Myth 4: Stiffer Frames Are Faster

To celebrate 15 years of Bicycle Quarterly, we are examining 12 myths in cycling – things we (and most others) used to believe, but which we have found to be not true. Today, we’ll look at frame stiffness.

When we started Bicycle Quarterly, the thinking about frame stiffness fell into two camps. The majority of cyclists subscribed to the notion that frame flex wastes energy and that stiffer frames are faster. A few scientific types believed that the energy lost to frame flex was small, and thus frame stiffness probably does not matter. There were a few builders, like Bill Davidson, who extolled the ‘lively ride’ of lightweight tubes, but they were mostly ignored.

At Bicycle Quarterly, we mostly subscribed to the notion that it didn’t matter. And so we were happy riding relatively flexible frames… Sure, stiffer frames might offer marginally better performance, but seeing pros win on Vitus and Alan frames that had a reputation for being ‘noodles,’ we figured that if a frame was stiff enough for Tony Rominger and Sean Kelly, it would be stiff enough for us.

Then we tested a bike that didn’t perform well for us. It seemed to bog down on the climbs. It was harder to maintain a high cadence. It wasn’t as much fun to ride. I described this to the framebuilder Peter Weigle, adding: “The frame is made from heavy-wall, oversized tubing, so it must be plenty stiff. I can’t figure out why it doesn’t perform.” Peter paused for a while, then he said: “What if the frame is too stiff for you?”

That was something I’d never considered! It was like saying that my bike was too light, or that I had too much power. But it got me thinking.

Along came another Bicycle Quarterly test bike (above). This one performed better than expected. It wasn’t particularly lightweight, and our initial expectations weren’t all that high. And yet, whether it was me or Mark (our second tester) riding it, this bike climbed faster than our other bikes. It turned out that it was made from very thinwall, and thus flexible, tubing.

So we had tested one bike that was stiffer than our own, and it didn’t perform as well. A second one was more flexible, yet it performed better. Even more startling was the difference in feel. On the flexible bike, pedaling faster didn’t seem as hard. We were out of breath, but our legs didn’t hurt. Once we got in sync with the frame, its response to our pedal strokes felt like a boat rising out of the water, going faster with only a little extra energy input. “You mean, it ‘planes’,” said Matthew Grimm of Kogswell, when I described the phenomenon to him. Deciding that the phenomenon needed a name, we used the term ‘planing’ to describe it.

We could only guess at the physical explanations for what we observed, so a term that was purely descriptive of our observations seemed best and most honest. Sort of like saying that a bike ‘flies’ up a hill, when in reality, its tires don’t leave the ground…

How to test whether our experience was real, and not just our perception? (Perhaps Mark and I just liked red bikes?) All the other magazines were still talking about ‘laterally stiff and vertically compliant frames’ as the ultimate goal… We decided to do a double-blind test with four identical frames, made from three different tubesets. (The duplicate frame served as a control.) Apart from the top and down tubes, the bikes were identical down to the last component. Their weights were equalized to make them truly the same – except that their flex characteristics were different.

The only way to identify them was by their stem cap, and that was switched by the test adminstrator between test runs. And of course, the testers weren’t allowed to talk to each other until the experiment was unblinded at the very end. This test was a huge (and expensive) undertaking for a small magazine, but we felt it was important to do this right.

The results confirmed our previous impressions: Two of our three testers could identify which frame they were riding with 100% accuracy, just based on how the frames performed under hard pedaling. Not only that, these riders were consistently faster on the more flexible frames. Power meters showed that they put out up to 12% more power on the frames that ‘planed’ best for them, yet they felt easier to ride hard. (Our third tester couldn’t tell the – very small – differences. All the bikes in this test were relatively flexible by today’s standards.)

Whether it’s 12% more (for Mark) or a bit less (for me), the difference in power output was very significant! What is happening when a frame planes? A frame that is too stiff apparently ‘pushes back’ against the rider’s pedaling. The rider cannot apply maximum power during the downstroke before their legs start hurting. Imagine pushing against a brick wall – the wall doesn’t move, so no work is done, yet your legs fatigue quickly.

If the bike ‘planes’ in sync with your pedal strokes, then your legs no longer are the limiting factor. Now your cardiovascular system determines how fast you can go: Your maximum heart rate is the limit.

On the stiffer bikes, our legs hurt, but we never reached our maximum heart rate. On the more flexible bikes, our legs didn’t hurt, but we were completely out of breath when we reached the top after putting out significantly more power on the climb.

In the decade since we published our double-blind tests, the belief that stiffer frames are better has lost a lot of traction. Experts finally have tried to measure the energy lost to frame flex, and they came up empty-handed. When Damon Rinard, Road Engineering Manager at Cannondale, proclaims, “I no longer believe that the ultimate rigidity defines the ultimate bike!” you know that the world is changing.

The challenge for the future is to fine-tune frame stiffness to the rider. It’s not simply that ‘more flex is better.’ Our tests indicate that more powerful riders may benefit from (slightly) stiffer frames. It all depends on your pedal stroke and power output.

Our subsequent research shows that flex needs to be in the right places for the frame to get in sync with the pedal strokes, so that the rider can reach their maximum power output. We now realize that the frame doesn’t just serve to connect the parts, but that it is literally the heart of the bike. Like the right amount of flex in a gym floor allows you to jump higher, or the right amount of flex allows pole vaulters to reach incredible heights, the right amount of frame flex allows cyclists to reach their maximum potential.

Further reading:

Acknowledgments: We thank framebuilder Jeff Lyon who made the frames for the double-blind test, Hank Folsom of Henry James who donated the True Temper tubing, and Hahn Rossman who administered the experiment.

About Jan Heine, Editor, Bicycle Quarterly

Spirited rides that zig-zag across mountain ranges. Bicycle Quarterly magazine and its sister company, Compass Cycles, that turns our research into high-performance components for real-world riders.
This entry was posted in Testing and Tech. Bookmark the permalink.

40 Responses to Myth 4: Stiffer Frames Are Faster

  1. Petar says:

    Excellent idea and, of course, the execution on the intriguing topic – The Myths. As an purely amateur rider, can not say much about the matter, but I agree with all the statements. I ride 62 cm Pinarello Montello and in spite of lightweight SLX tubing, and no chainstay bridge, it is flexible on the most beautiful way. Riding is pure pleasure – as they say: riding on the cloud! Longing for the next article

  2. Steve Skwarlo says:

    I’m enjoying “Myths” very much.  You explain what I know from riding the bike.  For quite awhile my tires of choice, a good place to start, have been Stampede Pass 700 X 32 or Barlow Pass 700 X 38 nothing less.  My Ti road bike with Chinook Pass 700 X 28 tires tends to languish.  It’s an excellent choice of the day twenty years ago.  Considerable frame modification was required to allow the minimum tolerable tire size.  Long gone are 700 X 25/23 tires so fashionable then.  My enjoyment of cycling has improved thanks to Bicycle Quarterly and the radical rediscoveries well described.  The wonderful adventure stories are a bonus!

    I’m eager to read “Myths” 5 thru 12.  I can’t imagine the list of commonly held beliefs about to be deflated!  Excellent!!

    Thanks, Steve Skwarlo

    • marmotte27 says:

      “Modern bikes are faster” would be the all-encompassing myth. The busting of it in BQ, in what was it 2009 (?), is still my favourite article.

      • Ah, yes, that was in Bicycle Quarterly 32: We compared the increase in speed of bike races with medium-distance running, and found that runners had a bigger increase in speed than bike racers, indicating that training and other human factors were responsible for the faster bike races today… That was a fun study!

        That said, many modern bikes are truly excellent. There are a few reviews coming up in BQ where I was once again tempted to buy the test bikes…

  3. Ray Varella says:

    Even though the third tester could not tell the difference between the frames, did the power meter show any difference in his performance?

    • We only had two power meters, so we only used them only with the two testers who could tell the difference between the frames. It would be interesting to repeat the tests for the third tester with frames that had larger differences in stiffness. All the frames in this test were flexible compared to most ‘modern’ frames: The ‘standard’ one used the similar tubing as Columbus SL (standard-diameter 9-6-9 tubing), the ‘superlight’ ones used the most flexible tubing ever made (standard-diameter 7-4-7). If we had included a ‘modern’ steel production frame with oversized 9-6-9 tubing, I would not be surprised if the third tester also could have told them apart.

  4. Paul says:

    Pantani could pedal his steel bike in the big ring up the Tourmalet and today’s riders need a 34 ring to do it. That says it all.

    • I think Pantani raced an aluminum Bianchi. However, you are right about the difference in style. Modern pro-level bikes seem to reward a high-cadence spin, whereas classic bikes often climbed best when the rider used their entire body (and a bigger gear). Just watch a video of Eddy Merckx…

    • Samuel Atkinson says:

      Although that impacts the subject matter, it’s a mostly separate matter of effort management. Today’s professional racers don’t strictly need a 34T ring to climb the Tourmalet, they use a 34T ring to climb the Tourmalet so that they don’t need to force their quads into a zone that’ll cause rapid muscular fatigue.

      In the recent BQ article about Jack Pass, I was happy to see the steep 100-foot climb early along Ben Howard road get mentioned. It’s a very short climb, but it has a striking effect when riding that road, in either direction. The fastest I’ve ever ascended that thing was on the 41″ granny gear of my 1983 Miyata 710, when I was relatively fresh and wanted to avoid bottoming out the gearing. The slowest I’ve ever ascended it was also on the 41″ granny gear of my 1983 Miyata 710, when I was less fresh and incapable of keeping it turning powerfully. The first situation looked “impressive”, but neither was anywhere near optimal for performing well in the respective group ride. A lower gear would have let me avoid both problems.

  5. Max Sievers says:

    Thanks for the pioneering work in the fields of researching and education of cycling physics. It’s a shame that almost no producer or constructor knows what he is doing. The marketing uses the wrong and misses the right terms like. They design tandems with the same angles to get the same handling like the single bike.

    • It’s quite complex, and for the big makers, it’s hard to figure out what is right and what isn’t. Not everybody has the luxury of testing identical bikes with just two different tubes…

      • Rick Thompson says:

        How can the big bike manufacturers not be capable of this kind of testing? Trek is privately held so profits are not known, but revenue is around $1 billion according to Forbes.

      • In Trek’s defense, I read somewhere that when they made a stiffer frame for Lance Armstrong, he actually preferred the previous model that was a little less stiff. That is why it’s so important to work with people who compete and take bikes to the limit – to fine-tune the design and optimize its performance. The Trek Madone we tested climbed really well, and if I had the power of Lance Armstrong, it probably would have been even better. (As it was, it worked well for me for the first 50 miles of an ultra-fast ride.)

  6. Mitch Harris says:

    Great review, thanks. Can you say something about differences among the three riders? If one of the three did not feel the difference the other two did, was it perhaps because he was had a different body weight or power level? For example if he was significantly lighter or less strong, he may not have initiated planing in the lighter gauge frames.

    • The two riders who could tell the difference (Mark and I) were closely matched, strong riders. When riding next to each other on the same frames, it was a close call on the sprints up the hill. When one was on the ‘less-optimal’ frame, they had a hard time. I once actually beat Mark up the hill on the ‘less-optimal’ frame, but from the effort it took, it was clear that I wasn’t on the ‘superlight’ one. And even when riding alone, we could tell how the frames responded.

      The third tester had a somewhat lower output. I suspect, as you say, that the bikes felt similar to him because he didn’t flex them quite as much. Also, he rode alone, so he didn’t have a direct comparison.

  7. Timothy Nielsen says:

    Great stuff, thanks. I have slowly accumulated many old steel bikes, and ride them all occasionally. It is amazing how much I can get some bikes to flex now. These are bikes I’ve riden for decades. But the difference now is my weight, it has increased by 60 lbs over my racing weight. I’ve actually really enjoyed the difference, it must be this “planing” you speak of. Wet Noodle Bikes should be a brand name by now, I love em.

  8. RickH says:

    This is doing my head in when looking for a responsive bike frame. How would one go about getting such a frame when their weight and power output is different to the manufacturers or tester.
    I read a tire resistance test recently where the performance of the tires changed between rough and smooth surface conditions. Would these “planing” frames be the same. How could the average rider “feel” if it is a good frame? And do you have to pedal in a specific way to get the frame to react to your energy input.
    Too many questions in my head. Need paracetamol.

    • Do you have to pedal in a specific way to get the frame to react to your energy input.

      It’s like dribbling a basketball – you need to work with the ball to get in sync. The same applies to the frame. When I first rode oversize frames, it sometimes took a hundred miles until my pedal stroke adjusted to the different flex characteristic, and then the performance suddenly became much better. It was almost like an afterburner switched on… But then, those frames weren’t optimally suited to my pedal strokes. When you ride a frame that really works for you, I think you’ll feel it. In the past, some called it ‘lively,’ and the French used the term ‘nervous’ (as with racehorses).

  9. Stuart Fogg says:

    Have you tried building (tangential) compliance into a crankset, something like the springs in an automobile clutch disc? That might be easier to adjust than frame stiffness and should have no effect on handling.

    • We did a similar experiment with a Trek 2100C that had an elastomer in the rear triangle. It was intended to absorb shocks, but the suspension travel was only a few millimeters. The bike came with three elastomers: hard, medium and soft. We made a hard plastic block that replaced the elastomers to lock out the suspension entirely. We rode the bike with all four inserts in the rear triangle. What we found was that without suspension and with the hard elastomer, the bike felt like a stiff aluminum frame (which it was).

      With the medium elastomer, the performance was transformed, and the bike ‘planed’ great for us. It climbed much faster. The transformation was quite a surprise. When we used the soft elastomer, the bike felt out of phase with our pedals strokes and started to ‘pogo’ a bit like a bad full-suspension bike. Clearly, the amount of ‘give’ has to be tuned to the rider’s power output and pedal stroke. The full report is in Bicycle Quarterly 16.

  10. Andy Stow says:

    I’m guessing that the rear triangle is relatively stiff, laterally, compared to the front triangle? Is the front triangle stiffness what primarily needs to be tuned? I think there is at least some rear triangle flex as I’ve had one bike which would rub the rim on the rear brake only when I pedaled hard. Or was that more likely wheel flex? Anyhow, between needing to keep the brakes centered and the gears aligned, I’d think that rear triangle flex would be undesirable.

    It would be interesting to design a frame with variable stiffness, perhaps with adjustable tension wires parallel to the tubes, and see if a relationship could be established between frame stiffness and rider power and/or weight and/or cadence.

    • It would be interesting to design a frame with variable stiffness

      The Trek 2100C (see comment above) offered adjustable suspension that changed how the frame reacted to pedaling inputs. It was an interesting experiment. The alternative probably would be to make many frames with just slightly different tubing, so you can adjust the stiffness in small steps by moving from one frame to the next. We did that in our initial double-blind test.

  11. I read a 1974 test of a Gitane Get-away traveller’s bike, and it seems frame stiffness was something NOT DESIRABLE back then!

    Here’s the quote: “The salesman called after me that I would probably find the bicycle stiffer than my import. But the test ride revealed no instability, very little stiffness, and no unusual handling characteristics” (BUSICK, DON M.:”The Getaway”, Bicycling! February 1974, p.47)

    After reading that, I wonder if the search for rigidity was a trend born out of the carbon fiber craze.

    By the way, the Gitane Get-away approach for travelling was very close to a modern “Rinko” bike. Check it out:
    http://www.velo-pages.com/main.php?g2_itemId=1625

    • Very interesting! However, the belief that frame flex wasted energy is very old. When Nicola Barra introduced his aluminum bikes in the 1930s, he tried to dispel the idea that they were flexible by publishing test results of frame stiffness. The illustration of the frame stiffness test is taken from his article.

    • George VZ says:

      Wow , the Get-Away !

    • Russ Paprocki says:

      Click “Next” 7 times on your link and check out the “Gran Tourisme” model. Missed one of these in my size on E-bay a few years ago, still kicking myself.

  12. mikebike says:

    Jan,
    One thing you mentioned in a previous post is that titanium could be beneficial to small (less heavy) riders because it has the potential to flex more than steel – since steel tubing less than 7/4/7 is not practical. Could you elaborate more on this and how would one go about finding a bike with this characteristic. Have you spoken with “small/light” folks who have found a particular bike that works well for them?
    Always enjoy geeking out on your articles, they are much appreciated!

  13. larryatcycleitalia says:

    Thanks for busting another myth. If I had a dime for every time I tried to explain “the frame doesn’t just serve to connect the parts, but that it is literally the heart of the bike.” in the bike shop I’d have a big bag of dimes!
    Even today how often do we see “Ultegra bikes or $XX!” as if the components were all that mattered? And of course with the big brands all having their stuff made in just a few Asian factories,
    all of ’em in t-shirt (too big, too small and close enough) sizes, what else are they going to point out when it’s time to tout “value.”?
    Keep up the great work, your message is (slowly) getting out there.

  14. Baldo says:

    Great post and good job.
    But It makes me wonder if you measured the stiffness only with the sensation given by the riders or also with the sophisticated method described in the drawing. Besides that, maybe the behaviour of the frame is given not only by the stiffness but also by its frequency response or natural oscillation frequencies (??). Maybe someone could be interested in the very simple method for measuring the stiffness of a frame that I propose in this post: http://www.enbicipormadrid.es/2018/01/como-medir-la-rigidez-del-cuadro-de-una.html
    Sorry it’s in Spanish, but Google may help in that

    • We also measured the stiffness of the test bikes and quite a few others using the method in the drawing. There were no surprises – the thinner-wall tubing made the frame significantly less stiff. It was on par with the most flexible frame we could find – an old Barra aluminum frame. Interestingly, my Alan cyclocross frame was far from a ‘noodle,’ testing as stiff as our ‘Standard’ frame made from 9-6-9 tubing and the ‘Oversize’ frame made from thinwall 7-4-7 OS tubing.

      • Baldo says:

        I have a doubt regarding the rotational stiffness measuring with the method in the drawing, In my opinion, the rear wheel doesn’t collaborate at all when submitting the bike to a side torque while riding. You grip the handlebar and press (and pull) the pedals, but the rear wheel is free to move sideways. So, for measuring stiffness and to make it comparable to the real riding forces applied, ¿shoudnt you leave point B free?. That’s what I did in the method I used in the aforementioned link, holding only fork and handlebar. Thank you.

  15. Thomas Gandesbery says:

    Jan:
    What’s next? “Myth: lighter bikes are faster on climbs” Seriously though, thank you for these “myth buster” articles- very informative. As far as your side-by-side frame testing, you know the Big Three (Trek, Specialized, Cannondale) have done this sort of thing over and over in-house and have kept the results secret. The stiff is better message just sounds right and is easier to sell to the public. And so it seems the theme is that conventional bicycle R and D considers the machine as tested on a bench (quantified), more than what is happening to the rider (subjective).

    • There is no doubt that lighter bikes are faster on climbs, all other things being equal. Just how much faster… It’s less than most people think.

      The hardest decision with any R&D program is what you test. For many years, the assumption was that stiffer was better, so companies tested how to make their frames stiffer without increasing the weight…

  16. Raymond says:

    I’m lucky enough to have had four frames built by one builder. The mountain bike is not relevant here. Of the other three, the “gravel” bike was the first (and the first of the four) though there was no such term for it in the 90s when it was built. It is made out of Tange prestige and a few aircraft spec chromo tubes – down tube and seat tube I believe. The second one was a traditional road bike with super thin French tubing (Excel(?), I think). The third is also a traditional road bike that was built with “modern” OS tubing, bc I wanted a “stiffer” bike. Both road bikes had the same geometry. Guess which bike I still ride (and would need therapy if I ever ruined it) and which of the other two would be my second choice . . . I am 6’3” and weigh btw 200&205, depending.

  17. Who did the buildup on the Terraferma? If the stem is not inserted past the top headset race, it presents a very dangerous situation where forces on the handlebar can break the fork steerer clean off at its weakest point — on the threaded portion below the bottom of the stem.

    • You are right, the stem should be inserted beyond the threads of the steerer tube, especially when the bike is delivered to a customer who may tighten the stem bolt with force. The Terraferma was sent to us as a bare frame, with the fork uncut. Since it wasn’t our bike, we didn’t want to cut the fork. It was ridden like this for only 200 miles, with the stem only tightened enough to hold. That posed no problem.

  18. scott says:

    As a tall 230lb rider with a 370w ftp i enjoy your blog because it’s a great counter point to my own experience. For me, the frame tubes you love feel like flexy garbage. Stiff frames feel great, but some of them chatter through corners on descents, so how the stiffness is applied is important. I don’t think stiffer/softer frames are better but that it’s all about matching the rider to the bike. Hard to write articles with that much nuance, though.

  19. antimony says:

    So how do you avoid ghost-shifting with super-flexy bikes?

    • We’ve never experienced ghost shifts on any of theses bikes, nor our own, despite sprinting up hills with power outputs of 800 Watt and more. For rear derailleurs, it’s only an issue if your shift levers are badly worn.

      On the front, derailleur rub actually occurs more often with stiffer tubes, because of how tubes are butted. A flexible 7-4-7 tube is 70% as thick at the ends as a stiff 10-7-10 tube, but only 57% as stiff in the center. This means the flexible tube will flex mostly in its center, whereas the stiff tube will flex all over, with more flex between the front derailleur and the bottom bracket. That matches my experience on the road – the only bike I ever rode that had bad front derailleur rub (with different components) was made from stiff ‘Super Tourist’ tubing.

Comments are closed.