Training · Pro: every answer, checked

Mechanical tension. Reviews of the molecular literature converge on tension sensed by the muscle fiber, transduced partly through mTORC1-mediated increases in protein synthesis, as the prime driver. Metabolic stress and muscle damage, once framed as co-equal mechanisms, are now viewed as secondary or indirect at best.

Contested and downgraded. Metabolite accumulation and cell swelling lack causal evidence as direct growth signals. The mainstream reading is that metabolic fatigue works indirectly: it forces recruitment of high-threshold motor units, which then exposes more fibers to mechanical tension. The pump is a marker of the session, not a mechanism of growth.

That damage does not mediate or potentiate growth. Early cross-sectional area increases in new trainees are largely damage-induced swelling, and the elevated protein synthesis after early sessions is directed at repair, not accretion. Measurable hypertrophy only tracks protein synthesis once damage has attenuated over repeated sessions.

Because early post-exercise synthesis is spent on repairing and remodeling damaged tissue rather than adding contractile protein. Studies tracking integrated myofibrillar protein synthesis found it only relates to hypertrophy after damage attenuates, typically several weeks in. Early MPS is a repair bill, not a growth receipt.

That only the final repetitions before failure, roughly the last five, stimulate hypertrophy, because those are when motor unit recruitment is near-maximal and fiber shortening slows enough to generate high per-fiber tension. Earlier reps in a submaximal set are framed as non-stimulating.

Contested. It extrapolates from isometric, single-joint recruitment data; in multi-joint lifts, near-full recruitment can occur at lower loads and further from failure than the model assumes. Direct longitudinal tests are lacking, and critics note hypertrophy data fit a graded dose-response to proximity to failure better than a sharp effective-rep cutoff.

Motor units are recruited in order of size: small, low-threshold units serving slow fibers first, then progressively larger, higher-threshold units as force demand rises. Recruitment is governed by required force, not by movement speed or load per se. In most large limb muscles the recruitment pool is fully engaged below maximal force, with rate coding covering the rest.

As fatigue accumulates in a light set, maintaining force output requires recruiting progressively higher-threshold motor units. Near failure, recruitment is largely complete regardless of load, so the fibers that matter end up under tension either way. This is why effort, not load, is the gatekeeper for growth across a wide rep spectrum.

Motor unit discharge rates increase, recruitment thresholds drop, voluntary activation rises, and agonist-antagonist coordination improves. Corticospinal excitability increases within about four weeks. These changes lift force output before any measurable muscle thickness change, which typically lags several weeks behind, and early apparent size changes are confounded by damage swelling.

Neural adaptation. Early studies show maximal voluntary force can rise by double-digit percentages within the first month alongside increased corticospinal excitability, while true muscle growth takes several weeks longer to appear and early size changes are confounded by swelling. The early gain is the nervous system learning to drive the existing muscle harder and coordinate the lift better.

Muscles do not grow uniformly along their length. Training studies show different proximal versus distal growth within the same muscle, and EMG shows activation itself is regionally non-uniform. Exercise selection shifts the pattern: in one RCT, leg extensions grew all three rectus femoris regions while squats grew mainly the central vastus lateralis.

Favorable but still maturing. A meta-analysis of studies comparing exercise at longer versus shorter muscle lengths found greater hypertrophy at longer lengths, and lengthened-position work appears to preferentially grow distal regions. Effect sizes are small to moderate and the study pool is young, so the mainstream position is a probable advantage, not a settled law.

Roughly equivalent, possibly a small edge to lengthened partials. Meta-analytic comparisons of full versus partial ROM find trivial overall differences, with subgroup analyses hinting at a benefit when partials are done at long lengths. A 2025 within-participant trial in trained lifters found similar adaptations between the two. Neither approach is clearly wrong.

An hour of daily static stretching at high discomfort, applied via an orthosis for six weeks, produced meaningful gastrocnemius thickness increases, around 5-15%. It is proof of concept that stretch alone can grow human muscle. The dose was extreme, the finding is so far specific to the calf, and shorter conventional stretching does not reproduce it.

Preliminary. The rationale is adding passive tension time without lengthening the session, and small trials report modest or mixed effects, with any early advantage tending to shrink over the study period. Longitudinal inter-set stretching data remain sparse. It is a plausible, low-cost addition, not an evidenced requirement.

Rising with diminishing returns. The 2025 Pelland meta-regression across 67 studies found hypertrophy keeps increasing as weekly set volume increases, with no clear plateau inside studied ranges, but each added set buys less than the one before. There is no single magic set count; there is a curve that flattens.

Both rise with volume, but strength's diminishing returns are considerably steeper. In the same meta-regression dataset, hypertrophy keeps buying measurable size from added sets far longer than strength does; strength gains flatten at comparatively low weekly volumes. Chasing strength with hypertrophy-level volumes mostly adds fatigue, not force.

Little to nothing. Schoenfeld's 2019 meta-analysis of 25 studies found no significant hypertrophy differences across frequencies once volume was matched, in trained and untrained lifters, upper and lower body. The 2025 dose-response meta-regression reached the same conclusion: frequency's marginal effect on size is compatible with zero.

Distributing volume into higher-quality sets. Splitting 20 weekly sets across more sessions keeps per-session fatigue lower, so late sets keep their rep quality and load. Meta-analytic data also hint higher frequency modestly helps strength, consistent with more frequent skill practice on the lifts. Frequency serves volume and skill; it is not its own stimulus.

No difference in whole-muscle hypertrophy, but clearly greater 1RM strength gains with heavy loads. Across 21 studies with sets taken to failure, size responded to effort across the load spectrum while maximal strength tracked the load lifted in training. Load selection is a strength decision more than a size decision.

Specificity. Testing a 1RM is partly a skill: expressing force in a specific coordination pattern under maximal load. Heavy training rehearses that exact task, driving neural adaptations, high-threshold recruitment at low fatigue, and confidence under load, that light training does not practice. Same muscle, better software for the test.

Modeling reps in reserve as a continuous variable across 55 hypertrophy studies, growth increased as sets moved closer to failure, with no clear threshold or binary switch at failure itself identified. Closer helps; touching absolute failure adds little beyond its fatigue cost.

Barely. Across 67 strength studies, differences in proximity to failure produced negligible differences in 1RM outcomes. Strength appears to be driven by lifting heavy and practicing the movement, and grinding sets to failure adds fatigue without adding measurable strength. This is a clean divergence between the two adaptations.

The 20% group did about 40% fewer reps yet gained similar 1RM strength, improved countermovement jump more (about 9.5% vs 3.5%), and preserved their type IIX fiber percentage. The 40% group, doing far more fatiguing reps, got more quadriceps hypertrophy but shifted away from the fastest myosin phenotype.

It depends on the goal, with one surprise: strength gains were similar across thresholds. The meta-analytic finding is that velocity-loss cutoffs did not influence 1RM strength or muscular endurance gains, while lower thresholds better preserved jump, sprint, and bar-velocity qualities and higher thresholds produced greater hypertrophy. The threshold is a dial between speed preservation and growth volume.

A small advantage for periodized training on maximal strength. That is the headline finding of the comparative meta-analyses. The result is heavily qualified, though: the comparison groups were almost always constant, unvaried programs, and the follow-ups were short, so the effect may reflect variation itself rather than planned periodization.

That the literature never tested periodization's actual claim. Across the meta-analyses' included trials, none compared periodized programs against varied but non-periodized programs, and none tested whether the planned timing of adaptations occurred as predicted. So the data cannot separate the benefit of pre-planned structure from the benefit of variation itself.

No meaningful difference. For hypertrophy, Grgic's volume-equated meta-analysis found an effect size near zero between the two models. For strength, Harries and colleagues likewise found no significant difference. The choice between undulating and linear structure appears to be preference and logistics, not physiology.

With everything else identical over 8 weeks, the 3-minute group gained more 1RM strength in squat and bench and showed greater muscle thickness gains, significant in the anterior thigh with trends elsewhere. The short-rest group still progressed, only measurably less. Longer rest won on both outcomes in trained men.

Volume maintenance. Longer rests restore phosphocreatine and clear fatigue, so subsequent sets keep their load and rep count, preserving the total mechanical work that drives growth. Short rests compress the session but bleed performance from every set after the first. The old hormonal rationale for short rests does not survive the outcome data.

Essentially none, per current evidence. Large cohort work by West, Phillips, and Morton found no correlation between exercise-induced hormone elevations and gains in muscle or strength, and protocols engineered to spike hormones grew no more muscle. Local factors like androgen receptor content track outcomes better than the systemic spike.

Growth is similar across repetition durations from about 0.5 to 8 seconds when other variables are equated, while deliberate super-slow training around 10 seconds per rep is inferior. Time under tension per se is not the stimulus; sufficient tension across enough hard reps is. Tempo remains useful for control and skill, not as a growth multiplier.

That endurance exercise activates AMPK, an energy-stress sensor, which suppresses mTORC1 and therefore blunts the protein synthesis response to lifting. Early rodent and pharmacological work supported it: chemically activating AMPK blocked mTOR signaling after resistance exercise. It remains the standard mechanistic story taught for concurrent training.

Not convincingly. In trained humans, moderate cycling before lifting does not suppress mTOR signaling, and studies finding differential AMPK and mTOR activity see no antagonistic relationship with adaptation. Growing consensus holds that whatever interference exists is not mediated by this axis; candidate culprits are residual fatigue, energy deficit, and glycogen depletion.

Across 43 studies, concurrent training cost essentially nothing for maximal strength or whole-muscle hypertrophy versus lifting alone when frequency and volume were matched. The one reliable casualty was explosive power, especially with same-session stacking. Modern pooled data shrink the interference effect to a power-specific, dose-dependent concern.

High frequencies and high intensities of endurance work, same-session stacking, and running more than cycling, plausibly via impact-related damage and fatigue. Acute signaling mirrors this: mTOR suppression appears after maximal sprint work but not after 45 minutes of moderate cycling. Interference is a dose and modality problem, not a property of aerobic exercise itself.

Because a snapshot of a signaling cascade is not the adaptation. Signaling correlations are inconsistent across studies and time points: some report post-exercise p70S6K relating to growth while others find nothing, and acute myofibrillar protein synthesis after an unaccustomed session did not correlate with training-induced hypertrophy. Chronic outcomes integrate recovery, volume tolerance, and repeated exposures that one acute biopsy cannot capture.

Shifts within the fast-fiber pool: training of nearly any hard type moves IIX toward IIA, and hybrid fibers resolve toward pure types. Detraining and low-fatigue training bias the pool back toward IIX. These are real, measurable myosin heavy chain changes, seen for example in the velocity-loss trials where high-fatigue training reduced IIX percentage.

Limited and contested. Single-fiber analyses show most adaptation happens along a continuum of hybrids (I/IIa, IIa/IIx) rather than as a binary switch, and clear, replicated I to II conversion from voluntary training is scarce; some sprint data suggest bidirectional movement toward IIA. The conservative mainstream position is that the slow-fast boundary is largely resistant to training.

Ask for longitudinal outcome data. Acute signaling and even acute protein synthesis responses have repeatedly failed to predict training-induced hypertrophy, partly because early responses fund repair and because single time-point biopsies miss the integrated response. A protocol earns superiority claims through measured growth over weeks, not through one cascade lighting up.

Central oxygen delivery, dominated by maximal cardiac output. Most reviews place the large majority of the limitation, on the order of 70 to 75%, in cardiovascular oxygen transport rather than in the muscle's capacity to extract oxygen. In elite endurance athletes the delivery side stays the constraint, and some of them add a pulmonary limitation that untrained people do not have.

Isolated small-muscle-mass exercise. When one small muscle group is exercised alone and overperfused, as in the single-leg knee extension work of Andersen and Saltin, its oxygen consumption per kilogram of muscle runs two to three times what the same muscle reaches during whole-body maximal exercise. The muscle's oxidative capacity is not the ceiling when delivery is not the constraint.

Maximal cardiac output, driven mainly by stroke volume, in most accounts. The size of the split is contested: the original Dallas Bed Rest data credited increased cardiac output and a widened arteriovenous oxygen difference roughly equally, and a later re-analysis put even more weight on muscle oxygen diffusion. Maximum heart rate does not rise with training and drifts down slightly.

No. Reported plateau incidence swings from under 10% to 100% of subjects depending on the plateau criterion, the sampling and averaging interval, and the exercise modality. One study found 8% on a cycle and 58% on a treadmill. Secondary criteria such as RER cutoffs and age-predicted heart rate were chosen largely by convention, and a supramaximal verification bout is increasingly recommended as confirmation.

LT1 is the first sustained rise in blood lactate above resting baseline during an incremental test, the top of the fully aerobic domain and a close functional match to the gas exchange threshold. It is a lactate-appearance marker, not a steady-state limit, and its position depends on the detection method used.

LT2 is estimated from an incremental test as the intensity above which lactate rises without stabilizing. MLSS is measured directly with repeated constant-load trials of about 30 minutes on separate days, taking the highest load where blood lactate rises no more than about 1 mmol per litre over the final 20 minutes. LT2 is an estimate of the boundary; MLSS is a direct measurement of it.

FTP is a field construct for sustainable one-hour power, not a metabolic boundary. The 20-minute test with a 5% deduction commonly overestimates MLSS in trained cyclists, and changes in FTP after training do not reliably track changes in MLSS. The size of the bias varies with population and with the comparison measure used.

Critical power is the asymptote of the hyperbolic power-duration relationship, the highest power at which a metabolic steady state is still attainable. Above it, VO2, lactate, and phosphocreatine drift until exhaustion. It is derived from a set of maximal efforts of different durations, not from a single test.

W' is the fixed quantity of work available above critical power, expressed in joules. It behaves like a finite tank that drains above CP and recharges below it, and exhaustion coincides with its depletion. The classical reading attributes it to phosphocreatine, stored oxygen, and glycolytic capacity; current work ties it more closely to accumulating fatigue-related metabolites.

At both ends of the duration range, and in the same direction. Below roughly 2 minutes it overpredicts, because the two-parameter model has no ceiling on instantaneous power and lets predicted power run toward infinity as duration approaches zero. Beyond roughly 15 to 20 minutes it overpredicts sustainable power, since it ignores glycogen depletion, thermal load, and central fatigue.

Critical power, or critical speed. In the heavy domain, above LT1 but below CP, VO2 and lactate rise then stabilize at a delayed, elevated steady state. Above CP in the severe domain nothing stabilizes: VO2 climbs to VO2max, W' drains, and exhaustion arrives on a predictable schedule.

It becomes evident around 2 to 3 minutes into constant-load work, superimposed on the primary rise, and pushes VO2 above what the work rate should cost. In the heavy domain it stabilizes after roughly 10 to 15 minutes at a delayed steady state below VO2max. Efficiency is lost but a steady state is still reached.

The slow component never stabilizes above critical power. VO2 keeps climbing until it reaches VO2max, and W' depletion coincides with that point. Across subjects, W' and the size of the slow component track each other closely, with about three quarters of the variance shared in the study that tested it directly.

Around 86% of it originates in the working limbs, which rules out ventilation and cardiac work as primary drivers. The cause within the muscle is contested: progressive recruitment of less efficient higher-threshold motor units is the classic account, but a slow component appears in conditions where extra recruitment is unlikely, and recent evidence weights a rising oxygen cost inside already-active fibers.

Not in every model. Mice with muscle-specific PGC-1alpha knockout have shown intact mitochondrial biogenesis and normal exercise capacity after voluntary wheel running. Other knockout work reports blunted mitochondrial and capillary adaptation, and the discrepancy tracks the model and the training protocol. PGC-1alpha coordinates the transcriptional program, but calling it the indispensable master switch overstates the evidence.

Less consistent than textbook diagrams suggest. AMP and ADP concentrations rise with intensity and duration, but measured AMPK activation varies between studies, with some supporting an intensity and duration effect and others not. CaMKII findings across intensities and durations are similarly inconsistent.

The limiting step differs. Human whole-body exercise is limited mainly by oxygen transport to muscle, and human muscle carries mitochondrial capacity in excess of what delivery can supply. Small quadrupeds sit closer to a mitochondrial limit, so a manipulation that raises rodent mitochondrial capacity can raise rodent performance without moving a human's ceiling.

Training reliably raises the capillary-to-fiber ratio, with a pooled increase of 0.33 across 47 trials and moderate heterogeneity. Gains cluster in the first weeks and appear mainly in untrained to moderately trained people. Both continuous moderate work and interval work drive angiogenesis, while sprint interval training has not reliably increased capillary density.

The endurance half holds up better than the strength half. Endurance athletes do show eccentric remodeling with a larger left ventricular cavity and higher LV mass. The prediction that resistance training produces concentric hypertrophy is what prospective training studies have failed to reproduce, and the original evidence was cross-sectional with limited echo precision.

Contested. Untrained subjects typically plateau early, around 120 bpm or roughly 40 to 50% of VO2max. Several studies report that trained endurance athletes keep increasing stroke volume toward maximal heart rate, with augmented diastolic filling as the proposed mechanism. Other studies report plateaus or declines, and measurement method explains part of the disagreement.

Fast. Expansion begins within hours of a session and typical training-induced increases land around 9 to 25%, roughly 300 to 700 ml. Increases of 12 to 20% have been reported after only 4 to 8 days of intense training. This is the earliest cardiovascular adaptation and it contributes to the early drop in exercise heart rate.

Almost entirely plasma. Nearly all of the blood volume increase in the first 14 days comes from plasma expansion, with red cell volume essentially unchanged. The rise in erythrocyte volume lags the plasma response by roughly 2 to 3 weeks, after which further gains distribute more evenly.

Group mean increases of roughly 3 to 5% after camps of about 3 to 4 weeks at moderate altitude, whether natural or simulated. One 21-day normobaric protocol reported 3.8%. Group means hide wide individual spread, and whether hypobaric and normobaric hypoxia produce equivalent responses is still argued.

The evidence says no. Hemoglobin mass responses do not repeat consistently from camp to camp in the same athlete. In one elite cohort followed across repeated camps, 27% responded positively every time, 13% negatively every time, and 60% went both ways. Hypoxic dose, meaning altitude reached multiplied by hours of exposure, and iron availability are the clearest modifiable determinants.

Stiffer triceps surae tendon-aponeurosis tissue is associated with better economy. In a 14-week loading study a roughly 16% increase in tendon-aponeurosis stiffness accompanied a roughly 4% reduction in the oxygen cost of running at a given speed, which the authors linked to greater energy storage and return plus a redistribution of muscular output within the lower limb.

Plasma volume goes first, falling within about a week and taking stroke volume with it. Mitochondrial enzyme activity such as citrate synthase declines with a half-life near 12 days and settles well above sedentary values. Capillarization is the most stubborn, showing no decline across 12 weeks of cessation in the classic detraining series.

The maximal rate of lactate accumulation, expressed in mmol per litre per second and estimated from a 10 to 15 second all-out sprint. Modeling platforms use it as the glycolytic counterweight to VO2max, combining the two to predict threshold and fat oxidation. The measured quantity is blood lactate accumulation, not glycolytic flux itself.

Glycolytic flux is unlikely to be maximal in the test and peak flux is brief, there is no criterion for confirming VLamax was reached, there is no correction for lactate clearance between the sprint and the blood draw, and the alactic time correction is error-prone. Reliability also varies across exercise modalities, so it does not hold up as a standalone diagnostic.

A small advantage for polarized distribution on VO2peak, a standardized mean difference of 0.24 across 11 studies and 284 participants, rated high certainty. The effect was clearer in shorter interventions and in highly trained athletes. Seventeen studies and 437 subjects were pooled in total.

Not reliably. Polarized and other distributions produce similar time-trial outcomes in the pooled data, and head-to-head trials in well-trained runners and national-level rowers found pyramidal distribution was not inferior. The VO2peak signal and the performance signal do not match.

Intensity relative to the VO2max work rate. HIIT is repeated near-maximal work below or up to the power or velocity at VO2max, commonly run in intervals of about 1 to 5 minutes. SIT is supramaximal all-out work above that work rate, usually 30 seconds or less. SIT loads glycolysis and the neuromuscular system harder; HIIT accumulates more time near VO2max.

The recent evidence favours HIIT. A meta-analysis of randomized crossover trials found HIIT outperformed SIT for both VO2max and maximal aerobic power or velocity, and a controlled trial in well-trained men found 4 x 4 minute HIIT raised VO2max more than 8 x 20 second or 10 x 30 second sprint protocols. Longer work intervals and longer programs showed the clearest effect.

Resistance to the deterioration of physiological variables and performance during and after prolonged exercise, covering both when the decline starts and how large it gets. It is measured as the shift in thresholds, economy, or power after several hours of work rather than in the fresh state, and it has been proposed as a fourth performance parameter alongside VO2max, threshold, and economy.

There is no consensus protocol. The fatiguing bout defines the answer, since greater relative intensity and longer duration produce more deterioration, so durability scores from different protocols are not comparable. Nutrition, environment, and starting glycogen have to be controlled for the number to repeat.

Plasma volume expands, exercise heart rate at a fixed work rate falls, core and skin temperature at that work rate drop, and sweat rate rises with earlier sweat onset. Sodium conservation in sweat develops more slowly. Meaningful heat tolerance can appear within a few days, and most adaptations are largely complete inside two weeks.

Meta-analytic estimates put the loss at roughly 2.5% of the adaptation per day without heat, about 2.3% for end-exercise heart rate and about 2.6% for end-exercise core temperature, which returns heat tolerance to baseline somewhere around 5 to 7 weeks. Decay is not uniform across variables: core temperature adaptation was only around 5% decayed at 12 days in one protocol while heart rate adaptation had already lost far more.

The endurance evidence is mixed and much weaker than the resistance-training case. Reviews describe the effects of regular post-exercise cold water immersion on aerobic capacity, lactate threshold, power output, and time-trial performance as equivocal, with time-trial performance and maximal aerobic power left uncompromised in the available trials. Markers of angiogenesis and mitochondrial biogenesis have also gone both ways.

Signaling responds, performance does not follow reliably. Low glycogen availability augments post-exercise AMPK and p53 signaling, but several studies showing that signaling difference found no matching advantage in VO2max or time-trial performance. A 3-week home-based sleep-low protocol did improve FTP in trained cyclists, in a small feasibility study.

It escalates. Low-threshold type I units are recruited first per the size principle and deplete their glycogen fastest at submaximal work rates. As those fibers become glycogen-limited, higher-threshold type IIa units are progressively recruited to hold the same power.

Central adaptations transfer better than peripheral and neuromuscular ones, and specificity matters more the more trained the athlete. A systematic review and meta-analysis of running and cycling used as substitutes for one another found no clear improvement or decrement in sport-specific VO2max or running performance over short to moderate periods, with wide confidence intervals and few studies.

You cannot hide from someone whether they are lifting heavy or light, or training at all. Participants know their group, so expectancy and effort differences contaminate outcomes. Meta-epidemiological work shows inadequately blinded trials report inflated effects on average, with the worst distortion on subjective, self-reported outcomes.

The 'optimized' group can outperform on strength and perceived recovery despite identical training. Expectancy is an active ingredient in exercise interventions. Belief changes effort, arousal, and self-report, which is why open comparisons of training methods partly measure marketing, not physiology.

Most training studies run roughly 8-20 participants per group, often convenience samples from one university or team. A meta-review of 266 musculoskeletal randomized trials found a median statistical power of 42%, and rehabilitation trial reviews report similar or worse. Small true differences between programs are usually undetectable.

In an underpowered study, only samples that got lucky with a large observed effect cross the significance threshold. If significant results are what gets published, the literature fills with inflated estimates. This is the winner's curse: the smaller the study, the more the published effect exceeds the true one.

No. A 2022 Sports Medicine review concluded acutely measured surface EMG amplitude is not a validated predictor of hypertrophy. Amplitude is altered by fatigue, electrode position, subcutaneous fat, and signal cancellation, independent of motor unit recruitment. Training studies show similar growth across loads despite different EMG amplitudes.

Essentially not at all. MPS rose over 200% after the first session, and quadriceps volume grew about 8% over training, but the correlation between acute MPS and eventual growth was near zero (r = 0.02 to 0.16). An acute anabolic response to novel exercise is not a growth forecast.

It substitutes a surrogate for the outcome. Activation, acute hormone spikes, lactate, soreness, and pump are all acute signals that correlate poorly with measured long-term growth. The only defensible evidence for a hypertrophy claim is a longitudinal study measuring muscle size change directly.

Ultrasound thickness correlates well with MRI cross-sectional area changes (around r = 0.7 in validation work) but poorly with whole-muscle volume (r near 0.33). It is cheap and sensitive enough to detect training-induced growth, but it samples one site and is operator dependent: transducer pressure alone can shift readings.

DXA lean mass correlates strongly with MRI at a single time point, but for measuring change after training the agreement drops to moderate (around r = 0.5 in comparison work). Its error also moves with hydration and glycogen. A few weeks of creatine or a carb load can shift 'lean mass' without any new contractile tissue.

A 2020 systematic review found test-retest ICCs with a median of 0.97 and a median CV of 4.2%, across trained and untrained groups. Counterintuitively, it also found reliability was similarly high with or without familiarization sessions. The 1RM is one of the more dependable tests in the field.

Novices improve 1RM partly by learning the test: coordination, bracing, confidence with maximal loads. This neural and skill component inflates strength change in short studies, especially when the training resembles the test. A program can look superior on 1RM change while producing no more muscle or force capacity than the comparison.

Peak force in the isometric mid-thigh pull is highly reliable (ICCs commonly 0.9 or higher). Early-phase measures like force at 50-100 ms and rate of force development are much noisier, with CVs reported up to 20% or more. Peak force is a solid monitoring number; early RFD needs many trials and strict protocol control to be usable.

It superimposes electrical stimulation on a maximal voluntary contraction; any extra force evoked means voluntary activation was incomplete. It shows most people cannot fully activate on every attempt, which is useful for separating neural from muscular limits. But activation estimates vary between trials, the force-activation relationship is nonlinear near maximum, and results depend on stimulation setup.

Flight-time devices (contact mats, most apps) compute height from time in the air and assume you land in the takeoff posture. Tuck the knees or plantarflex at landing and flight time stretches, inflating the score. Force plates using the impulse-momentum method compute takeoff velocity directly and avoid that assumption. Same jump, different physics, different number.

Surprisingly valid. A 2023 meta-analytic review of My Jump validation studies reported high agreement and reliability (ICCs above 0.8, many above 0.9) against force plates, contact mats, and photocell systems, with near-perfect correlations in several individual validation studies. The app reads flight time from slow-motion video, so it shares flight-time assumptions, but as a field tool it holds up.

Group-level correlations are strong, roughly r = 0.84 to 0.92 in validation studies, with standard errors of estimate around 4-5 ml/kg/min in adult validation work. That is good enough for ranking a squad, but prediction equations are population specific: equations built on schoolchildren or general adults misestimate trained athletes, and shuttle tests reward change-of-direction ability that a treadmill test never sees.

Probably indistinguishable from noise plus test learning. With prediction error of several ml/kg/min and improved pacing and turning economy on retest, a 2-point rise is within the uncertainty band. The performance itself (levels completed) is the more trustworthy signal, because it is a direct measurement, not an estimate.

Yes. Ramp slope and stage length change test duration, and protocols that end too fast or drag too long can yield lower peak values; the common recommendation is a protocol reaching exhaustion in about 8-12 minutes. Comparing your VO2max across labs or years means little unless the protocol and modality match.

Plateau incidence during ramp tests is low, and whether one appears depends heavily on data processing. In one analysis, plateau detection ranged from 100% of subjects with breath-by-breath or 15 s averaging down to 8% with 1-minute averaging. The same test can show a plateau or not depending on the smoothing window.

A constant-load effort performed after the incremental test, commonly prescribed at around 105-110% of the peak power reached. If it reproduces the same VO2 peak, the incremental value looks like a true maximum. Recent work complicates the practice: supramaximal bouts often end too soon to confirm anything and can undershoot the ramp value, while intensities of 85-95% of peak power confirm more often, and several studies in trained adults find verification adds little over the ramp alone.

The gold-standard protocol needs several constant-load trials of at least 30 minutes on separate days, hunting for the highest intensity where blood lactate rises no more than about 1 mmol/L between minutes 10 and 30. That is three to five lab visits with repeated blood sampling. Every practical threshold test (FTP, lactate ramps, critical power) exists to approximate this without the burden.

Contested. Some studies in trained cyclists report near-perfect correlation with MLSS and small mean bias; others find FTP95 sits above MLSS, with MLSS averaging roughly 89-93% of the 20-minute-derived value, and training-induced MLSS changes not tracked by FTP. Group correlation hides individual disagreement of many watts either way.

A short recording (around 1-5 minutes after stabilization) taken on waking, in a consistent posture, before caffeine, food, or screens, using the natural log of rMSSD as the metric. The point of the ritual is not precision of one reading but holding every condition constant so day-to-day change reflects the nervous system, not the routine.

HRV differs between supine, seated, and standing because autonomic balance shifts with posture, and respiratory sinus arrhythmia couples HRV to breathing rate and depth. Slow breathing inflates variability. rMSSD is less breathing-sensitive than frequency-domain measures, but switching position or breathing style between days still fabricates trends.

No. Devices differ in sensor (ECG strap vs optical PPG), recording window, artifact correction, timing (spot morning reading vs nocturnal average), and some rescale rMSSD onto proprietary 0-100 scores. Nocturnal and morning rMSSD track each other reasonably well in studies, but absolute values across platforms are not comparable. Trend within one device is the only safe use.

Ultra-short recordings of about 60 seconds after a stabilization period can agree acceptably with standard 5-minute criteria for rMSSD, and stabilization itself can be shortened to around 60 seconds in endurance athletes per validation work. Cut the stabilization and the reading is contaminated by the transition to lying down.

Homogeneous convenience samples. Recruiting one team or one training class truncates the spread of ability, shrinking the standard deviation that standardized effects divide by. A modest raw improvement over a tiny SD produces a huge Cohen's d, sometimes physiologically implausible. Effect sizes from narrow samples do not transfer to broader populations.

Studies finding significant improvements or differences get written up and published; null comparisons stall in drawers. Combined with small samples, the surviving literature overstates how much protocol details matter. Funnel plot asymmetry and implausibly high positive-result rates in some S&C topics are the fingerprints.

That observed variation in training outcomes is not the same as true individual differences in trainability. Measurement error, random within-subject biological variability, and regression to the mean all masquerade as response heterogeneity. Without a control comparison, labeling someone a non-responder from a single pre-post change is statistically unjustified.

Replicate designs: give the same person the same intervention more than once, as in replicate crossovers or repeated training blocks separated by washouts. If someone responds well both times, that is evidence of true individual trainability; if response does not repeat, the first result was likely noise. A 2025 repeated resistance training study found strength and size responses do show within-person reproducibility.

MDC is the smallest change exceeding what measurement noise alone could produce with a chosen confidence, commonly computed as SEM x 1.96 x the square root of 2 for 95% confidence. If a test's MDC for 1RM is 5 kg, a 3 kg improvement cannot be distinguished from retest noise, whatever the athlete or coach feels about it.

No detectable change. A fluctuation inside the error band is expected from noise on a flat true score, so it cannot carry a fatigue interpretation on its own. Defensible options: wait for the trend across several tests, tighten the protocol to shrink the error, or corroborate with independent signals before acting.

That classical periodization rests on an assumption of predictability that biology does not honor. Fixed long-range plans presume you can forecast how an athlete will adapt weeks ahead, while actual responses to training are variable and context dependent. Kiely argued the tradition is culturally inherited rather than evidence-led, and that planning should be responsive to the athlete's emerging state.

Many comparisons confound periodization with other differences: the periodized arm often trains with varied loads while the control repeats one scheme, so variation and load progression are entangled with 'periodization' itself. Studies are short, samples small and often untrained, and different periodization models rarely separate from each other. The label gets credit that the design cannot assign.

Transfer is the degree to which improvement in a trained exercise carries over to a target task. It depends on how closely the exercise matches the sport action in movement pattern, velocity, force direction, and coordination demands. A bigger squat loads the same muscles a sprint uses, but sprinting is a high-velocity, unilateral, elastic action; the overlap is partial, so the carryover is partial.

Transfer shrinks as athletes advance. In novices, almost any strength gain improves sprinting and jumping because general force capacity is the limiter. In trained athletes the general qualities are closer to sufficient, so further gym gains buy less sport improvement, and specific, velocity-matched work carries more of the load. Research on well-trained athletes shows progressively weaker correlations between added strength and sport outcomes.

Study conditions rarely resemble athlete life. Interventions run 6-12 weeks with supervised, fully standardized sessions, fixed diets or none tracked, and subjects who are often untrained students. Real training runs years, unsupervised, alongside sport practice, stress, and self-selected nutrition. Effects measured in the clean box may shrink, vanish, or reverse in the wild.

Chest-strap accelerometer estimates of vertical oscillation track video-based analysis reasonably and repeat reliably, but they can differ from true center-of-mass displacement because the sensor rides on the torso, not at the body's center of mass. The number is fine for tracking your own trend and useless for comparing against another runner's absolute value or a universal 'ideal' range.

Device dependent and imperfect. Validation of the Garmin Running Dynamics Pod found ground contact time differed significantly from force plates at all speeds, with bias reaching roughly 80 ms at fast step rates; chest-strap systems fare better but still deviate. Since real contact times span about 200-300 ms, errors of that size are a large fraction of the signal.

Novices adapt to almost any stimulus, which compresses differences between protocols: sophisticated and crude programs look similar over 8-12 weeks. Their responses are also amplified by learning effects. Findings from untrained cohorts describe the start of the adaptation curve, where an experienced athlete no longer lives, so 'no difference between programs' in novices says little about what separates programs at high training age.