AHK presses “T” to open the garage, polls a pixels color where the bike’s button resides, it only turns to blue when the rider completely stopped. When the button becomes blue it gets clicked, then sent a Down to select the last bike, then Esc 3 times to get out of the setting.
There’s actually two active colors for the button, and some delay is needed inbetween the button presses but it can be very fast and reliable, i didn’t even get dropped.
AI can easily write it.
Yesterday, way before this article, I won my A race swapping from the sl9 to the cannondale, but no brakes so had to wait to roll to a stop.7th before the swap, 1st at the summit. Awesomeness ^^
Alrighty. Just did the whateveritiscalledwearelost race. Set up the fastest (non-upgraded) gravel frame and wheelset then jumped back on the Tarmac SL8 (upgraded) and those Lava wheels. Stayed with the front 3 riders to the turn onto the gravel, switched to the gravel bike at the first orange uphill before entering the bamboo forest and stayed on it the rest of the race. Results, I dropped back 7 seconds on the brake/exhange/accelerate portion, stayed steady at my threshold and gradually gained on the front 3 as they pushed watts to open gaps. Lost up to 3 seconds on the front 2 racers on the last bit of tarmac, but when we hit the Temple KOM for the last time, it was just a matter for when I passed for the win to the finish.
Finish
Well done, sir! “Knowledge is power!”
Thanks, Eric. Your details on the route and bike swap made a big difference. And Dr. Joe Noack dictating the pace at the start sorted the size of the front. The dude is strong. If he and Sumrak also made a bike swap, I wouldn’t have made my comment. 😂
Being a big fan of bike swaps I feel that everybody is focused on how long the bike swap takes and nobody mentions that the bike swap time is spent resting. Real energy has been saved. To that end I always swap when I am going slowly (braking wastes energy). Also for almost bullet proof one button bike swaps get a stream deck. The swaps take about 2 seconds after the bike is stopped, and there is no skill involved. The only failure I have had with it is when the mouse is not over plain scenery (menus, buttons …). I have a button that swaps with the second bike, and another that swaps with the third bike. I have one that swaps socks as well (lightning to confetti and back).
Regarding it technically you do waste energy and not save it. As you say, braking wastes energy and so does coming to a halt. You deacellerate and have to accelerate to get back to your initial speed. So from that standpoint every bike stop itself is wasting energy. However, especially in this route you safe a lot more energy going up on a gravel bike. So in total you do save energy.
aside from that, I like your thought about time resting a lot, even though for me, the couple of seconds don’t give a lot of rest to me. 😅
Any energy used for braking is wasted. Any energy used to coast to a stop is used to make the bike go forward (not wasted). Yes you do have to put energy in to get back up to speed, but if you coast to a stop on a steep hill you didn’t waste any energy, you just didn’t add any to keep up your speed.You actually use a little less energy for the distance you go because the wind resistance and rolling resistance are less when you are going slowly.
i would find it funny if such stages had a bike-swap station. on the course coming closer to the bike swap station you could select the direction tab (e.g. go to the right). like a pit stop in F1 or gravel unbound. breaking on the course and switching bike is obviously unrealistic. having a virtual crew on the course with a spare bike is much less 🙂
As a back of the pack – caboose rider: I just go with the frame that needs upgrading.
So I rode a Trek Emonda to get climb meters.
For this week’s Tiny Races I might go with a gravel bike – Pinarello Dogma GR.
Having the wheels “sticky” – remain paired with a frame is great.
It would be nice to be able to archive unneeded frames – for last Saturday’s Tiny Race somehow I started with Zwift Steel.
Hello
Many thanks for the advice on this route. I’ve not tried switching bikes mid-race before. It worked really well for me during the race this morning (10:10am BST, 30 August).
I’m a 67.3 kg rider, B cat, Amethyst. The set ups I used for this race were:
Canyon Aeroroad 2024 (upgraded to level 5) with Princeton CarbonWorks Wake 6560 wheels.
Gravel bike: Cannondale SuperX LAB71 (upgraded to level 1) with Roval Terra Aero CLX wheels).
I watched your video before the race and i’d practiced switching bikes. I’ve got a Zwift Ride so the controllers help of course.
I went into the right hand bend with the front group of 12 riders. Once I’d switched bikes, I was off the back of them, but not by too much. I could perhaps have started pedalling slightly sooner when I switched (I was concentrating too much on leaving the garage). But it didn’t matter because I soon caught up the group and rode through them – the performance difference was marked. I then thought: I am not sure I want to be at the front with no draft with still a decent way to go, so I didn’t push as hard as I could. Anyway long story short someone ultimately shot off up the field. I finished 3rd.
For appropriate races, I’d definitely try this tactic again. But you need to have confidence, when you drop off the back of the pack that you’re going to make up the time.
Thanks
I rode the course today in recon mode . Forgot to change bikes. Have my Cannondale Gravel @ level 5. My question is this . Have a TT race , if I have a choice between Level 5 Cadex & level 5 SL9 , which should I choose ? I have swapped bikes on Road To Sky , super fast now .
In a TT race, I would swap bikes at the bottom of the full Temple KOM near the end of the route. That would give you the maximum time advantage.