What a 2K Result Actually Tells You — A Research-Grade Synthesis

Abstract

A 2K result is one number on one day under one set of conditions — and it is the most over-interpreted number in indoor rowing. The number tells the rower what they could hold for roughly seven minutes under that day's sleep, hydration, warm-up, mental state, and pacing plan. It does not tell them what they are capable of holding across a season, what they will hold in a year, or whether they are actually improving. The physiological-determinants literature shows the 2K is a synthesis of power at VO2max, VO2max, lactate threshold, and maximal power ([1] Schabort et al. 1999, Level 2b; [2] Ingham et al. 2002, Level 2b; [3] Russell et al. 1998, Level 2b; [4] Riechman et al. 2002, Level 2b; [5] Cerasola et al. 2020, Level 2b; [7] Ingham et al. 2013, Level 2b; [8] Bourdon et al. 2009, Level 2b). The reliability literature shows day-to-day variation in trained rowers is roughly two percent of mean power, so a 2K improvement of one second per 500 m sits inside the noise band of a single test ([1] Schabort et al. 1999, Level 2b; [9] Stine et al. 2019, Level 2b). The pacing literature shows rowers settle into a conservative-start, end-spurt pattern after a habituation trial — a first 2K underestimates what a familiarised rower can hold ([6] Gee et al. 2013, Level 2b; [18] Concept2 — Pacing blog, Level 5). The ranking-and-percentile literature shows the 2K is only meaningful when read against the right reference population: age, sex, weight class, and adaptive category all matter ([13] Concept2 — Rankings, Level 5; [14] Concept2 — Ranking help, Level 5; [15] World Rowing — Indoor rules, Level 5; [16] British Rowing — Go Row Indoor tests, Level 5). The bottom line: the 2K is a useful anchor for the AI coach's reference pace, but it is one data point among many. The rower's full test history, training distribution, and stated context all reshape how the coach reads it ([10] Seiler & Kjerland 2006, Level 4; [11] Sylta et al. 2014, Level 4; [12] Rosenblat et al. 2025, Level 2a; [26] Ebert et al. 2007, Level 2b; [27] Otter et al. 2015, Level 2b; [28] Filipas et al. 2022, Level 2b; [29] Bouchard et al. 2015, Level 5).

What the 2K is and isn't

The 2K is a six-to-eight minute maximal effort on a rowing ergometer, governed by World Rowing's indoor-rowing rules ([15] World Rowing — Indoor rules, Level 5). Concept2 is the standard equipment at World Rowing indoor events and at virtually every national indoor championship, and the ruleset enforces the conditions that make a 2K comparable: fixed-distance pieces on a level surface, a non-moving flywheel start, no performance-altering modifications, free damper settings, and a preset distance or duration ([13] Concept2 — Rankings, Level 5; [14] Concept2 — Ranking help, Level 5). The piece is short enough to demand a maximal effort, long enough to require sustained aerobic power, and constrained enough that the result is reproducible when conditions are matched ([1] Schabort et al. 1999, Level 2b; [13] Concept2 — Rankings, Level 5).

The 2K is not a VO2max test, although it correlates with VO2max at the population level. It is not a lactate-threshold test, although lactate threshold contributes independently to the prediction. It is not a maximum-power test, although peak anaerobic power contributes independently too ([2] Ingham et al. 2002, Level 2b; [4] Riechman et al. 2002, Level 2b; [7] Ingham et al. 2013, Level 2b). It is a synthesis test: the result is whatever combination of aerobic power, anaerobic power, lactate tolerance, pacing strategy, mental toughness, and test-day conditions the rower brings to the piece. Reading a 2K well is reading the synthesis.

The most common mistake is treating the 2K as a single physiological number — "my VO2max is X" or "my lactate threshold is Y." The 2K gives you the score, on the day, under the conditions; the VO2max, the lactate threshold, and the reference pace have to be inferred, and the inference is best when it uses the 2K alongside the rower's other test history and training log ([21] Hagerman 1984, Level 5; [16] British Rowing — Go Row Indoor tests, Level 5).

The 2K as a synthesis test: what the physiological literature actually says

The 2002 [2] Ingham, Whyte, Jones, and Nevill paper is the strongest single anchor for what a 2K actually measures (Level 2b). The authors took finalists from World Championship rowing and sculling — nineteen male and thirteen female heavyweight rowers, plus four male and five female lightweight rowers — and ran them through a discontinuous incremental rowing test to exhaustion (blood lactate threshold, VO2max, power at VO2max), a maximal-stroke test (maximal force, maximal power, stroke length), and a 2K ergometer time trial. Power at VO2max, maximal power, and maximal force correlated with 2K speed at r = 0.95 (p < 0.001). VO2max alone correlated at r = 0.88, and VO2 at the blood lactate threshold correlated at r = 0.87. A stepwise regression model — power at VO2max, VO2 at lactate threshold, power at 4 mmol/l blood lactate, and maximal power — explained 98% of the variance in 2K speed, and validation in eighteen elite rowers gave limits of agreement of −1.5 to +6.9 seconds on the 2K time. The read is unambiguous: the 2K is dominated by power at VO2max, with VO2max and lactate threshold adding real but smaller contributions, and the noise band on a single test is a few seconds either way.

The 1998 [3] Russell, Le Rossignol, and Sparrow paper (Level 5) tested nineteen elite schoolboy rowers and found body mass, VO2max, and knee-extension strength each correlated with 2K time (r = −0.41, −0.43, −0.40); the combined anthropometric-and-metabolic model reached R = 0.82. The 2002 [4] Riechman, Zoeller, Balasekaran, Goss, and Robertson paper (Level 5) tested twelve competitive female rowers and showed peak Wingate power, VO2max, and Wingate fatigue together explained 96% of 2K variance. The 2020 [5] Cerasola et al. paper (Level 5) found sixty-second all-out mean power was the single strongest predictor of 2K time in elite youth rowers (r = −0.943). The 2013 [7] Ingham et al. paper (Level 5) reported maximal minute power correlated with 2K power at r = 0.98, and power at VO2max at r = 0.96. The pattern is consistent across populations and decades: the 2K is a synthesis of aerobic and anaerobic power, with the anaerobic contribution (peak power, sixty-second mean power, maximal stroke force) carrying more weight than a "VO2max test" framing would suggest.

The 2024 [22] Astridge et al. paper (Level 5) added a cross-distance comparison, showing how the 2000m pacing strategy and energy contribution differ from the 1500m. The 2019 [23] Turnes et al. paper (Level 5) showed the deoxygenated-haemoglobin breaking point correlates with 2K performance, which supports the case that the 2K recruits both central and peripheral oxygen-delivery systems. The 2018 [24] Turnes et al. paper (Level 5) — the ischaemic-preconditioning paper — is an existence-proof that even small acute interventions can shift 2K time by several seconds, which is a useful reminder that the result is not a fixed physiological number.

The honest read: a 2K is a window into roughly four physiological capacities at once, with peak power and power at VO2max carrying the most weight and VO2max and lactate threshold contributing meaningful but secondary variance. A reader who treats the 2K as a single number misses what the literature actually says; a reader who treats the 2K as a multi-system synthesis is reading the number correctly ([2] Ingham et al. 2002, Level 2b; [4] Riechman et al. 2002, Level 2b; [5] Cerasola et al. 2020, Level 2b; [7] Ingham et al. 2013, Level 2b; [21] Hagerman 1984, Level 5; [23] Astridge et al. 2024, Level 2b).

The 2K as a lactate-tolerance test

The 2K is more than a power test. Lactate accumulates substantially during a six-to-eight minute maximal effort, and the rower's ability to tolerate that lactate — to keep producing high power as pH drops — is part of the result ([21] Hagerman 1984, Level 5; [24] Turnes et al. 2019, Level 2b). The 2002 [2] Ingham et al. paper (Level 2b) showed VO2 at the blood lactate threshold and power at 4 mmol/l blood lactate each contributed independently to the 2K prediction in elite rowers — the threshold matters, and not just as a proxy for VO2max. The 2009 [8] Bourdon, David, and Buckley paper (Level 5) — the 2-in-1 test — showed lactate-threshold testing and 2K performance testing can be combined into a single session with no significant change in either measurement, which is the practical anchor for reading both lactate threshold and 2K from one test session.

The consequence for the rower is direct: a high VO2max with a poor lactate threshold produces a disappointing 2K; a moderate VO2max with a strong lactate threshold produces a 2K that exceeds what the VO2max would suggest alone. The AI coach that anchors reference pace purely on VO2max misses this. The coach that anchors reference pace on the 2K directly, and lets the rower's lactate threshold emerge from the training log, reads the rower more honestly. The 2003 [21] Yoshiga and Higuchi paper (Level 5) added the sex-specific data: lactate kinetics during a 2K differ between male and female rowers, and within-sex variance is large enough that a single 2K does not pin down the rower's lactate-tolerance ceiling.

The practical read: the 2K is a lactate-tolerance test whether or not the rower measures blood lactate, because lactate tolerance is built into the result. A rower who has done too much lactate-tolerance work and too little aerobic base will see a 2K that improves faster than their 5K; a rower who has done the opposite will see the reverse pattern. The 2K is a window into both, and the AI coach that reads the 2K alongside the 5K is reading both windows together ([2] Ingham et al. 2002, Level 2b; [8] Bourdon et al. 2009, Level 2b; [22] Yoshiga & Higuchi 2003, Level 2b).

The 2K as a pacing test

The 2K is more than a physiological test. The pacing literature is unambiguous: how the rower distributes effort across the piece is a substantial part of the result, and a poor pacing plan can cost more seconds than a poor physiological state ([6] Gee et al. 2013, Level 2b; [18] Concept2 — Pacing blog, Level 5).

The 2013 [6] Gee, French, Gibbon, and Thompson paper (Level 5) put fourteen well-trained male rowers through three 2K trials at three-to-seven-day intervals, measuring power, stroke rate, heart rate, and lactate per 500 m. Performance was unchanged across trials, but the pacing strategy shifted: trial 1 was a positive split (higher starting power, progressive decrease), while trials 2 and 3 settled into a reverse-J shape — conservative start, end-spurt with elevated power in the final 500 m. Typical error in 2K performance was 2.4%, and the authors concluded that intervention studies using 2K performance need a habituation trial before reliable measurement. The rower who has only done one or two 2Ks is leaving several seconds on the piece through pacing alone; the rower who has done five or more has a stable pacing pattern and is testing physiology, not pacing.

The 2019 [9] Stine et al. paper (Level 5) — the feedback-type paper — showed that feedback framing during a 2K effort shifted performance, supporting the case that the rower's mental context (cues, audience, framing) is a real source of variation. The 2024 [22] Astridge et al. paper (Level 5) compared 2000m and 1500m pacing and showed the 2K's pacing profile is a distinct pattern from other ergometer race distances. The 2007 [24] Ebert et al. paper (Level 5) — the dehydration paper — showed 2% body-mass loss reduced time-to-exhaustion by 28% in trained cyclists, which transfers to the 2K as a hydration-state variable.

The practical read: a 2K is a pacing test as much as a physiological test, and the pacing side is trainable. A rower who has done two 2Ks and is comparing them is comparing a familiarised pacing pattern against a less familiarised one; a rower who has done eight 2Ks over six months is comparing familiarised pacing patterns, and the difference reflects physiology rather than pacing skill. The AI coach that anchors reference pace on a single recent 2K is anchoring on whatever blend of physiology and pacing the rower happened to bring. The coach that anchors reference pace on a moving average of the rower's most recent two-to-three 2Ks is anchoring on the more stable signal ([6] Gee et al. 2013, Level 2b; [9] Stine et al. 2019, Level 2b; [18] Concept2 — Pacing blog, Level 5; [20] Concept2 — Warm-up blog, Level 5).

The 2K as a reliability test: how much does the result actually move?

The 1999 [1] Schabort, Hawley, Hopkins, and Blum paper is the reliability anchor (Level 5). Eight well-trained rowers (peak VO2 61 ± 5 ml/kg/min) rowed three 2000-m trials on a Concept II ergometer at three-day intervals. Performance improved 2.3% from trial 1 to trial 2 (95% CI 0.1 to 4.5%) and 0.9% from trial 2 to trial 3 (95% CI −1.4 to 3.3%). The coefficient of variation for mean power was 2.0% (95% CI 1.3 to 3.1%), the retest correlation was 0.96 (95% CI 0.87 to 0.99), and the variability in completion time was roughly one-third that of mean power — simulated velocity on the ergometer is proportional to the cube root of power. The read is direct: a rower who rows two 2Ks a few days apart and sees a one-second-per-500m difference is inside the noise band; a rower who sees a five-second-per-500m difference is outside it and probably has a real change.

The 2013 [6] Gee et al. paper (Level 5) corroborated: typical error across three trials in trained rowers was 2.4%, with low-to-moderate variability for physiological variables (typical error 1.4 to 5.1%) except peak lactate, which was 11.5%. The 2015 [25] Otter et al. paper (Level 5) reported a submaximal rowing test that predicted 2K time with ICC between 0.91 and 0.99 — the inverse question (how stable is the predictor, not the 2K itself), but reinforcing the same point: physiological measures around rowing have small noise bands, and the 2K is no exception.

The practical read: a 2K result is reproducible to within roughly two percent of mean power in trained rowers, which translates to roughly one-to-two seconds per 500 m on the time itself. Improvements smaller than that are inside the noise band of a single test; improvements larger than that are likely real. The AI coach that updates reference pace based on a single recent 2K may be updating on noise. The coach that updates on a moving average of the most recent two-to-three 2Ks is updating on signal ([1] Schabort et al. 1999, Level 2b; [6] Gee et al. 2013, Level 2b; [27] Otter et al. 2015, Level 2b).

The 2K and the 5K: the ratio that tells the training story

The 2K is not the only test. The 5K is the longer companion, and the 2K/5K ratio — expressed as a pace ratio or a split difference — tells the rower something the 2K alone does not. A rower whose 2K split is much faster than their 5K split is doing well on anaerobic capacity but may be under-trained on aerobic endurance. A rower whose 2K and 5K splits are roughly equal is doing well on aerobic endurance. A rower whose 5K split is faster than expected from their 2K is in the unusual position of being better at aerobic power than at anaerobic power — which is rare and worth investigating.

The polarized-training literature gives the framework. The 2006 [10] Seiler and Kjerland paper (Level 5) is the landmark: elite endurance athletes train roughly 75% below VT1, 7-8% between VT1 and VT2, and 17-22% above VT2. The 2025 [12] Rosenblat et al. network meta-analysis (Level 2a) confirmed that polarized distribution outperforms pyramidal and threshold models on VO2max and time-trial performance. The 2022 [26] Filipas, Bonato, Gallo, and Codella paper (Level 2b) compared pyramidal and polarized training directly over sixteen weeks in trained endurance runners and found polarized produced the better VO2max and time-trial outcomes. The 2014 [11] Sylta, Tønnessen, and Seiler paper (Level 5) validated self-reported training against HR-monitor data (r = 0.99 for duration) — the methodological anchor for trusting the rower's own training log when the coach evaluates whether the rower is actually training in a polarized distribution.

The practical read: a rower with a 2K in the top decile of their age-and-sex reference population and a 5K in the middle is doing too much lactate-tolerance work and not enough aerobic base. A rower with a 2K in the middle and a 5K in the top decile is doing the opposite. The AI coach that reads the 2K alone and prescribes more 2K-style work for the first rower will deepen the imbalance; the coach that reads the 2K alongside the 5K and prescribes more aerobic volume for the first rower will move the training distribution toward the polarized model the literature defends ([10] Seiler & Kjerland 2006, Level 4; [11] Sylta et al. 2014, Level 4; [12] Rosenblat et al. 2025, Level 2a; [28] Filipas et al. 2022, Level 2b).

Reading a Concept2 percentile ranking

The Concept2 Logbook rankings are the most widely cited reference population for indoor rowing ([13] Concept2 — Rankings, Level 5; [14] Concept2 — Ranking help, Level 5). The ranking system is stratified by ergometer type, movement type (RowErg vs. Dynamic vs. Slides), event distance (2000m, 5000m, and so on), season, age range, weight category (heavyweight vs. lightweight), sex, country, state, verification level (unverified, verified, race-only), and adaptive category (PR1, PR2, PR3, FES variants). The ranking rules enforce the conditions that make a 2K comparable: fixed-distance pieces, level surface, no performance-altering modifications, free damper settings, preset distance or duration, and a non-moving flywheel start. Verification levels matter: an unverified piece is the rower's word; a verified piece has been independently observed; a race-only piece is from an officially sanctioned competition ([13] Concept2 — Rankings, Level 5; [14] Concept2 — Ranking help, Level 5).

Reading a percentile ranking well requires reading it against the right reference population. A 2K at the 90th percentile of "all male heavyweight rowers aged 19-29" is a strong score but not exceptional. The same 2K at the 90th percentile of "all male heavyweight rowers aged 50-59" is exceptional. The same 2K at the 90th percentile of "all female lightweight rowers aged 19-29" sits in a different competitive context entirely. The age brackets are predefined (0-12, 13-18, 19-29, 30-39, and so on), the weight categories are binary (heavyweight vs. lightweight), and the sex categories are binary (male vs. female). The reference population is the filtered slice, not "all rowers" ([13] Concept2 — Rankings, Level 5; [14] Concept2 — Ranking help, Level 5).

The manufacturer guidance is consistent: the 2K test is a check of overall fitness, the percentile is a population comparison, and the more specific the filter the more meaningful the comparison ([17] Concept2 — 2K test blog, Level 5; [18] Concept2 — Pacing blog, Level 5). The 2015 [27] Bouchard, Blair, and Katzmarzyk paper (Level 5) is the public-health framing: cardiorespiratory fitness is a mortality marker, and a low 2K score is associated with elevated all-cause mortality risk. The population-level reading is real; the individual-level reading still requires the rower's full test history and the AI coach's context ([15] World Rowing — Indoor rules, Level 5; [16] British Rowing — Go Row Indoor tests, Level 5; [29] Bouchard et al. 2015, Level 5).

The 2K under different conditions: what shifts the result?

The 2K is reproducible when conditions are matched, and the day-to-day variation in trained rowers is roughly two percent ([1] Schabort et al. 1999, Level 2b). What shifts the result beyond that noise band is a combination of controllable and uncontrollable variables.

Hydration is the most-studied controllable variable. The 2007 [24] Ebert et al. paper (Level 5) showed 2% body-mass loss from dehydration reduced time-to-exhaustion by 28% in trained cyclists; the equivalent effect on a 2K is several seconds. The [13] Concept2 — Rankings (Level 5) and [20] Concept2 — Warm-up blog (Level 5) reinforce that pre-test hydration and warm-up are the rower's controllable variability sources.

Damper setting is a rower-choice variable that shifts the result by several seconds even when physiological state is unchanged ([19] Concept2 — Damper setting blog, Level 5). Concept2's own guidance is that the damper setting should be matched to the rower, not to the test, because the test standardises distance and time but not damper choice.

Warm-up structure is a controllable variability source. The [20] Concept2 — Warm-up blog (Level 5) gives manufacturer guidance on a structured warm-up before a 2K; the rower who rows the 2K cold leaves seconds on the piece through muscle temperature and activation state, not through physiology.

Acute interventions like ischaemic preconditioning have been shown to shift 2K time ([25] Turnes et al. 2018, Level 2b), which is the existence-proof that the 2K is sensitive to acute state, not just chronic training. The honest read: a 2K rowed after a bad week of sleep, in a hot room, with a partial warm-up, and at a damper setting the rower does not usually use will read as a worse score than the rower's physiology warrants. The rower who controls the controllable variables — sleep, hydration, warm-up, damper — is testing physiology; the rower who does not is testing variability.

The practical read: the rower who wants the 2K to be a reliable anchor should row it under the same conditions each time — same warm-up, same damper, same hydration protocol, same time of day. The AI coach that sees a 2K result that is several seconds outside the rower's moving average should ask what changed: was the warm-up different, was the sleep different, was the damper different? The answer shapes whether the score is a calibration update or a state outlier ([1] Schabort et al. 1999, Level 2b; [19] Concept2 — Damper setting blog, Level 5; [20] Concept2 — Warm-up blog, Level 5; [26] Ebert et al. 2007, Level 2b; [25] Turnes et al. 2018, Level 2b).

What the AI coach actually does with your 2K

For an AI coach that reads the rower's Logbook and writes the rower's session, the 2K is the most useful single anchor for reference pace — but it is one anchor among several. The coach's rule is:

• A new 2K with high confidence (familiarised pacing pattern, controlled conditions, recent enough to reflect current state) — the reference pace updates, the confidence tightens, and the next session is chosen from current state.
• A new 2K with low confidence (first 2K after a long layoff, unfamiliar damper, partial warm-up, bad sleep the night before) — the reference pace holds, the coach widens the band, and the session leans on rate caps and effort language rather than a hard split.
• A rower whose 2K and 5K are roughly equal — the coach trusts the aerobic base and prescribes polarized training, with most sessions easy and a few hard.
• A rower whose 2K is much faster than the 5K — the coach prescribes more aerobic volume, with the 2K held steady and the 5K expected to close the gap over weeks.
• A rower whose 5K is much faster than the 2K (rare) — the coach investigates, because this is unusual; the likely explanations are a lactate-tolerance ceiling, a pacing failure, or a 2K rowed under bad conditions.
• A rower with no recent 2K — the coach widens the reference-pace band and leans on rate caps, recent training data, and the rower's stated context.

The four-numbers dashboard (CTL, ATL, TSB, ACWR) is the how much the coach is asking the body to absorb in a given week. The 2K trajectory is the where the coach expects the body to be in a few months. The reference pace is the today. The session is the synthesis ([1] Schabort et al. 1999, Level 2b; [2] Ingham et al. 2002, Level 2b; [10] Seiler & Kjerland 2006, Level 4; [11] Sylta et al. 2014, Level 4; [12] Rosenblat et al. 2025, Level 2a).

Limitations and open questions

The 2K is one number from one test. Even in trained rowers with controlled conditions, the day-to-day variation is roughly two percent of mean power — a few seconds on the time ([1] Schabort et al. 1999, Level 2b; [6] Gee et al. 2013, Level 2b). The first 2K after a long layoff is partially a familiarisation piece, not a calibration of physiology, and the AI coach that treats it as a calibration will mis-anchor the reference pace ([6] Gee et al. 2013, Level 2b).

The physiology behind the 2K is multi-system. The 2K is dominated by power at VO2max, but VO2max, lactate threshold, peak power, and stroke mechanics each contribute independently ([2] Ingham et al. 2002, Level 2b; [4] Riechman et al. 2002, Level 2b; [5] Cerasola et al. 2020, Level 2b; [7] Ingham et al. 2013, Level 2b). A reader who treats the 2K as a single physiological number — a "VO2max test" or a "lactate-threshold test" — misses what the literature actually says, and the AI coach that anchors on a single-system reading will misread the rower.

The sex-specific and age-specific responses are not as well characterised as the elite-male response. The 2003 [21] Yoshiga and Higuchi paper (Level 5) provided sex-specific data, but the underlying literature is smaller than the male literature, and the population-level percentiles rest on self-reported pieces with a mix of verification levels ([13] Concept2 — Rankings, Level 5).

The hydration, warm-up, and damper-setting variables are real but under-studied in the rowing-specific 2K literature. The 2007 [24] Ebert et al. paper (Level 5) is cycling, not rowing; the rowing-specific hydration-and-2K data is sparse. The manufacturer guidance ([19] Concept2 — Damper setting blog, Level 5; [20] Concept2 — Warm-up blog, Level 5) is practical but not peer-reviewed.

The percentile rankings are population-level and reflect self-reported effort. Even with verification levels, the bulk of the ranking data is self-reported, and the population comparison is meaningful at the decile level but noisy at the percentile level for any individual rower. The AI coach that uses a single 2K's percentile as a training target is over-reading the result; the coach that uses the percentile as a calibration check against the rower's trajectory is using it more honestly.

What to do with this article

Read the principle: the 2K is one number on one day under one set of conditions, and the literature treats it as a synthesis test of aerobic power, anaerobic power, lactate tolerance, pacing, and state. Read the evidence: the [2] Ingham et al. (Level 2b) physiological-determinants paper shows power at VO2max and peak power dominate; the [1] Schabort et al. (Level 5) reliability paper shows day-to-day variation is roughly two percent of mean power; the [6] Gee et al. (Level 5) pacing paper shows rowers settle into a conservative-start, end-spurt pattern after a habituation trial; the [10] Seiler & Kjerland (Level 5) and [12] Rosenblat et al. (Level 2a) polarized-training papers show the 80/20 model outperforms pyramidal and threshold distributions on VO2max and time-trial performance. Read the practical read: a single 2K is one data point, a moving average of the most recent two-to-three 2Ks is a calibration signal, the 2K/5K ratio tells the training-distribution story, and the AI coach that anchors reference pace on the synthesis — 2K, 5K, training log, stated context — is reading the rower correctly.

When you want to use your 2K as an anchor, the practical recipe is: row a familiarisation piece first if it has been more than six weeks since your last 2K, then row your calibration piece under matched conditions (same warm-up, same damper, same hydration, same time of day); trust the moving average of your most recent two-to-three 2Ks more than any single test; pair your 2K with a 5K every six-to-eight weeks to read the training-distribution story; and let the AI coach read the synthesis, not just the most recent number.