Myth 15: Marginal Gains

‘Marginal gains’ are the latest buzzword in cycling. The idea is that many tiny improvements can add up to make a meaningful difference. Make 10 changes that each save 3 Watts, and you’ll have gained 30 Watts…

Think of Greg LeMond winning the 1989 Tour de France by eight seconds… If the second-placed rider, Laurent Fignon, had used ceramic bearings, he might have won that year.

Chasing these marginal gains, cyclists put bigger pulleys into their derailleurs to reduce the bending of the chain and make other tiny improvements.

Marginal gains may be appealing when you feel that they are all that is left, after you’ve optimized everything else on the bike. And yet most riders still can make major improvements that will outweigh the sum of all the marginal gains. Here are two examples:

  • Switching to truly fast tires gives you the biggest edge. We are talking 5% in speed gain when compared to most racing tires – more if you currently ride stiffer gravel or touring tires.
  • Getting a frame that ‘planes’ and gets in tune with your pedal stroke can increase your power output by 5% or more.

Even Greg LeMond won the ’89 Tour not because of marginal gains, but because his aerobars reduced his wind resistance by at least 10% compared to Fignon’s traditional bike.

Aerodynamics enable you to go faster without spending any money: Get into the aero tuck on downhills, and you’ll reduce your air resistance by about 30%. Not only will you go faster on the downhill, but you’ll coast further on the flat (or up the next hill), before your speed drops back to where pedaling is faster than coasting. Coasting more allows you to pedal harder the rest of the time. This is one of the secrets behind riding fast across rolling terrain.

There are other gains that we may consider marginal, but each will make a bigger difference than just a few Watts:

  • A fork that absorbs ‘road buzz’ can save 20-30 Watts on smooth roads by reducing the suspension losses, yet most modern forks are stiff and absorb little shock (above).
  • Wide handlebars increase our air resistance. Some pros use ultra-narrow bars, but even going from 44 to 42 cm bars will make a difference.
  • When we optimize the aero tuck, our knees touch the top tube. Modern frames have wide top tubes, which means that our knees can’t get as close together. Modern cranks put our feet further outward. Both increase our frontal area. It’s probably the reason why BQ’s carbon test bikes descend slower in the aero tuck than our randonneur bikes.
  • Cogs smaller than 14 teeth significantly increase a drivetrain’s resistance, because the upper chain run is under load. By comparison, there is almost no load on the lower chain run, so the savings from extra-large derailleur pulleys are much smaller.

Combining all these small, but significant, gains does make a difference. They are the reason why our fully equipped randonneur bikes are as fast (or faster) than modern carbon bikes on real roads, even though the carbon bikes are a bit lighter.

There is one place where very small gains matter: the weight of the bike. Because a bike consists of so many parts, the way to make a lightweight bike is to reduce the weight of every part as much as possible. The remarkably light weight of the J. P. Weigle for the Concours de Machines (above) – 20.00 lb (9.07 kg) fully equipped – did not come from a few superlight parts, but from every part being as light as possible without giving up strength. Does it matter? Well, every bit counts!

Further reading:

Photo credit: Nicolas Joly (Photo 6).

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 the high-performance components we need for our adventures.
This entry was posted in Uncategorized. Bookmark the permalink.

34 Responses to Myth 15: Marginal Gains

  1. Gert says:

    Are we talking TRUE marginal gains or TUE marginal gains?
    My feel is about 1km/h with Compass Tires at 25km/h compared to other tires. I have no controlled test result.
    Riding as mush as possible in the drops also gives a lot, but it takes practice.

    I have so much weight on my body, that I should start there, and not on the bike

    • When we tested positions in the wind tunnel, riding in the drops reduced the air resistance by 7% compared to the hoods. Plugging those values into http://www.analyticcycling.com, you get a speed increase of 0.75 km/h. Like the faster tires, it really makes a difference.

      Compare this to the power savings of derailleur pulley (reported to be about 2.4 W) – the model at analyticcycling says you’ll go about 0.1 km/h faster.

      • larryatcycleitalia says:

        Interesting point – did your testing use “real” drop bars vs these modern things? The modern “compact” bars seem to offer no real change in position form the hoods to the drops so you see pros stomping into the wind on flat roads with hands on hoods or draped over the bar tops while I can see an instant speed difference on my own “real” drop bars going into the wind vs up on the hoods.

      • We tested a bike with classic deep-drop Philippe Professionel bars. (This shape inspired our Compass Maes Parallel bars.) You are right, modern bars are so short that the drops don’t change the angle of your back, and the two positions are basically the same. With modern bars, it’s possible that riding on the hoods is more aero, as your arms are further stretched forward, which reduces your frontal area.

        What modern racing bikes give up is the more upright position for lower efforts. Modern races are relatively short, and pros don’t need a ‘slower’ position any longer. For the rest of us, it means we have to choose between a ‘fast’ or a ‘slow’ setup, when classic handlebars offer both on the same bike. Especially for long-distance riding, it’s great to have the drops for headwinds and descents, the hoods as an intermediate position, the ramps for slightly more relaxed cruising, and the tops for an upright position.

  2. Johan Brox says:

    Is the argument here that there are so many non-optimized elements in an average road bike that the really small stuff (like pulleys) comes further down the line?

  3. tom larson says:

    Serious question and this is not meant as an insult but im genuinely curious- what’s the motivation for marginal gains at all on a rando bike? As far as im aware, brevets and other such rando events are not races…why the obsession with speed at all? Why promote marginal gains in some areas such as tucking down a hill with knees inward, but then wear a baggy jersey and baggy shorts? And if you are really promoting marginal gains on rando bikes, why not promote the biggest marginal gain of all, aero bars on rando bikes? I guess Im confused with the inconsistencies or what the point is.

    Thanks.

    • Randonneuring is often misunderstood. It’s not a race, but many riders try to set personal bests. It’s like a Marathon run, where most riders aren’t racing for a win, but still try to beat their previous best, or get under 6, 5, 4 or even 3 hours… And randonneuring does have a time limit, so performance is especially important for slower riders. Gaining free speed by tucking more effectively is something many randonneurs appreciate.

      Beyond that, there is a joy in going fast. It also allows us to roam further on our rides.

      By the way, the jerseys aren’t baggy – in part because a close-fitting jersey is more comfortable in addition to being faster. We also tested the knickers: The lightweight fabric doesn’t flap in the wind, and it doesn’t affect aerodynamics much. Even so, when I try to set a time in a brevet, I leave the knickers at home.

      You mentioned aerobars. I used them a lot in the past on shorter distances, where they provide a big advantage. I found that over long distances, their lack of comfort nullifies the speed gains. With the elbows resting on the pads, the suspension of the arms no longer works. In the cross-state races that were popular in Washington, I found myself going faster with aerobars, but requiring longer stops to recover. In the end, my time was the same, and I enjoyed the ride less.

      The quest for most riders is how to go faster without detracting from the enjoyment of the ride. Faster tires and a frame that planes increase the joy of cycling, whether you are trying to set a personal best or not. That they are faster is a nice bonus.

      • Steve Palincsar says:

        At the other end of the pack, slower riders on brevets are concerned with meeting the time limits to avoid disqualification. Quoting from RUSA’s “Rules for Riders,” these are: “(in hours and minutes, HH:MM) 13:30 for 200 KM, 20:00 for 300 KM, 27:00 for 400 KM, 40:00 for 600 KM, and 75:00 for 1000 KM. Additionally, riders must arrive at each checkpoint between the opening and closing time for the checkpoint. These times are noted on the brevet card with the information for the checkpoints.” 13 hrs 30 min for a 200km brevet may seem generous, but that’s overall time, not just riding time and for many at the back of the pack presents a serious challenge.

    • Conrad says:

      I’m with you in that if it isnt a race, I want to enjoy the scenery and sniff daisies and all that. But still, I use nice tires and handlebars, even on the daily commuter bike. Because they are comfortable, the tires have an amazing ride quality and shave maybe 5 to 10 minutes off my commute 🙂 I would never use aero bars unless I’m in a time trial. One because I like to enjoy the scenery and two it is flat out dangerous riding in a pack with aero bars. I know people do it but its a bad idea. If you are in no mans land or off the front in a race or spirited ride you can always just utilize the “phantom aero bars”

  4. Owen says:

    Could you expand on cogs smaller that 14t increasing drivetrain resistance? Isn’t this resistance offset by the increased power these smaller cogs provide (apologies if my terminology is incorrect)?

    • When you think of a cog as a polygon, it’s easy to see that the more corners (teeth) your polygon has, the closer it approximates a circle. And the chain runs smoothest if it forms a circle. Tests by Frank Berto and others have shown that efficiency really drops off below 14 teeth.

      Your power is independent of your gearing. To go faster with the same cadence, you need bigger gears. You can get those either by making the rear cog smaller or by making the front chainring bigger. In other words, a 56×14 is the same 4:1 gear as a 44×11, but it has less resistance.

      • Rick Thompson says:

        Do you have a link to the Berto efficiency data? Since I’m now running an 11 tooth small cog, kinda curious about how bad that is.

      • I only had read the paper version, but it’s now also online at http://www.ihpva.org/HParchive/PDF/hp52-2001.pdf.

      • John C. Wilson says:

        The problem is we are still stuck with 1/2″ pitch chains. And it is not just that small cogs have extra resistance, The high load on each tooth and on each chain pin makes the drive train wear out faster. As early as 1909 Coventry Chain was making 8mm pitch chains and sprockets to suit. Chainrings had 60-70 teeth. Used to great effect by Freddie Grubb and Hubert Opperman. Out of production 1939.

        Any who have used antique 1″ pitch or 5/8″ pitch know they are plain rough. 1/2″ is about the largest that might work at all. Far from optimum. Outside of bicycles and toys chain drive will rarely use a drive sprocket smaller than 16T.

      • Shimano tried 10 mm pitch chains in the 1970s. They offered them for track bikes, but they found that shifting with a 10 mm-pitch chain didn’t work well. I suspect that is the reason why the 1/2″ pitch chain became the prevalent choice.

      • zigak says:

        Friction facts have (I think) better methodology, but have arrived at similar conclusion as Berto in HP (2001). Berto’s article focused more on internal hub gear efficiency and only touched on a classic transmission as a reference point.

        Friction facts on the other hand have focused solely on classic transmission. Takeaways:
        1.) efficiency rises with power and reaches a plateau at around 250W regardless of the gearing combination used.
        2.) cross chaining contributes to additional losses, around 2W at the extremes versus the chain being straight.
        3.) Larger chain ring/sprocket is more efficient than smaller, given the same power, cadence, gear ratio and no cross chaining, e.g. 53×15 vs. 39×11.

        I will not post the link to this excellent article, as the comments with links do not go through the spam filter. Google “friction facts cross chaining chainring size”

      • I couldn’t find the original post – it seems that friction facts now automatically redirects to ceramic speed, the maker of pulleys etc for marginal gains. I found articles summarizing the Friction Facts findings. It seems that they confirm what Berto and Kyle found.

        The big-big combination is useful to get over the top of a rise to avoid a front shift. Efficiency may be slightly smaller, but a front shift will lose more time than 2 W in power losses. It also stresses the drivetrain components more, so climbing mountain passes in the big-big combination is best avoided.

  5. marmotte27 says:

    Once more an example of a buzzword from professional cycling that doesn’t make any sense for the cycling of 99% of all riders, just like ‘stiffness’ in the latest craze but one. Thank you for calling this out as the simple marketing ploy it is.

    • Mike says:

      Agreed. I’d heard of marginal gains in the books that I’d read on pro cycling. I’d never apply such a thing to my own riding. I’m more concerned with things like reliability, robustness, and cargo capacity than in squeezing every last drop of performance out of my bike. Trust me, your bank account will drain a lot faster than your times will improve if you take “marginal gains” to their (il)logical conclusion…

  6. karenashg says:

    Re: compact bars vs. traditional bars, I think it’s important to consider the cyclist’s size. At 5′ 4″ with a short torso, compact bars give me the option of going from being pretty upright on the tops, to laying my torso flat along my thighs in the drops. On a bike that had traditional drop bars, I was either ridiculously upright on the tops if I was nice and low in the drops, or if I put the tops at a good height, could barely reach the drops (and certainly couldn’t bend my elbows for a better aero position through my arms).

    For me, the very short reach from the tops to the hoods or to the drops allows for a very distinct change in my torso angle–if I ride with a taller person who has traditional drop bars, the angle changes are very similar.

  7. Cary Blackburn says:

    To be fair to Sky and British Cycling who started the Marginal Gains bandwagon they always did the basics right before doing the marginal stuff. It also encompassed every aspect of the team like taking along riders own mattresses and pillows to the overnight stops on the Tour. Anybody who’s attempted to sleep in some of the French motel chains would probably think that was more than a marginal gain!

  8. Stuart Fogg says:

    Have you tried the Lauf Grit fork? it offers only 30mm travel but doesn’t weigh much more than a stiff road fork and might be good on relatively smooth roads.

    • Hoping to try one soon, even if it looks like a complicated way to achieve what an optimized steel fork does naturally.

    • zigak says:

      I would be really interested in performance gains from a a specialised road suspension fork like cannondale slate has.

      Re: Lauf fork, I read a comment somewhere that it is a great fork for road and gravel, but for something rougher like xc it lacks the damping aspect of a regular mtb fork.

      • Damping is necessary when you have a lot of suspension travel, but it also consumes energy. (Damping converts the energy of bouncing into heat.) A steel road fork doesn’t need damping, because the suspension travel and the frequency at which it tends to vibrate don’t cause bouncing. I suspect the Lauf fork to have similar characteristics.

  9. jasonmiles31 says:

    In my riding I find that very small changes to the bike can have a very large impact on the way I think during my ride. Something as simple as a headset top cap color or a leaf stuck to my fender can really distract me. These distractions can add up to a big change in performance.

    I would argue the marginal real gains from oversized derailleur pulleys can be totally overshadowed by psychological gains.

    Hopefully all the racers at Kona don’t read this blog, you might ruin their snake oil right before the big race.

  10. PedalWORKS says:

    I have your 35 mm tires on an all road bike made by Mariposa and love them. You are right. Wider is faster. I am so impressed I have ordered your wider tires for my other road bikes. Well done!

  11. Andy says:

    Fo the non-racers like me out there: A bit of headwind or tailwind, whether I feel well rested, or mentally in the zone: all these things make far more difference to my personal records than ceramic bearings and aero wheels. A good day is a combination of strength and luck. There are just way too many variables like road surface and stoplights that go into a real world ride, so if your mission is to squeeze out 0.1% of an advantage on a Strava time, there is always that person who can do 5% because they are simply stronger.

  12. Conrad says:

    The wonderful thing is that the things that matter the most are inexpensive. truly supple tires are not quite as durable but when run at a reasonable pressure the price per mile is still way under any 23mm tire out there. Ever notice how nobody uses the drops on modern bars? Its because they are uncomfortable and unnatural. With a good handlebar you will actually use the drops, which is a bigger aerodynamic improvement by far than any set of aero wheels available. Big props to Compass for delivering high performance products that are much more reasonably priced than the current standard fare!

  13. pdd83 says:

    Agree on the narrow bars, but I do prefer a modern compact bend and very thick bar tape. I know for many, thin bar tape helps them keep a loose relaxed grip, but I prefer thick for the same reasons. One thing that I have wondered about is the aerodynamic impact of the front bag. I run a front bag on my long adventure rides because of their practicality, but have recently wondered what the impact is as I consider some mass start events next season.

    • We tested a handlebar bag in the wind tunnel, and the aerodynamics were surprisingly good – much better than a seatbag. It seems that the bag works partially as a fairing – it’s especially effective when you’re in the aero tuck. The full details were reported with our wind tunnel tests – see the link in the post above to the back issue.

  14. Eric Daume says:

    I’m with you on this post right up until the closing comment. Isn’t weight itself a marginal gain? As you’ve pointed out in the past, the bike is maybe only 15% of the total (bike+rider) package. Saving a pound or two off the bike isn’t much in terms of overall percentage.

Comments are closed.