The 30-Minute Rowing Critical Power Test: How to Measure CP and W' on the Erg

Ask a rower what their threshold is and most will hand you a percentage of 2k split, discounted by ten or fifteen seconds per five hundred meters because a plan they downloaded once said so. The number rarely comes from anything the athlete's body actually produced under controlled conditions. It is arithmetic layered on top of a single all-out race effort, one that measures total work capacity far more than steady-state output, then gets treated as a threshold reading.

Critical power (CP) and its partner metric W' close that gap. CP is the highest power a rower can hold in true steady state, the ceiling before anaerobic metabolism has to cover the difference. W' is the finite tank of work, measured in kilojoules, that can be spent at any rate above CP before the split falls apart. A single 30-minute test on the erg, paced deliberately and read correctly, produces both numbers in one sitting, and a full stack of individualized training zones falls out from there with almost no guesswork left.

Why 2k-Pace Percentages Don't Give You a Real Threshold

The flat-percentage approach assumes a fixed relationship between 2k power and sustainable threshold power, and that relationship is not fixed. It depends on how much of a 2k score comes from CP versus how much comes from spending W'. Two rowers can post the same 2k time and share almost nothing else physiologically.

A rower with a large W' and a modest CP, the classic sprint-built athlete, gets a 2k score propped up by a big anaerobic tank drained across six minutes. A flat percentage of that score overstates what they can hold at steady state, so their threshold pieces end in survival mode by minute twelve. A rower with a smaller W' and a higher CP, the engine-first type, gets undershoot from the same math, and their threshold sessions feel more like a warm-up. Neither is training against their actual physiology; both are training against a number borrowed from a test that measures something else.

Equipment and Test-Day Setup

The test needs nothing exotic: an ergometer with a monitor that logs power per minute, a fixed and recorded drag factor, and about an hour of uninterrupted time.

Drag factor matters more than most rowers treat it. It works like resistance on a bike trainer: a setting of 130 versus 110 changes what a given stroke rate produces in watts, so a CP number recorded at one drag factor isn't directly comparable to one recorded at another. Set it once, write it down, and match it on every retest.

Step-by-Step 30-Minute Protocol

1. Warm-up (12-15 minutes): Easy rowing, then three progressive 250m efforts at roughly 70%, 80%, and 90% perceived effort, each followed by 90 seconds of rest.
2. Set and record the drag factor before the clock starts. Most club-level rowers land between 115 and 135; the exact number matters less than testing at the same one every time.
3. Controlled start: Build into the piece over the first 60-90 seconds instead of attacking it like a 2k start. A hard opening sprint burns W' before the test can register it as usable data and wastes the first several minutes of the calculation below.
4. The 30-minute effort: Row at the hardest pace sustainable for the full 30 minutes, aiming flat-to-slightly-positive rather than fading hard late. A small fade in the final five minutes is normal; a collapse means the opening pace was too aggressive.
5. Record power every minute for all 30 minutes, from the monitor's interval log or an app pulling data directly from the erg.
6. Cool down for 8-10 minutes easy.

Total time, warm-up and cooldown included, runs about 55-60 minutes. This is a maximal-effort test; schedule it on a day without a heavy session before or after, since testing during a fatigued week suppresses both numbers below the athlete's real capacity.

Turning 30 Minutes of Splits into CP and W'

CP: average the recorded power across the final 20 minutes, minutes 11 through 30. The opening 10 minutes get left out of this part of the calculation because they still carry an anaerobic contribution and the oxygen-uptake transition into steady state; folding them in would inflate the estimate.

W': for each of the first 10 minutes, subtract CP from that minute's average power, keep only positive results, multiply each by 60 to get joules, sum them, and divide by 1,000 for kilojoules.

W' = Σ(power at minute i − CP) × 60, for i = 1 through 10, positive terms only, converted to kJ.

Worked example: a rower's per-minute watts read 220, 210, 205, 200, 196, 195, 194, 195, 193, and 194 across the first ten minutes, then average 182W for minutes 11-30. CP is 182W. Summing the excess above CP across those ten minutes and multiplying by 60 seconds gives about 10.9kJ of W'. A different rower holding that identical 182W CP but opening harder at 240W and settling more slowly could post a W' north of 16kJ from the same CP, a real difference in anaerobic reserve that a flat percentage of 2k pace would never surface.

What the Research Actually Shows

Cheng, Yang, Lin, Lee, and Wang (2011), testing 18 trained male rowers for the European Journal of Applied Physiology, compared CP from a single 3-minute all-out rowing test against CP derived the traditional way, from three separate maximal constant-work-rate trials on different days. The end-power from the single test (269±39W) came out statistically indistinguishable from the multi-trial CP (272-276W, depending on the model used). That is rowing-specific confirmation that a single maximal erg session can produce a valid CP number, not an extrapolation borrowed from cycling.

Karsten, Petrigna, Klose, Bianco, Townsend, and Triska (2020), in Frontiers in Physiology, ran a comparable comparison on 17 trained cyclists and triathletes: three-trial CP against a single 20-minute time trial discounted by the standard 95% factor. The two tracked tightly (r=0.969, p<0.001), but CP came out significantly higher by a mean of 7±13W, a trivial effect size (d=0.04) with individual gaps as wide as -19 to 33W. A single sustained effort is a legitimate way to estimate CP without three separate maximal visits, but a 20-minute window still undershoots slightly, which is why the protocol above reads CP from the last 20 minutes of a 30-minute effort rather than applying a flat discount to a shorter one.

The caution belongs on W', not CP. Gama, dos Reis, Sousa, and Gobatto (2018), in PLOS ONE, ran the same style of comparison for tethered running and found the CP-equivalent estimate held up well (p=0.486) while the W'-equivalent estimate did not (p<0.001), correlating only moderately with the multi-trial reference (r=0.37-0.52). Treat the W' number from this test as a way to track one athlete's own trend, not a figure to compare against a teammate's or a number pulled from a different published protocol.

Turning CP and W' Into Training Zones

Once CP and W' exist as measured numbers, zones stop being arbitrary percentages of a race score and start being percentages of a real steady-state ceiling.

Zones 5 and 6 are where W' personalizes the session. Two rowers with an identical 200W CP but W' values of 10kJ and 18kJ need different prescriptions at 130%CP (260W): the smaller tank empties in about 167 seconds, the larger one in about 300 seconds, before power has to drop toward CP regardless of effort. The same 4-minute interval asks one of them to hold a pace their own physiology can't sustain past minute three.

W' also refills, slowly, once power drops back below CP, which is the physiological basis for rest periods between intervals. How quickly it recovers varies enough between athletes that it's worth tracking individually rather than assuming one textbook time constant fits the whole roster.

Mistakes That Wreck the Test

Most bad tests trace back to one of five errors, all avoidable with a bit of discipline on setup and pacing.

What Low CP or Low W' Actually Means

A CP that lags teammates with similar 2k scores points at an underbuilt aerobic engine; the fix is more zone 2-3 volume before adding threshold or sprint work. A W' that lags relative to CP points the other way: the aerobic engine is fine, the anaerobic tank is small, and short, hard efforts with full recovery, 6-10 x 30-45 seconds well above CP with 2-3 minutes of rest, move that number faster than another long steady row.

Retest every 6-8 weeks rather than after every block. CP shifts gradually with sustained aerobic and threshold training; W' responds more to high-intensity work and can move faster in either direction, including downward during a base-building phase. CP rising while W' dips slightly during that phase isn't a red flag, it's the expected trade-off.