This is great input.

I would also add that I assume the OP is talking about bike FOP. Because being run FOP is what REALLY matters and CDA is irrelevant ;)

Personally this is why the "ban tri/TT bikes" arguments always get me so riled up. I enjoy a sport or activity where I have to THINK and problem solve and not just hammer.

It amazes me how much aero unconsciousness actually exists in the MOP. When literature always states the biggest drag comes from the rider, and the internet is full of pictures of "fast positions", so many are giving up on free speed. Of course testing is the best method, but so many people would see improvement with judicious use of a mirror and smart phone recording to the side, and a bit of tinkering.

This is a great point. The less aero unconscious you are the more you're going to penalize yourself. The aero conscious know that they need to be still/locked in/dialed in, whatever you want to call it. The unconscious don't think about these things. That's probably worth .015, minimal, right there.

If you're tunnel position is .200 then you're real life position is probably .205-.210. If you're more conscious when racing maybe you get that down .003 from .210 to .207, if less concerned maybe it rises another .004 to .214. Still a ton better than .224

To jkhayc point: Running fast is still the #1 way to KQ, win your AG, win a race, etc. being aero & aero conscious allows you more leeway in case your run isn't there, you have a bad run, drop nutrition etc.

Never underestimate what riding an IM at 500kJ less than you usually do can/will do for your run performance.

---

Brian Stover USAT LII
Accelerate3 Coaching
Insta

One of the things I notice is in out and back courses or courses with multiple loops is that there is a direct correlation between a cyclist's speed and how often they're in the aero position. FOP are almost always in the aero position, and then the percent of cyclists who are not goes up the further you are to the BOP.

Never mind optimizing position once you're aero, for most people I suspect the real gains are just being disciplined about training the position and then holding it through a race.

After that, you can worry about some of the easy stuff, that goes even beyond body position.

• Round bottles on the frame -> move them to BTA and BTS
• Aero Helmet
• Non-baggy clothes
• Not having a million gels taped to the frame
• etc.

The idea with the drag coefficient is NOT that it must be speed-invariant. But it should be a function of some nondimensional property of the flow (or some set of them) -- that is, a number that describes how the flow behaves that has no units or dimensions. In fluid mechanics, use of nondimensional numbers is extremely important because it exploits dynamic similarity -- the equations of fluid motion don't directly care about speed or how big you are or how viscous the fluid is; instead they care about the relative sizes of different terms in the Navier-Stokes equations which are determined by ratios of different dimensional quantities ("Navier-Stokes" = Newton's laws as applied to fluids, including pressure and an equation for mass continuity, amount to four partial differential equations in four variables, or five in five variables if there are also density/temperature variations). These nondimensional numbers include the Reynolds number (a measure of how unimportant viscous forces are relative to the inertia of the flow), the Mach number (how close the flow is to the speed of sound in the fluid, and consequently whether shock waves develop), the Rayleigh number (how strongly thermally forced a fluid is relative to the viscosity and thermal diffusion), the Rossby number (how unimportant rotation is relative to nonlinear terms in the equations), the Richardson number (importance of stratification relative to shear kinetic energy) and many others.

Dynamic similarity of fluid dynamics is what allows engineers to test small-scale models of cars, planes, etc in a wind tunnel and extract meaningful information about aerodynamics -- or to use working fluids different from those that the objects will operate in.

When talking about the drag coefficient in the context of cycling, the primary nondimensional number of interest is the Reynolds number (flows are well subsonic, so the Mach number is too small to be relevant, and temperature differences and rotation are irrelevant so the others I mentioned above don't matter either). Reynolds numbers = U L/\nu, where U is a typical flow speed, L a length scale, and \nu the kinematic viscosity of the fluid, are typically on the order of 10^6 for cyclists -- which corresponds to strongly turbulent wakes (though arguably potentially near the "drag crisis" point others have referred to) and flows where "form drag" dominates (drag caused by pressure perturbations in the flow around the object, rather than by viscous shearing forces in the fluid). A smooth sphere at a certain Reynolds number has the same drag coefficient regardless of its size, the speed of the flow, or the fluid that is flowing around it -- but the drag force may of course be orders of magnitude different due to these factors. (side note: "slippery" is not really the right term for a cyclist who has a low CdA, because skin drag is much smaller than form drag... "streamlined" is more accurate)

Finally, the drag coefficient we usually use, which describes how drag forces (dominantly form drag) scales with flow speed, is itself a function of the Reynolds number (these are the graphs shown above) -- but the variation of drag force with speed U due to the dependence of Cd on Reynolds number is usually much smaller than the dependence of drag force (except, e.g., at very low Reynolds numbers, or near the "drag crisis" point) directly on U^2.
Hope some of this helps.

That was a lot to digest. Wow.

Either way, is it possible folks misconstrue "CdA changes with speed" with "yaw angle changes with speed"?

If you've got a 10mph wind coming at a 45 deg angle to you and you're going 20mph that's a lot different than you going 30mph at it.

I think folks see the delta in speed and CdA and go "aha!!!!". When it's really just a totally different test case, the yaw angle is lower.

To be fair, you'd have to increase the wind speed OR the angle to make the resultant the same for both rider speeds.

If a person always tests even with a Notio at Z2 pace at a higher yaw angle and optimizes for that then goes to try a TT at literally 5mph faster.......that's not the same test point.

This is why I still claim it's a lot easier for a triathlete to aero test than a TT racer. I feel as a TT racer my data needs to be at least done at sweetspot. Which as a very very non-pro person limits me to 60min or so of total test ride time per session. If I want to test and not wreck training plans. Whereas a triathlete could do hours and hours of z2 testing in a day and it just be part of a normal zone 2 training day.

Yes, you're certainly right that yaw angle is another crucial parameter for cycling drag since the bike-rider system is not a sphere (or other shape with rotational symmetry)! I'm not an aero expert as applies to cycling, so I don't know to what extent people try to disentangle yaw angle and speed (via Reynolds #) effects on Cd -- so I'll let others speak more to this. (I'm an atmospheric scientist/climate scientist, so mostly know about all this fluid dynamics as applies to much larger scales!)

This is a good short answer.

In general, the practical benefit of resorting to Cd, Cl, etc., is that we assume that it is close-enough to constant over the speeds of interest, for a fixed configuration, although we know these values change for sideslip angles for cyclists. If this assumption doesn’t hold, then it does you no real service to use the coefficients rather than the measured forces.

I really doubt that a very carefully done wind tunnel experiment, where positions, clothes and straps aren’t uncontrolled, would show anything dramatic in terms of CdA vs velocity from say 15-30mph… slight slope maybe. Primarily because cyclists are never really experiencing laminar flow. We’re riddled with sharp edges, wrinkles, hair, spokes, and cycling through atmospheric air, which has some turbulence intensity itself… so any “large flow changes” aren’t going to be due to mitigating laminar separation as on the smooth sphere.

wow I am learning a whole new set of things with respect to aerodynamics... my mind is blown! I almost stopped reading when you said partial derivatives.... it sent me back to physical chemistry and the horrors of partial derivatives which I wiped from my mind as soon as possible!