Choosing a Rowing Drill: Start From What the Drill Changes

The problem with a drill menu

Almost every rowing resource offers a table. Catch too shallow, do this. Finish too fast, do that. Rush the recovery, do the other one. The manufacturer’s own drill list is the clearest example of a menu with no decision procedure attached ([30] Concept2). The table is easy to use and it is not wrong, but it treats a drill as a name, and a name tells you nothing about why it works or when it will stop working.

The evidence suggests the name matters less than something else. The same drill, taught with a different focus of attention, produces a different result. A drill the rower chooses produces a different result again. A drill that removes a constraint produces a different result again. So the useful unit is not the drill; it is the change the drill makes to how the rower is learning.

There are three such changes, and they are worth separating because they fail in different ways.

1. It changes the constraints — what the rower is physically able to do while rowing.
2. It changes the feedback — what information about the movement is available, and from where.
3. It changes who decides — whether the rower or the coach chooses the difficulty.

A good drill is one where you know which of the three you are buying, because that tells you how to run it, how long to run it, and what failure will look like.

Change one: the focus of attention

This is the strongest and most replicated finding in the whole of motor learning, and it is the one most often lost in rowing advice.

An external focus of attention directs the performer toward the effect their movement has on the environment. An internal focus directs them toward their own body. The classic rowing application: "keep the handle level" is internal — it is a fact about your arm. "Send the handle away from you, out to the end, and let it come back" is external — it is a fact about where the handle goes.

The external focus wins, and it wins by a lot. Directing attention to movement effects enhances learning ([1] Wulf and Prinz). External focus raised automatisation in a complex sport skill rather than only in laboratory tasks ([2] Wulf and McNevin). Moving the external target further away increased the effect ([3] McNevin and Shea). A comprehensive test with an explicit automatisation measure found the same ([4] Kal and van der Kamp). And external focus reduced muscle activity as well as improving accuracy ([5] Zachry and Wulf), which is the detail that matters most for a sport where fatigue ends sessions.

That last point deserves emphasis. The internal cue "keep the handle level" asks the rower to hold something rigid, and holding things rigid costs muscle. The external cue "send the handle away" does not require a rigid arm; it gives the arm a job. The lower EMG in Zachry and Wulf is not a curiosity — it is the likely reason the effect shows up in performance, not just in laboratory measures.

The current synthesis joins this to motivation, which matters because the two interact ([6] Wulf and Lewthwaite). A rower told to focus externally but who does not understand why will drift back to internal cues within a session, and a coach who has not noticed will assume the rower is not trying.

There is a rowing-specific complication, and it comes from the fact that the handle path is already fixed by geometry. The handle hangs from a chain under tension running to a fixed sprocket, so it travels in a straight line whether or not the rower attends to it. An external cue about handle position is therefore partly redundant: the machine enforces it. The useful external cues are the ones the machine does not enforce — where the handle goes relative to the rower's body, how far away it finishes, how quickly it comes back.

Change two: the feedback

Feedback is where coaching most often goes wrong, and the reason is a frequency problem that has been known for decades.

Constant feedback improves performance and harms learning. A rower told every stroke whether the catch was right performs better in that session and learns less, because what they acquire is a dependence on the coach's eye rather than a feel of their own stroke. Winstein's distinction between knowledge of results and knowledge of performance is the origin of the practical advice here ([8] Winstein). The evidence on extrinsic feedback frequency has been reviewed directly for motor learning ([7] van Vliet and Wulf), and the broader human evidence base is set out by Halsband and Lange ([9] Halsband and Lange).

The workable version for a rowing session is simple and slightly counter-intuitive: give feedback rarely, make it specific, and leave the rower alone long enough to find the answer.

The current consensus on delivery and timing supersedes most older prescriptions ([21] Otte and Davids). Two practical consequences for a rower on an ergometer. First, the monitor is a feedback channel like any other, and a constant glance at the split during a technique piece is external feedback at high frequency, which means it is subject to the same discount. Second, the fields available are limited, so a rower self-correcting is doing something narrower than it looks ([33] Concept2, [34] Concept2). Drive ratio and force curve are useful because they are specific; average watts are much less useful as technique feedback, because they move for reasons that have nothing to do with the fault you are drilling.

Change three: who decides

Self-controlled practice means the rower chooses the difficulty. The finding is robust, and the reason is motivational rather than mysterious: choosing your own difficulty gives the learner autonomy, and autonomy support improves motor learning ([16] Lewthwaite and Chiviacowsky). Sanli and Patterson set out the self-determination account directly ([14] Sanli and Patterson).

The boundary condition is what makes this useful rather than a slogan. Self-controlled learning helps when the learner perceives the choice as real ([15] Chiviacowsky and Wulf). A rower told to "pick your own rate" and then corrected for going too hard has not been given control, and the effect will not appear. Self-selected difficulty has been shown to improve learning of a complex task ([17] Andrieux and Danna), and the everyday rowing version is choosing your own pace for a long steady piece — which is why the manufacturer’s pace-finding work functions as a self-controlled drill ([32] Concept2).

Change four: variability, and why you are not trying to remove it

This is the finding most likely to contradict what a rower believes they are doing.

Averages hide the adaptation. Individual trials are noisy in a way that is not error but information ([20] Davids and Glazier). Practice variability improves transfer rather than performance ([13] Newell and Shapiro, [12] Lee and Magill), and the schema account explains why: a schema built from variable conditions has to capture what is common across them, so it transfers to a new condition, while a schema built from one repeated condition captures that condition and does not ([10] Schmidt; the boundary conditions are in [11] Sherwood and Lee).

The practical translation for a rower is uncomfortable. If your technical work is a hundred identical strokes at a fixed rate with a coach correcting every one, you are practising performance and building a schema for that single condition. If instead you vary something each stroke — the rate, the split, the piece you are doing — the average may be slightly worse and the adaptation is much better.

The rowing-specific version of representative practice follows the same logic. The ergometer and the boat are different movement problems ([25] Lamb), so a schema built only on the ergometer transfers to the water partially and a schema built with some on-water reps transfers better. Representative learning design in sport is the formal version of this argument ([19] Pinder and Davids), and the practice-ecology framing behind it is older than the current literature ([18] Handford and Davids). The coach’s role under this view is to design the environment the rower practices in, which is a design problem rather than a correction problem ([22] Woods and McKeown).

Putting it together: a procedure that produces a drill

Now the menu can be replaced by a sequence of questions. The order matters, because each answer rules out whole categories of drill before you spend any time on one.

One: is the fault real, or is it a setup problem? A rower whose foot stretcher is wrong, or whose seat height is wrong, will keep missing the catch while drilling catch. The stroke has a physiological order that a setup fault can break ([24] Secher; [31] Concept2), and equipment setup is cheap to check before anything else. A monitor reading can be wrong in ways that send you drilling a fault you do not have ([33] Concept2).

Two: is the fault a coordination fault or a load fault? If the rower can do the movement correctly at a lower rate, it is coordination, and it belongs to change one or two. If they cannot do it correctly at any rate but can at a higher one, it is a load or sequencing problem, and drilling the pattern will not fix it.

Three: what are you buying? Decide which of the three changes you want, then pick the drill that makes that change and no other. If you want a feedback change, the drill that removes a constraint will not deliver it, and the rower will get a cleaner stroke with no better feel.

Four: what does failure look like? A drill whose failure mode is invisible is not a drill. If you cannot tell from the erg whether the rower is doing the thing the drill was for, you need a cue that makes it visible — which is usually a stronger external target ([3] McNevin and Shea).

Five: how long, and who decides? Set the volume before you start, and let the rower own part of it. On a technique piece, choose the target shape and let the rower hold it; on a long aerobic piece, let them choose the split. The injury literature is the reason volume is not a free parameter ([26] Rumball and Lebrun; [27] Bahr and Andersen), and it is worst when juniors add drilling volume fastest ([28] Smoljanovic and Bojanic).

What drills cannot fix

Three things, because a piece about drill selection is worth little if it implies drills are the answer to everything.

Fatigue. No drill addresses a rower who is too fatigued to execute the stroke correctly. Monitoring shows what is measurable and leaves technique to a coach ([29] Mäestu and Jürimäe), and the injury data makes the distinction consequential.

A stroke that is wrong at the top end. Peak-force sequencing under a 2K is a different problem from a stroke at 20 spm, and intensity work changes the trade-off in ways that need to be trained at intensity ([35] Concept2). What the biomechanics says an ideal stroke is ([23] Soper and Hume) is a target you approach, not one you reach by repetition alone.

Pain that is not a technique fault. Rib stress and lower-back problems in rowers are load problems with a technique component, not pure technique faults ([26] Rumball and Lebrun; [27] Bahr and Andersen). Drilling through pain is the one intervention on this list that reliably makes things worse.

What to do tomorrow

Pick your current pet fault and write down which of the three changes you are trying to buy, before you pick a drill. If you cannot name it, you are choosing a drill by name, which is the thing this article is arguing against.

Then change one cue. Find the rower’s most common internal cue — "keep the handle level", "squeeze the handle", "sit tall" — and replace it with an external one about where the handle goes. Nothing else changes: same drill, same volume, same intensity. Run it for two weeks and compare the force curve rather than the watts, because that is where the EMG difference in Zachry and Wulf should show up.