By Jon Johnson

It’s been awhile since my last post. There’s a reason for that.  I got confused, and perhaps a little obsessed, with the current topic, drafting.  Confusion and obsession are nothing new with me, but drafting played to both of these unfortunate tendencies because: a) it’s really important to any kind of bike riding that allows it, b) it has to do with aerodynamics (a pre-existing obsession), c) it’s really complex, and d) no one has yet researched or written about it in a particularly clear and compelling way.  The latter is a little surprising, because I’d argue that drafting is almost singularly responsible for making mass start road races what they are.

Without drafting, all road races would simply be variations on time trials.  There’s nothing wrong with time trials—they are the races of truth after all, at least if the question is “How hard can you go for x kilometers?”—but the dynamics introduced by drafting make things much more interesting, and entertaining, especially if your forte happens to be something other than super high lactate threshold.  Drafting is why we ride around in cohesive clumps of humanity and machinery.  It accounts for the sprinty-marathon nature of our sport.  Because of drafting, we can ride much faster together—even when riding with competitors—than alone.  “Free riding” isn’t just a metaphor for us.  It also makes us more interesting people than our dedicated time trialist brethren (and sistren), in a volatile, hollering, conniving, and conspiratorial kind of way.

So I was surprised to find very few good studies on the subject.  It wasn’t for lack of looking.  That’s not to say that a lot hasn’t been written about drafting.  It has, but no one has come up with a solid empirically validated model to help understand what exactly is going on, what matters, and how by much.  Several basic, useful generalizations did emerge, many of which are summarized in the diagram below.  For instance, there is no doubt that you get more benefit by riding closer to a leading rider, though just how much isn’t clear—maybe up to 50%, but you’d have to ride dangerously close to get that.  (More on this below.)  This is easy to imagine if you’ve ever caught a whitewater eddy.  Drafting is essentially catching a wind eddy, after all.  The current immediately behind the rock is strong, and tails off as you move downstream.

I was surprised to learn that some aerodynamic advantage may be found as far as 10 meters back.  Equally surprising: The lead rider may actually realize small aerodynamic savings from having a close trailing rider, who can provide a slight “push” by filling in some of the rearward low pressure pocket.  Position in the group in relation to the rider ahead of you also matters.  Although there is some disagreement on the particulars, there appears to be a small additional advantage for riders in rearward positions in a pace line—an extra 5% for rider #3 over rider #2 in the diagram—but this levels off quickly, by about rider 4 or so.  Riding in the middle of a group can save a boatload of watts, anywhere from 40% to upwards of 65%.  Absent a side wind, it’s best to ride aligned directly behind the cyclist ahead of you—as little as 10-20 cm (4-8 in.) to the left or right can scrub off considerable savings.  (I’ll have more to say about drafting in windy conditions in a separate post.)  Finally, cyclists riding too closely side-by-side, less than about a half a meter, may actually increase each others’ drag.  It’s best to stay at least a meter to the side of a rider you’re overtaking.

I would love to be able to tell you precisely how many watts you can save by riding at a certain distance and velocity, but there simply isn’t a great predictive model out there.  (Much more on the subject is over at the ovto.net site.)  Conventional wisdom holds that drafting behind a single rider gives about a 30% advantage, which is more-or-less in line with the few empirical studies that have been conducted.  Most have been conducted in carefully controlled conditions in wind tunnels and indoor and outdoor velodromes.

I was curious about how these findings would generalize to real world rides, with their swirling and gusting winds, continuous changes in position and distance between bikes, variations in speed, and the like.  So I did a little experiment of my own, collecting data from four of our group rides that I also videoed.  I then coded the files for periods where I was pulling and drafting.  (For those who worry about whether I have a life, this didn’t take much time.  Video editing software allows you to move through files very quickly.  This doesn’t necessarily mean I actually have a life, of course, but it’s not because I’m staring at hours of video while coding 0’s and 1’s.)

So anyway, after controlling for things like gradient, velocity and the like, at the speeds and in the formations and conditions we typically ride, I found that on average, drafting gave about a 25% power savings.  Some days were a little more, some a little less, and we don’t frequently ride in bunches that completely surround cyclists, so I didn’t get good estimates for those conditions.  That said, when considering all the sources of aerodynamic noise we encounter out on the road, 25% strikes me as a realistic estimate for riding in an actual Ozark paceline.  It’s the estimate I’ll use in the examples below, but keep in mind that there is an inherent nondeterministic component—your mileage may vary.

At 25%, where a lead rider would be pushing about 300W to ride 25 mph, a drafting rider would needs only about 225W.  This simple fact has all sorts of consequences for riding bikes, one of which is our decided tendency to clump together.  Generally speaking, as long as a rider can maintain 75% as much power as the stronger riders in the group, he or she can adhere to the pack.  It’s only when aerodynamics become inconsequential that groups become less adhesive.  Big climbs serve to break up the pack because, as we learned before, when velocities slow, aerodynamic drag is quickly dwarfed by gravity, and there is no such thing as gravity drafting.  (If only.)  It’s also why breaks so often begin with rapid accelerations that go wide of the field.

The flip side of clumping is rapid separation.  Drafting is why getting popped can be such dramatic event.  Generally speaking, once off the back an individual is done, at least until the next regrouping point, which in a race is just past the finish line.  For a redlined rider, 12 meters off the back might as well be 12 kilometers.  Unlike in time trials, where differences in fitness will create predictably proportional differences in distances covered, the dynamics of drafting create discontinuities that don’t neatly correlate with conditioning.

A paceline is another obvious draft dependent phenomenon.  Pacelines are great examples of nonzero sum games, instances where cooperation, even among competitors, can benefit all.  At 25%, a two man paceline riding at 25 mph will average (300W + 225W) / 2 = 262.5W, versus the 300W each cyclist would need if riding alone.  If they’re working well together, that’s 37.5W savings for both.  And the advantage improves with additional riders—three riders average 250W, and by the time you get to 20 riders, the average power required is 229W, only 4W more than required to draft full time.

This is all well and good, but it happens nonlinearly, and each rider added to the group contributes relatively less than the one before.  As the graph below shows, the marginal advantage of adding riders to a paceline flattens quickly.  You get less than two additional watts savings by adding a seventh rider.  Though the specifics vary a little, the same principles apply to double and rotating pacelines—large initial power savings flatten as more riders are added.

In addition to lowering average power requirements, larger groups provide another important advantage, at least up to a point.  Humans generate power through three basic energy systems—creatine-phosphate, anaerobic, and aerobic, the first two of which are very short lived.  You can ride anaerobically for only about 2 minutes before requiring a few minutes of aerobic recovery, which allows for another anaerobic surge.   Two riders taking 1 minute pulls will only have 1 minute of recovery while drafting—less, actually, because it takes time to switch places—not nearly enough time to recover from an anaerobic effort.  But recovery time goes up with every additional rider.  Riders sharing 1 minute pulls in a 6 person paceline will have 5 minutes of recovery for every 1 minute of anaerobic effort.   As with average power, marginal gains for recovery go down with additional riders—the difference between 5 and 6 minutes of rest isn’t nearly as important as the difference between 2 and 3.

Pacelines form in all kinds of situations, but they play an especially important role in breaks, instances of cooperating competitors that introduces yet another important factor.  The popular vision of an ideal paceline—as seen, for instance, in team time trials—consists of cyclists going as fast as possible because every rider takes equal turns riding hard in the wind.  Outside of team time trials, of course, this isn’t all that common. Free riding happens.  Decades of research has shown that the emergence and maintenance of the kinds of cooperation necessary to make paceline-like efforts in breaks work is very sensitive to free-riding.  People generally do not put up with freeloaders—surliness ensues, and it usually only takes one or two slackers to trigger a cascade of defection that can bring cooperation to a grinding halt.  For this reason, and because marginal advantages flatten out relatively quickly, breaks become geometrically more difficult to form and manage as group size grows beyond just a few riders.  Riders’ willingness to cooperate also declines as they approach a finish line.  (Breaks are classic game theoretic dilemmas, about which I’ll have more to say later.)

Finally, there are pelotons, which are in some ways more complex than pacelines, and in other ways not so much.  Complexities aside, because the drafting advantages from riding inside the pack are so substantial, and because expectations for taking pulls in pelotons are more relaxed, it’s much easier for a rider to get to the end of a long race with fresh legs.  (Jeukendrup reported that one cyclist, riding at 25 mph in the middle of the pack on a flat stage in the Tour de France, averaged just 98W.  That’s a savings of 65%.)  At any point in time, a large peloton will usually have strong, unspent riders available to head up a chase, if it can get organized to do so.  More commonly in amateur races, on flat courses that keep the group together, there will often be lots of riders who get to the final 200 meters with plenty of energy for the final sprint.

These are just the basics, of course, and drafting plays into plenty of other tactical considerations and nuances.  The tactics of attacks on flats and on hills, launching final sprints, and when and how to take pulls and sit in are all fundamentally dependent on drafting, some of which I’ll take up in future posts.  In the meantime, I suspect there are lots of strategic advantages to be had by racers who pay appropriate attention to gaining full advantage from surfing others’ wakes.

For more on Jon and his writing check him out click here.