Aerodynamics: Rando vs Racer

Aerodynamics: Rando vs Racer

A few weeks ago, we talked about the difference between the soon-to-be-illegal ‘Super Tuck’ and the more stable ‘Aero Tuck.’ In that discussion, we touched upon the idea that a handlebar bag can act as a fairing and make the bike more aerodynamic.

For years, I’ve noticed that I coast at the same speed as some of the other riders on the Bicycle Quarterly Team when we’re in the aero tuck – if I’m on my randonneur bikes with handlebar bags. When I’m on a BQ test bike without a bag, I usually coast quite a bit slower. We use the same Rene Herse Extralight tires on our test bikes, and bearings and other ‘marginal’ resistances don’t really play a role at these high speeds – we’re talking 65+ km/h (40+ mph) – so the difference is all in the aerodynamics. Can a ‘boxy’ handlebar bag really act as a fairing and improve the rider’s aerodynamics (in the aero tuck).

To examine this idea, I looked at my Strava speeds on a course I ride from time to time. The ‘Mountain 100k’ has two long downhills where the bike reaches ‘terminal velocity’ – that means it no longer accelerates, but rolls at a steady speed. I descend those hills in the aero tuck, coasting. For aerodynamics, high-speed testing is useful, as all other resistances (bearings, etc.) don’t matter much compared to wind resistance. I rode the course on three different bikes, with and without handlebar bags. Before we go any further, I want to emphasize that all three are wonderful bikes, and it would be hard to pick one over the other for a spirited ride in the hills.

The four runs are comparable: My riding position on these three bikes is identical. Since I’m in the aero tuck, the width of the handlebars (wider on the MIN.D.) shouldn’t affect my speed, either. There was no significant wind on any of these days, and the two hills face in different directions, so if we see similar speed differences on both, then they aren’t caused by tailwinds. Here’s the data:

BikeTire WidthWheelsHandlebar Bag?Lakemont DescentTiger Mtn. Descent
Rene Herse42 mmClassic, generator hubYes58.6 mph(44.5 mph)*
Firefly54 mmBladed SpokesYes58.6 mph47.2 mph
Firefly54 mmBladed SpokesNo53.5 mph44.1 mph
OPEN MIN.D.32 mmAeroNo53.0 mph44.1 mph
* cold pavement limited traction and required slowing for a corner in the upper part of the descent

First of all, it came as a surprise that the racing bike, with its aero-shaped frame tubes, aero wheels and narrow tires, was actually the slowest on both hills. Looking at the data, the width of the tires doesn’t seem to have much of an effect on aerodynamics. Nor does the type of wheels (classic, bladed spokes or aero rims and bladed spokes). What seems to determine the downhill speed is whether the bike has a handlebar bag or not. In fact, the same bike (Firefly) with a bag was much faster than without.

The above suggests that, in the aero tuck, the handlebar bag makes the bike significantly faster. Both on Lakemont and on Tiger Mountain, a bike seems to roll 7-10% faster with a bag. You’ll notice the qualifiers ‘suggests’ and ‘seems.’ These runs were done on three different bikes, on four different days – and there are just four runs total. And Strava uses GPS to infer speed, and that’s not always reliable. So this is the start of a research project, not the end result. The next step is to use the same bike, do multiple runs on the same day (making sure that there is no wind), with and without a bag (equalizing the weight), and measure speed directly. Once we have a good set of data, we’ll have to do a statistical analysis to see whether the differences are real or just ‘noise’ in the data… and then we can say something with more certainty.

In the meantime, here are a few observations on why the aero tuck and and randonneur bikes appears to be so effective in improving aerodynamics.

If you need any convincing that it’s more efficient to tuck rather than pedal on steep downhills, the photo above should do the trick. My camera has a self-timer function which takes 10 shots in quick succession. I rolled past the camera, and copied two images into one.

I’m relatively upright in the front photo – I’d get in the drops and bend my elbows more when pedaling downhill – but even so, it’s easy to see how much smaller the frontal area is in the aero tuck. What’s also remarkable is how much bike, bag and body become one shape. My arms almost touch the bag, and my knees almost touch my elbows. That’s much more aero than pushing the bag, arms, chest and legs through the air separately, with turbulence surrounding each element. In cycling terms, my arms are drafting off the bag, my legs are drafting off my arms, and so on.

From the front, you can see how narrow the rider’s profile gets in the aero tuck – compare that to my 42 cm-wide handlebars! This is where the randonneur bike has an added advantage over the modern bikes: The top tube is level and sits higher, so my knees can touch each other below the top tube. That means they are about 50 mm (2″) closer together than on the MIN.D., where my knees touch the top tube on either side.

50 mm may not sound like much, but it’s actually quite significant. When we tested bikes in the wind tunnel, we found that lowering the stem by just 20 mm decreases wind resistance by 5%. Shoulders are wider than legs, but bringing the knees 50 mm inward will have a significant effect. (This is something we can test by putting a 50 mm-wide piece of foam between the legs and comparing the downhill speed.)

The Q factor on my rando bike is narrower, too – by about 10 mm – which puts my feet and legs closer together. Combined, getting the top tube out of the way and putting the feet closer together reduces the frontal area of the rider quite a bit – and our wind tunnel tests have shown that reducing the frontal area is the most effective way of improving a cyclist’s aerodynamics.

Getting the legs closer together not only reduces the frontal area, it also makes the rider shaped more like a teardrop: wide in front and narrow at the back. A teardrop is the ideal aerodynamic shape. This photo also shows how the bag helps achieve that teardrop shape. Without it, you have very un-aero gaps between arms, shoulders and legs. The bag is like a small fairing to complete the teardrop and close off those gaps.

We’ve long known that the rider’s aerodynamics are far more important than what you do to the bike. That’s why aerobars are so effective – but the aero tuck is even more aerodynamic. Of course, you can’t pedal when you’re tucked, but that’s OK, because the gains from improved aerodynamics increase exponentially the faster you go. On a hilly course, it’s not just the higher speed on the downhill, but how far you can roll on the flats (or up the next hill) before you need to start pedaling again. The aero tuck is the secret for moving fast with minimal effort on a hilly course. Over a long ride like Paris-Brest-Paris, the result is quite noticeable.

How about racing? While handlebar bags would probably count as fairing and thus be illegal in racing, getting legs and feet narrower is something that works on a racing bike, too. And during that long, lonely breakaway, it might just make the difference between winning or getting caught by the pack. And once we’ve got some real testing, we’ll know how big that difference really is!

Update 3/16/21: After posting this, I realize that there is potential for a misunderstanding. The handlebar bag acts as a fairing only in the aero tuck. When you look at the second photo with the rider in a more upright position, you see that the bag is too far from the rider’s body to act as a fairing. Our wind tunnel testing showed that a bag or small fairing adds drag in that position, especially in a crosswind. For a record attempt with a support car, you wouldn’t use a handlebar bag. However, if you have to carry things, a handlebar bag is probably the best place – not just because it’s more aero when it matters the most (when you go ultra-fast on downhills). Mostly, handlebar bags make it easy to access your stuff without dismounting the bike. And since we’re going to carry a bag during long rides like Paris-Brest-Paris, we’ll optimize its aerodynamics as much as possible…

Further reading:

For a comprehensive overview about bike performance, check out our new book:

• The All-Road Bike Revolution: Make Your Bike Fast, Comfortable and Reliable

Share this post

Comments (99)

  • ToujoursOnRoule

    Hi Jan, man thanks for following up on this. Cheers, Christopher

    March 15, 2021 at 12:48 am
  • Jenny Brien

    I’d be interested to see see if any further gains can be made by:

    1 Dropping the seatpost, perhaps moving it backwards in an arc so that your shoulders come more level to the
    bars. I have an idea of a bike where most positional changes are done by changing the seat post angle from
    Rans Dynamik to Obree Mk 1

    2 When you are riding with flat pedals, moving your feet inboard to rest on the crank arms and bottom bracket.

    March 15, 2021 at 2:11 am
    • Joe

      That’s a good point, with a dropper post you’d think the aero tuck would be even faster

      March 15, 2021 at 8:57 am
      • Tim

        It’s way faster with a dropper. If I drop the post on my gravel bike when I’m at maximum speed going downhill in a tuck, I can feel it accelerate.

        March 15, 2021 at 10:24 am
    • Brad Irvine

      The UCI will surely ban the rando-tuck with hands on bag. That can’t be safe. 🙂

      March 15, 2021 at 11:22 am
      • Jan Heine

        The hands are on the bars, but fairings are already banned!

        March 15, 2021 at 12:23 pm
    • Andy Stow

      GCN just did a video where they tested a dropper.

      Spoiler: it was nearly as good as the banned supertuck of death.

      March 16, 2021 at 7:30 am
  • Julie jacobs

    Thanks for the great info. One comment is that in your aerotuck pics you’re not really holding onto the bars and nowhere near the brake levers. It seems to me that it’s important to have a good grip on the bars at all times (especially at the speeds you’re going). I can’t see myself not holding onto the bars somehow when going fast or downhill. Holding onto the bars or hoods or drops will change aerodynamics some.

    March 15, 2021 at 8:04 am
    • Jan Heine

      As you say, the speeds you achieve in the aero tuck are only for roads where you don’t need to brake. Lakemont is a new road that has four lanes and no cross streets. Tiger Mountain also is a road with no cross streets and just a few driveways that have great visibility. If there’s a risk of a car pulling in front of me, I won’t go 50+ mph!

      For the aero tuck, I place my hands next to the stem and rest my chin on my hands – or a little above if the road is rough. The bike is more stable with a light touch on the bars than if you grip them with force. (The exception are cross-winds on a high-trail bike.) Of course, when there are sharp corners, you need to more leverage on the bars than you get in the aero tuck – it’s really for straight, clear roads. And the biggest gains aren’t in the super-fast descents, but the small hills where you go 35 mph instead of 32… time and again.

      March 15, 2021 at 8:23 am
  • John C. Wilson

    Aero shapes are long and narrow. Human body is too short to be aero, all that can be done is making best from a weak starting point. Putting handlebar bag in front makes the form longer. Oscar Egg famously mounted an egg tail and it worked so well it was immediately banned. Current style ‘bikepacking’ saddlebags would do some of same. Combine the handlebar bag with the saddlebag and it starts to get long enough.

    Having another rider in slipstream always improves aero. My saddlebags are big chunky Karrimors owned for over fifty years. Too wide to be useful on descents, flatland bags. Many riding companions over the years have commented that when the saddlebags are worn it feels as if they are drafting much closer than they are.

    March 15, 2021 at 8:25 am
  • Masato Yoshihara

    Do you notice a difference when you’re riding with your CdM-edition Berthoud bag vs. the standard Berthoud bag? I remember reading your article about how eliminating the side pockets helped with aerodynamics, but I wonder how eliminating the front pocket helps with aerodynamics, if any?

    March 15, 2021 at 8:43 am
    • Jan Heine

      We haven’t tested that. We recently did test the side pockets – switching bags in mid-ride. The difference was very noticeable: The bike with the smooth-sided bag (as shown in the photo) was significantly faster.

      March 15, 2021 at 9:56 am
  • Ed

    Have you considered the weight equation? Maybe the firefly and Rene Herse bikes are heavier than the Open. Especially with the handlebar bag? I am not an expert but from experience, heavier riders always decent faster than lighter riders…

    March 15, 2021 at 8:52 am
    • Jan Heine

      That’s something we’d equalize in a ‘real’ test. But just looking at it, the weight difference amounts to just 1.5%. That can’t account for a 8-10% speed difference.

      March 15, 2021 at 9:32 am
  • Tom Anhalt

    Mass differences of the setups? Terminal velocity is dependent on that as well.

    March 15, 2021 at 8:52 am
    • Jan Heine

      Very little for the Firefly. The handlebar bag weighs as much as the seat bag I used on the other ride. My weight also doesn’t fluctuate a lot. The Open is lightest, but a 1.5% weight difference can’t account for a 8-10% speed difference.

      March 15, 2021 at 9:30 am
      • kai s

        a bigger than the smallest seat bag in itself is in my experience an effective way of slowing you down when fighting headwinds in a normal position.

        i have concluded right or wrong its important in that situation to not in any way hinder air flow between your legs.

        in that same position a tear dropped backpack really low on your back also is a significant benefit. typical commuting situation with straight bars.

        instead of a front bag in racing there sometimes used to sit two water bottles. may have been banned?

        March 15, 2021 at 10:18 pm
  • Dr J

    At this point, if you’re chasing aerodynamics, you should just install a plastic fairing with a windshield on your rando bike. I bet the results would be far better than with a rando bag.
    Of course, you would say that this is not what you like and it doesn’t match the style of your bike. Which is exactly the same answer you’re going to get if you ever suggest to a roadie enthusiast putting a front bag on his carbon road bike.

    March 15, 2021 at 9:05 am
    • Jan Heine

      We tested fairings, and unless they are very close to the rider’s body, they add drag, not reduce it. Sort of like motorcycle fairing – the racing fairings reach to within a few inches of the rider and make the motorbike faster. The touring fairing are to reduce wind buffeting, but they slow down the bike by quite a bit.

      March 15, 2021 at 9:57 am
  • Tom McGoldrick

    It would be great to see the same comparison for a flat section of the road where you are riding in your normal position. Riding at 15 MPH into a 15 MPH headwind produces 30 MPH of wind over the bike. It may just be spring in MN but this is where I really wish for an aero advantage.

    March 15, 2021 at 9:07 am
  • Eric Langley-Rokhlin

    Interested to see how the research shakes out. Even more interested to know how it feels to have your chin tucked into your fingers while descending a hill at nearly 60 MPH!

    March 15, 2021 at 9:13 am
    • Jan Heine

      It feels just fine. It’s not the most comfortable position, but you gain so much that it’s worth it.

      March 15, 2021 at 9:58 am
      • Peter

        Have you ever tried a zzipper faring , just so much easier you can still have your hands on the brakes.

        March 15, 2021 at 7:00 pm
        • Jan Heine

          We tested various fairings. They increase drag because they aren’t close enough to the rider’s body. So they’re just an additional shape to push through the air.

          March 15, 2021 at 8:28 pm
  • yipyf

    I doubt it’s a big factor here (for this data sample). But, all other things being equal, heavier is faster going downhill. Racers to start reloading their bottles at the top – 🙂

    March 15, 2021 at 9:24 am
  • marmotte27

    Hi, very interesting subject.
    I’ve been wondering about aerodynamics for a while now, and more so since last summer. I’ve got a randonneur bike and a classic Berthoud bag with side pockets. Last year I was out with a friend who is just two centimetres taller than me, has the same bike in the same size, allbeit with somewhat narrower tires (32 to my 38mm) and fenders (40 to my 50mm), and the same bag. He’s a bit heavier than me but we had Berthoud lowrider bags on the bikes and figured his must have been a bit lighter than mine (so maybe his weight was 2 or 3kilos more than mine. We were wearing similar clothing as well and both our lights were on. On a pretty long descent I went into a low position, sometimes but not always aerotuck but pretty low, whereas he stayed mostly on the hoods. Nevertheless he was always running up to me despite not really being in my slipstream. The one big difference between our bike was that I have a mudflap (90mms wide), he has none.
    We were planning to do a test of this, equalizing all the factors as far as possible, do a few rides down a slope, with and withpout mudflap for instance, and see what comes up. Well, now his bike has been destroyed in an accident so that won’t happen any longer.
    So I’m left wondering if a few mm wider tires and fenders, a mudflap and maybe 2 kilos more weight can really make such a big difference against my more aerodynamic position (I’d say 2 or 3 km/h or thereabouts)…

    March 15, 2021 at 9:32 am
  • Tom

    If you’re only (!!)) looking at aerodynamics, shouldn’t a windtunnel test (if „fundable“ ) be able to answer that question? (Despite it obviously being only one of many factors determining „terminal“ velocity.)

    March 15, 2021 at 9:41 am
    • Jan Heine

      We spent two days in the wind tunnel, but that was years ago. Yes, wind tunnel tests are very useful, but they’re also a huge effort, not just in money (wind tunnel time costs $ 500/hr) but also in preparation so you use that time well. Maybe some day we’ll return to the wind tunnel…

      March 18, 2021 at 11:11 pm
  • Evan Worthington

    What about adding a seat post bag? I’ve read that can reduce drag by reducing crossflow via a cleaner taper behind the rider. Maybe just stuff it full of bubble wrap?

    Many of my descents are dirt lately but my tuck is horrible, probably why my averages are so low right? (can’t be my climbing speed…)

    March 15, 2021 at 9:58 am
    • Ablejack BLRC

      Funny thing; Imagine I climb at 3mph and you climb at 6mph. But I’m fearlessly aero down the hills at 60mph and you are creeping downhill at 8mph. With the start and finish at the same elevation, and the same pitch up and down each hill; you will win. No matter the distance, no matter the arrangement of hills. Wouldn’t matter if I went downhill at the speed of light, I could never catch you because of that 3mph faster you go up the hills.

      March 15, 2021 at 4:32 pm
  • Jim

    I live in Lakemont and know the area well and ride many of the same hills with my lovely Rossman Rando.

    Very nice test! I’m not surprised by the results. I look forward to seeing it repeated with better controls (i.e., in series, same bike, same hill, with and without handlebar bag).

    When cycling, I avoid Lakemont blvd. like the plague, but I experiment with downhill speeds on other hills in the neighborhood and I am always surprised how fast my “Rossman Rando” with 650b 42 EL’s & Hondo fenders and Handlebar bag carries me down the hills compared to non-Rando bikes. After a year of such comparisons, I do believe the terminal velocity is higher on the Rando.

    Having said that, I do believe my Rando is slower on my standard paved “long” rides which involve a good amount of elevation gain as well. Since I do not run a power meter on the Rando, I cannot consider power differences when comparing rides and without the PM there are too many variables to have a highly informed opinion on the pace of the Rando vs. other bikes designed for tarmac.

    You are clearly a more skillful cyclist than I (and younger and more daring). No way would I go down Lakemont Blvd. at 58.6 mph, esp. with my hands on top of the stem with chin on hands (it’s a nice position that I’ve always wanted to employ, but I’m too chicken and instead I grasp the bar on either side of the stem, but otherwise do my best to emulate the position you show).

    I guess the good news is that given a speed limit of 40 mph on Lakemont Blvd., most cars won’t be coming up from behind you as you when you exceed 50 mph (haha). PS, now and then the motorcycle police are working the hill – they were last week. There are a couple of road cuts they can pull into to catch miscreant speeding cyclists and other unfortunate targets.

    BTW, unless you are on a section of Lakemont that I am not familiar with, it is a 3 lane rd., not a 4 lane rd. There is a one lane downhill, two lane uphill with a curve and bridge transition at the bottom shortly before the stoplight (I’m referring to the section between Village Park Drive and Newport Way). It is not a 4 lane road in that section. But perhaps there is another section that has been repaved that I am not familiar with.

    Second comment, since I ride some of the hills around Lakemont frequently and am quite familiar with wind patterns, I would not assume that wind direction cancels out between two separate hills. It can be quite variable in that area due to the hills.

    Keep up the great testing! I look forward to seeing the next report.

    March 15, 2021 at 10:07 am
  • Michael Wilson

    Why not test a bike with aero clip-ons? You can even add a brake lever to postpone your death if an elk or semi jumps out in front of you while descending. BTDT.

    I can say anecdotally that my 2018 Trek Domane is 3mph slower down Massachusetts hills than my BMC TMR01, the latter also having aero wheels, on a short descent speeds reaching about 40mph. Both have aero clip-ons with 3rd brake lever for the front brake, and similar riding positions. The stem on the Domane is 17 degree down to achieve that.

    My riding time is less than 4 hours, so the front bag is not something I would otherwise use.

    My Cervelo P4 with the Topeak rack trunk vertically mounted behind the seat was pretty good at descending. My intuition was that the rack trunk smoothed the rear airflow, but not so much that I actually tried that in a time trial. That was just for commuting and getting to and from the time trial.

    March 15, 2021 at 10:09 am
  • Cory Brenn

    Interesting to note, especially the comment about side pockets vs flat sides. I wonder how wind factors into this. Surely and small amount of wind and it’s direction will influence the results as well.

    March 15, 2021 at 10:13 am
    • Jan Heine

      Crosswind makes a big difference. When we tested in the wind tunnel, we looked at crosswinds, too. However, at 50+ mph, you’d need a pretty strong crosswind to change the resultant wind direction – and then you wouldn’t be in the aero tuck!

      March 15, 2021 at 12:25 pm
  • ydrive

    Ah, but it requires a good deal of flexibility to achieve an optimal position as displayed here, especially for any length of time in the saddle.

    March 15, 2021 at 10:16 am
  • Jeffrey Guild

    I believe it.
    There’s a “flow” section on the way to the City that my commuter buddies and I race down without pedaling. I’m the only one with a handlebar bag, and I routinely “win.” I’m also running rat trap pass tires on an old Bontrager drop-bar + fender conversion vs. my buddies who are all on skinny-tire road bikes and wear backpacks. Rider weights and skill-level are all comparable.
    Non-scientific, I know, but another anecdote that supports additional research for sure.

    March 15, 2021 at 10:36 am
  • Tom Anhalt

    Perhaps this is just pointing out that maybe the OPEN bike isn’t very aero?

    You know, these types of questions would be perfect to answer with coastdown testing with a Virtual Elevation (i.e. “Chung Method”) protocol. Don’t even need a power meter, since lack of pedaling means power = zero. Would have much better resolution/accuracy than your previous coast down test. Just a suggestion.

    March 15, 2021 at 10:58 am
    • Jan Heine

      The difference in the Firefly with and without the bag indicates it’s not the OPEN. In any case, the bike doesn’t really matter, if the rider position is similar.

      Yes, coasting works great for testing resistances, because gravity is a much more constant input than pedaling power. Having run tests with power meters, it’s almost impossible to keep a constant power output. Instead, we kept the speed constant – as much as possible – which gave better results, but gravity doesn’t change at all!

      March 15, 2021 at 12:32 pm

        The Chung method doesn’t require keeping power constant, in fact it’s easier to make it work well if you have variation. If you’re riding a flat course in calm conditions at constant velocity, it cannot distinguish between Crr and CdA.

        It involves using your velocity and power to infer a “virtual” elevation profile, and choosing a value for CdA that makes the virtual elevation profile make sense. “Make sense” meaning, for example, that a bunch of repeats of a loop all start and finish and the same elevation, rather than the inferred elevation showing that you’re losing a couple hundred feet of elevation on every repeat.

        March 15, 2021 at 7:50 pm
        • Jan Heine

          The Chung method is a really neat idea. Like all tests, it requires on limiting the amount of variables, but since you’re testing at different speeds, you can have more variables. The big question is how much noise there is in the data – can we resolve relatively small differences. For example, if you pedal, your position changes when your power output changes on up- and downhills… We’ll investigate this more!

          March 16, 2021 at 7:43 am
          • Tom Anhalt

            One way to evaluate the resolution is to perform something that Andy Coggan once called, “The Tom Compton” test (yes, Katie’s dad 😉 This involves doing runs with shapes of known, or easily calculable CdA. In my case, I used spheres of varying diameter on a rod. You can see the details here:

            In my case, I was able to reliably discern differences as small as .001-.002 m^2.

            March 16, 2021 at 11:10 am
          • Jan Heine

            I remember that test – it was incredible. When we went to the wind tunnel, even repeats of the same position, same rider, same bike, same day – everything the same – had variations of 3%. And this was in the best wind tunnel in the entire country… So fo you to discern changes that are several orders of magnitude smaller, on real roads, with wind and all was quite a surprise.

            Maybe I missed it, but I didn’t see any evidence that the results were repeatable. Did you run the same test multiple times and get the same results? Standard scientific practice is to do a statistical analysis to make sure the results you got are not just noise in the data. In other words: What’s the likelihood that you just got lucky?

            March 16, 2021 at 3:45 pm
          • Tom Anhalt

            …and to be fair, the Chung method doesn’t rely on _limiting_ variables, so much as it relies on measurements (such as speed, power, air density, temperature, mass). Without an on-board wind sensor, about the only variable that one would try to minimize, or eliminate, is ambient wind…oh, and braking (since that’s un-modeled). One of the things that’s cool about it is that Robert is used to dealing with “imperfect” data in his career, and so he applied those techniques to power meter data, and was quite surprised at how well it works…mostly because the data from a power meter is typically much better than what he’s used to having to deal with.

            March 16, 2021 at 11:15 am
          • Tom Anhalt

            Of the 3 of us (you, Robert, and me) I’ll defer the discussion of the details of the statistical analysis to him. In short, the analysis is a bit different than what you may be used to, and relies on evaluating the data as a series, rather than a single CdA or Crr value…and when done in laps, allows for the data to be evaluated as a LARGE number of overlapping “laps”. One way to think about it is that the consistency of the VE traces across laps is the measure of error. For example, in my “Tom Compton” tests, if unmeasured (and unmodeled) variables were changing wildly during the run, that would show up in the VE traces, and the smooth consistency shown across laps (there’s an example in that blog post) points to the precision of the result. In other words, the “evidence of repeatability” is in the runs themselves. It took me quite a bit to wrap my head around that as well.

            I don’t have the “statistical chops” to derive the error estimate myself on those plots, but I believe it was the same files from that test (or a similar one I did) that Robert analyzed to come up with his statement in his .pdf explanation that “direct assessment of the variability of VE-based estimates of CdA by “similarly” experienced users has produced CVs as low as 0.3% (i.e., sd= .0006 m^2 on an estimated CdA of around .2 m^2)” (Robert, please correct if I mis-remembered).

            Josh Poertner’s “Marginal Gains Podcast” interview with Robert is a must-listen (IMHO) to get a better feel for the methodology and logic.

            March 16, 2021 at 4:37 pm
          • Jan Heine

            I understand the idea of taking multiple data from one set to get more statistical significance. But in the very least, I want to see one totally independent repeat to show that you can recreate the experiment and get the same results. Repeatability is one of the basic standards of science. It’s not that hard – just go out again, measure again, and see how close you got. For our tire tests, we run each setup at least 3 times…

            When we tested positions, bags, etc., in the wind tunnel, we ran the same ‘reference setup’ (rider on the bike on the hoods) every day in the morning, at lunch and in the evening. That way, we could estimate how much variation we get just because the rider position changes slightly. As I mentioned, it was about 3%. The wind tunnel staff said that this was better than most riders they’d seen. Which means that you can assume that the aerodynamics of a cyclist vary by +/- 1.5% just due to involuntary changes in position, even when pedaling stationary on a trainer in the wind tunnel. On the road, I’d expect this to be greater yet.

            This means that discerning differences in wind resistance that are smaller than +/-1.5% with a rider on the bike is impossible, unless you do so hundreds or thousands of repeats. Or you test without the rider, but that’s meaningless for obvious reasons. Fortunately, for most people – even pro cyclists – +/-1.5% is good enough. After all, with the aero tuck, we’re looking at a 36% improvement in aerodynamics, and that’s before the bag comes in!

            March 16, 2021 at 5:00 pm
  • Paul

    Easy puzzle to solve.
    A bike which does not go in a straight line needs friction for steering. Speed scrubbed.
    Light wheels are super squirrelly thus they a. Need constant correction and b. actually make the trip longer by 1-2%. Next time you do a tire rollout to set your computer don’t till a straight line.
    A 500m sprint is often lost by inches. Longer chain stays ensure that the bike doesn’t lurch sideways with each powerful thrust. Heavier wheels keep the bike pointed at the finish line.
    Unfortunately, these ideas don’t sell. Which is, after all, the whole purpose of the Bicycle Industrial Complex.
    Jan, you really need to test the same bike in two runs. One with and one without 500g of lead tape under each rim strip. I think you’ll find a Baby Bear’s bed bell curve for wheels too light and too heavy.

    March 15, 2021 at 10:58 am
    • Jan Heine

      That test is easy to do – just run light or heavy tires. We did that when we tested wheel size and handling…

      March 15, 2021 at 12:33 pm
  • DoctorFairlane

    How did the tire widths compare the the width of the bikes’ down tubes? You’ve mentioned before that wide tires don’t substantially increase drag until they’re wider than the down tube. The MIN.D. with 32mm tires is quite a bit slower than the Firefly and the Herse with 54mm and 42mm wide tires. Is there a “sweet spot” ratio or range of tire width to downtube diameter that maximizes aerodynamics?

    March 15, 2021 at 11:41 am
    • Jan Heine

      It seems that tire width doesn’t make a big difference. As you mentioned, in another test, the Firefly was faster with 30 mm tires compared to the 54 mm it usually runs, but it was a tiny difference, noticeable only when coasting next to a rider who didn’t change their bike. What we’re seeing here are some pretty large differences of 8-10% in terminal velocity…

      March 15, 2021 at 12:35 pm
  • Jamie

    I have noticed my handlebar bag seems to make me faster on descents as well. My friends think I am crazy.

    March 15, 2021 at 11:49 am
  • Jim Melendy

    I have always felt that handlebar bags improved my aerodynamics. I ride recumbents, now, and I have an Easy Racers Javelin that has a cable bundle running from the top of its handlebar (about the top of my chest) out to the front of the bike, curling back under to the frame. It seems to me that it acts like a fairing, splitting the air so that it goes around me. On steep descents, I can feel the air on my face, which is above the protected area, but not on my chest and arms. As my feet leave and enter the envelope of protection, I can feel them catch the air, then lose it.

    March 15, 2021 at 12:39 pm
  • Gran

    Did the racing bike have fenders.. another variable that might have a surprising effect at these ultra-high speeds

    March 15, 2021 at 12:52 pm
    • Gran

      One more thing, I think the handlebar width may be more important than meets the eye. It’s more about the gap between the bag and the hoods than it is about the overall width – the larger that gap is, the more likely you are to create weird air eddies, and the bag width was probably constant in this test. I say this from the perspective of being a car enthusiast – the placement of small vents on the front of a vehicle can have a massive effect on the aero signature of the vehicle at high speeds (doesn’t matter much at “normal speeds”). What’s going to happen is that the air will enter that gap, narrow down, increase in speed, and then when it exits the gap and mixes with the air that went around the outside it will be forced to slow down and may create force in the opposite direction. A quick and dirty test would be to tape a dollar bill to the trailing edge of the bag and watch how violently it shakes around at moderate speeds (at higher speeds it’ll move around a lot no matter what)

      March 15, 2021 at 1:05 pm
      • Jan Heine

        Bar width was the same on both bikes with bags. But the idea of using some wool tufts to visualize airflow is good. Car makers used that in the past, driving the car at speed and taking photos. They arrived at very aero shapes that way, with drag coefficients that rival today’s cars. The problem is who’s going to take the photos. When going at high speed, I certainly can’t look at the bike – I’m looking ahead!

        March 15, 2021 at 1:19 pm
        • Tom Anhalt

          That’s what a GoPro is for! 😉

          March 15, 2021 at 1:57 pm
          • Jan Heine

            Next project: Design a GoPro mount that can film the bike from the side, about 2 m away. It gets complicated, because we’ll have to balance both the weight and drag with a mount on the other side, too. Maybe two GoPros on big outriggers?

            March 16, 2021 at 7:45 am
    • Jan Heine

      We tested fenders in the wind tunnel. The front shields the tires, but the rear catches the wind. Overall, the effects cancel mostly – there seems to be only a very slight advantage to having fenders. The Firefly doesn’t have fenders…

      March 15, 2021 at 1:17 pm
  • calcagnolibero

    More than 93 kmh max speed? Wow!!!

    March 15, 2021 at 12:57 pm
    • Jan Heine

      It’s downhill… Actually, I don’t have a computer on my bike, so I had no idea. The bikes (all three) were perfectly stable and pleasant to ride at those speeds.

      March 15, 2021 at 1:16 pm
  • urje

    Being a bit of a wimp when it comes to really fast descents, I often find myself sitting as much straight up as I can when I bomb down passes in the Alps – 40mph is plenty fast, thank you. But I do like my handlebar bag, and any reason to brag about it to other cyclists is welcome, so I’m interested in your upcoming experiment.

    I get the theory about the bag adding to the aerodynamics in the aero tuck position, but do you think it has any effect (positive or negative) while pedaling? I know, at 20mph aerodynamics matters less than at 40, but to me it actually matters more, because when I go 20 and work for it, I’d like to go faster; while if I go 40 ‘for free’, I’m OK with that.

    Perhaps when you do this experiment, what about adding a few runs in pedaling position? Just shift to some easy gear and pedal at a steady cadence without actually engaging the freewheel, so it’s still only gravity doing the actual work.

    March 15, 2021 at 2:03 pm
    • Jan Heine

      We tested bags in the wind tunnel. Those had side pockets, which appear to mess up the airflow quite a bit. Even so, drag was increased in all positions, unless the rider’s body and head were within a few centimeters of the bag. Even if you delete the side pockets, there’s too big of a gap between bag and rider – it’s unlikely to decrease drag. That’s the big problem with fairings on bicycles – if they work in one position, they will increase drag in all the others.

      However, a narrow Q factor reduces your frontal area even when pedaling, so that should be effective in all riding positions.

      March 16, 2021 at 7:43 am
  • Mark Beaver

    Not too surprised at the fairing effect of a bar bag. Do you recall the 1980’s front panniers made by Tailwinds? They claimed (truthfully) that they made a bike faster than a bike with no front panniers at all. They had a cool aero cross section too, and this was back in the days before lowriders were popularized, so the panniers helped aerodynamicize the riders legs, so went the legend. Always wanted a pair…

    March 15, 2021 at 4:42 pm
  • R. Chung

    How did you measure the terminal velocity? Are you recording speed with some sort of head unit like a Garmin or a Wahoo, and then downloading the ride data afterward? Is that how you got terminal velocity for Lakemont and Tiger Mtn? If so, then you already have collected the data you need.

    March 15, 2021 at 9:09 pm
    • Jan Heine

      The values are from Strava, but as mentioned, that’s just the start of the project. We just wanted to see whether there was something that was worth testing. Your method (assuming you’re the author and don’t just share a name) is certainly interesting.

      March 16, 2021 at 7:21 am
      • R. Chung

        Well, the slope of the road is always the same, and those tires are the same for both runs, so if the inflation pressure and temperature were close to the same, the Crr is the same for each pair of runs. Air density may have been different, but you could probably go to a weather site to get temp and baro pressure to double check. So using your already collected data to calculate Crr and CdA would be faster than doing more test runs. Of course, riding a bike is its own reward so data collection is more an opportunity than a burden, but you already have the data you need not only to calculate the drag parameters but also to assess how reliable the parameter estimates are.

        March 16, 2021 at 7:47 am
        • Jan Heine

          The plan is to use the same bike, same day, and then we get much cleaner data. If we choose the day carefully, there won’t be wind, and temperature won’t change. We’ll use your method and see how it works.

          By the way, temperature changes affect Crr to a significant degree, so that’s something to consider with your testing if you run multiple laps and the day gets warmer…

          March 16, 2021 at 8:47 am
          • Tom Anhalt

            In my own spreadsheet implementation of Robert’s VE method, I include a temperature compensation of rolling resistance, based on Crr differences measured during roller tire tests at different ambient temperatures (it turns out tire temperature usually follows ambient air temp quite closely on the road, which makes sense when you consider the convective heat transfer of all of the air flow).

            March 16, 2021 at 11:02 am
          • Jan Heine

            That matches our experience. For racing cars and motorbikes, tires warm up significantly due to the loads of accelerating, braking and cornering, but not for bicycle tires. It’s important, though, before testing, to make sure that the tires have a little time to adjust their temperature.

            March 16, 2021 at 1:25 pm
          • R. Chung

            Yes. That’s why I usually measure temperature — not just for the effect on air density, but also its effect on Crr. I used to bring a bathroom scale out with me for tests but now I just weigh everything in test condition before I leave the house. That way, I don’t have to equalize mass across test runs — if I measure the mass and use it in the calculation then I don’t have to control it. One more thing is that I soft-pedal when coasting — that way I don’t have to remember exactly where my feet and legs are, and that also improves consistency between test trials.

            Contact me if you have questions. You have my email.

            March 16, 2021 at 11:30 am
          • Jan Heine

            We’ll be in touch!

            March 16, 2021 at 1:27 pm
  • Mark

    But what about the effect when not riding/coasting down a hill in the aero truck? Surely when sitting up the space behind the bag creates eddies &c? And the square shape can’t be too efficient (or could be better)—? Isn’t the most time on a ride spent on the flat or uphill? Gains downhill would be more than offset by losses uphill or on the flat. Similarly, can the decrease in leg/pedal width make much difference—esp. on road (rather than track) where there’ll be many factors at play—since, as you say, it’s the overall width/cross section that’s the biggest factor? I write this as someone whose overall average speed on a rando bike with front bag & ‘guards is around 4% higher than my overall ave. speed on a modern carbon bike w/out front bags (but, due to strong Covid restrictions where I live I’ve only ridden about a quarter of the distance on the rando bike compared to the other one, and the longest rides have been only c.120 km vs 650 km). Certainly, riding into a head wind (or up a hill) on my 38 mm tyred rando bike with bag feels much harder than on my carbon bike with 32 mm tyres (all RH tyres, of course).

    March 15, 2021 at 9:16 pm
    • Jan Heine

      We tested bags in the wind tunnel. Those had side pockets, which appear to mess up the airflow quite a bit. Even so, drag was increased in all positions, unless the rider’s body and head were within a few centimeters of the bag. Even if you delete the side pockets, there’s too big of a gap between bag and rider – it’s unlikely to decrease drag. That’s the big problem with fairings on bicycles – if they work in one position, they will increase drag in all the others.

      You can read more about that in our book ‘The All-Road Bike Revolution.’

      March 16, 2021 at 7:19 am
  • Dan

    When I was racing, I tried every aerodynamic trick, because of my tall height and lack of training. One, of course, was the aero tuck. Just be careful not to bang your chin on the stem when hitting bumps!

    March 15, 2021 at 10:36 pm
  • Adam

    Jan, if you crash, for whatever reason, with that speed-you will be scraped from bitumen over long distance.
    Is it worth the risk ? Use some extreme, full body protection 🙁

    March 15, 2021 at 11:45 pm
  • Phlpp

    Ha, another common-sense-bike-myth to be dismantled, love it!
    Will you also be testing how handlebar bags affect aerodynamics when you are not in the aero-tuck but riding in a normal more upright position? Which is propably the more common position in a long ride.

    March 16, 2021 at 1:14 am
  • C.C

    I rekon more improvement can be made with a custom cafe racer front fairing shaped bag. Smooth surface on the bag vs coarse textile will probably help as well. As per studied aerodynamics on TT riders, frontal area / front turbulence is the main drag.
    Certainly interesting to think about it.

    March 16, 2021 at 1:29 am
    • Jan Heine

      A good fairing might help a bit, but key is to make the fairing so it doesn’t increase frontal area… All fairings I’ve seen for bicycles increase drag, because they don’t come close enough to the rider’s body.

      March 16, 2021 at 7:14 am
  • Jan

    Honestly I didn’t expect that. I remember the spike effect in aerodynamics which obviously explain the result. Curious about whether the difference would be the same in a normal riding position on the hoods.

    March 16, 2021 at 1:48 am
    • Jan Heine

      We tested bags in the wind tunnel. Those had side pockets, which appear to mess up the airflow quite a bit. Even so, drag was increased in all positions, unless the rider’s body and head were within a few centimeters of the bag. You can read more about that in our book ‘The All-Road Bike Revolution.’

      March 16, 2021 at 7:13 am
  • Ed Garnett

    There’s a glaring omission Jan in your Areo-Tuck position, and that’s the position of your head/helmet, in terms of minimising drag as I understand it. If you can lower the head/helmet to be in-line or below your shoulders and back then you will save 10’s of watts I believe. Otherwise a very nice position you’re in. I appreciate the head below your shoulders/back line would make it almost impossible for you to see where you’re going with your rando/bag combo… but it’s a major consideration in being aero, heads and helmets cause considerable interference.

    March 16, 2021 at 3:33 am
    • Jan Heine

      My chin rests on my hands, which rest on the handlebars. So unless I lower the bars, my head won’t get any lower. In any case, I’d have a hard time seeing the road.

      March 16, 2021 at 7:12 am
  • bob

    More “science” from rene herse… Yes sure rando bikes with square, wind-blocking handlebar bags are faster than road racing bikes. I’m sure protour teams will be running steel touring frames with dyanmo hubs in the tour any day now…

    March 16, 2021 at 8:33 am
    • Jan Heine

      This idea is just as crazy as the idea that 25 mm tires are faster than 20s and 23s. When we first figured this out, people laughed. And then the pro teams looked at our testing, confirmed it with their own tests, and… you won’t see a pro racers on anything narrower than 25s now.

      I think what pro racers use is a very complex issue. If you read our book ‘The All-Road Bike Revolution,’ it clearly explains why 25 mm tires make sense for pro races (lightweight riders, smooth roads, sprints are often won by half a wheel or less), but why the rest of us may be better on wider tires.

      As to generator hubs, not sure why you think pro racers would use them, since they ride only in daytime. Endurance racers like Lael Wilcox and Ted King prefer generator hubs over battery lights because they free them up from having to worry about charging lights.

      March 16, 2021 at 9:38 am
    • Owen

      Bob, this seems a little snarky. Contrary to most cycling equipment tests, BQ does an excellent job of designing their experiments and analyzing the data in a thorough and transparent manner, with both results and methodology presented for peer review. This is how real scientific research works and I’d challenge anyone to come up with real data from real roads showing that wider tires are slower. Indeed most magazine “tests” fall apart when you analyze the methodology. The cycling industry is bound both by consumer demand and tradition, two things which do not facilitate objective scientific research. In other words, pro racers will ride whatever they’re paid to, and most consumers will follow suit.

      I had my own doubts about wider tires being faster on the road, then I bought a pair of RH Babyshoe Pass Extralights and haven’t looked back. For the record I’m a BQ subscriber but have no financial interest in either BQ or RH; I’m just an average cyclist who appreciates thorough and objective research–especially when it leads to more speed and comfort on my own bike.

      March 16, 2021 at 3:43 pm
      • Michael Banks

        Okay, Bob was a bit “snarky”. The comparison is interesting and Jan has stated that it is preliminary. But Owen, lets be clear, it is not peer reviewed, nor does it rise to the level of a rigorous peer review process. BQ is a gem, and the results should be taken in the context of BQ.

        March 17, 2021 at 9:39 pm
        • Jan Heine

          Absolute – these aren’t final results. We reported them because there’s a lot of evidence pointing in this direction – not just the Strava numbers.

          However, I think it’s important to keep in mind that there isn’t much cycling-related research that meets a ‘rigorous scientific peer review process.’ But even in science, what matters in the end is the track record of the researchers. It’s hard to evaluate one study, but once you see that people’s research stands the test of time, you can assume that they’re using good practices and careful data collection. I hope our research about tires has stood the test of time by now…

          March 18, 2021 at 3:18 pm
  • Derek

    Does shoulder position make a difference? Without moving your head and hands, you can either drop your shoulders, or pull them in toward your ears. I think it could be significant, but there is no such thing as a long, fast downhill anywhere near where I live.

    March 16, 2021 at 4:43 pm
    • Jan Heine

      I don’t know. Just looking at it, it probably won’t make too much of a difference. The aero tuck reduces the frontal area of the cyclist by 33% – we measured that in the wind tunnel. There was also a small reduction in turbulence, and that seems to be where the bag helps – not because it’s super-streamlined, but because it closes the shape. Funnily, we tried a streamlined foam cone, and it was worse than the bag, since it seems to funnel the air into places where it creates more turbulence than a boxy bag. Aerodynamics are complex…

      March 16, 2021 at 4:49 pm
      • Derek

        When you’re already in the aero tuck, does pulling your shoulders in make a small improvement?

        March 16, 2021 at 5:13 pm
        • Jan Heine

          Good question – we don’t know. We could test this…

          March 16, 2021 at 6:16 pm
  • Pete Chesworth

    Holding the tuck position at speed on a low trail bike is not easy.

    March 16, 2021 at 6:58 pm
    • Jan Heine

      At speeds that high, all bikes are stable, because the gyroscopic forces of the wheels are huge. Trail doesn’t make a big difference, and on a calm day, you’d be hard-pressed to notice any difference in how high- and low-trail bikes behave at 40+ mph.

      The only difference is how bikes react to crosswinds. Trail provides a lever for the sidewind force to turn the handlebars, so a high-trail bike will not just move sideways, but its handlebar also turn. It veers off course quite suddenly. A low-trail bike is much less affected by crosswinds, so I find it much easier to ride in the aero tuck.

      March 16, 2021 at 8:47 pm
  • marmotte27

    Wonder what happened to my comment, a long story about how a friend and I have the same size, weight, bags and bike – well almost, I’ve got 6mm more tire width, 10mm more fender width and a largish mudflap and he doesn’t – and yet he was faster by 2-3 km/h on a descent, and the question if this can all be due to those differences in tires and fenders? (We were going to test this in a controlled experiment, but now his bike has been destroyed…)

    March 17, 2021 at 9:37 am
  • yipyf


    More fun with bags. How about some low riders too – front, back, front+back? You may not want to aero tuck for these. There’s some good turbulence coming off the legs. I can get an idea of how much by undoing the zips on my tights.

    March 17, 2021 at 9:50 am
    • Jan Heine

      We tested panniers in the wind tunnel. Surprisingly, front lower-rider bags were much more aero than rear panniers, especially in crosswinds. In fact, front low-riders in crosswinds shield the wheel, which makes up for most of the aero losses due to the larger shape. Rear bags were pretty bad in all conditions. The wind tunnel people pointed out that you want a teardrop shape to be narrow at the back, to to them the results made sense…

      March 18, 2021 at 3:14 pm
  • JG White

    Very interesting/fun post, thanks for sharing. Look forward to any further research/testing.

    So the fastest speed that I’ve recorded in recent times on a bike was around 53 mph. That was on my now retired late 1980s Ritchey Outback (expedition style frame w/ drop bars) on a wide open descent near Death Valley. Bike was unloaded but equipped w/ fenders, front and rear racks and, yes, a handlebar bag. 26 inch wheels running the first iteration of the then Compass 26 inch tires. I don’t recall doing anything special like a tuck. I was somewhat faster than my several companions who were riding mostly regular road bikes. Always attributed the speed on that descent to the Ritchey being on the heavier side, and the relatively ideal road conditions.

    Two days prior, rode the bike in the same configuration for what should have been a speedy, gradual downhill 50 mile day. But we were greeted by just brutal headwinds for the entire day. I fared somewhat worse than my companions on the road bikes. I’ve always felt that the combo of front rack, fenders and handlebar bag worked against me on that day. Perhaps I was wrong! I do know that it was a mentally tough day simply because I was expecting smooth/fast sailing due to the long, gradual downhill route.

    March 17, 2021 at 10:07 am
    • Jan Heine

      The bag helps only when you’re tucking and your arms and head are almost touching the bag. Otherwise, the bag will slow you down a bit.

      March 18, 2021 at 3:15 pm
  • Mark

    Re your addendum about bags on long brevets (mostly on paved roads, eg, PBP, LEL). There may be an upper limit to the size (& loaded weight) of a bag that determines its usefulness & the gains of convenience over aero. I’ve used bar, frame, and saddle bags of all sorts & combinations. I find the extra resistance of large bags (say 8-10+ litres) outweighs convenience; with 2-3 L bags you may as well put stuff in a frame bag, which generally holds more and, for me, is just as convenient. So, perhaps something around 5-7 L? Front bags might be better, too, if longer than wide—I’m making one to find out. I’m undecided, too, if the extra 100-200 gm of a rack is worth it (I use a RH one). Respectfully, I disagree being more aero when descending is when it matters most: far more riding time is spent on the flat, in head or side winds, or climbing. Finishing usually mid- to front-of-pack whatever the distance, it makes little difference to me where I store things (apart from snacks), but FOR ME frame bags work best & my fastest brevets have been with frame & top tube bags (and, sometimes, a small front bag, plus, where allowed, TT bars—of course, many other factors obtain). Bigger front bags work well for me at the slower pace of cyclo-touring or exploring—not riding ‘spiritedly’. I must admit I find claims that bikes with bar bags are over all more aero than those without one wishful thinking at best—as you say, going for a record you wouldn’t use one. But, though it doesn’t square with my experience, I may be wrong.

    March 17, 2021 at 8:30 pm
    • Jan Heine

      For long rides, most of the time lost is due to stops. So anything that means you can ride rather than stop is going to do more than small gains in aerodynamics. An experienced rider can access their food, clothes, etc. in a handlebar bag while riding, rather than needing to stop. That’s the biggest gain from a handlebar bag. That it’s more aero during the downhills is a (small) plus.

      March 18, 2021 at 3:16 pm

Comments are closed.

Are you on our list?

Every week, we bring you stories of great rides, new products, and fascinating tech. Sign up and enjoy the ride!

* indicates required