Place · Level 3
Recovery science
训练只是给身体下指令 · 真正长出来的部分发生在恢复里 — 超量恢复 · DOMS · 睡眠 · 减量 · 冷热
Story path
- 1SupercompensationSupercompensation
- 2DOMS is not lactic acidDOMS is not lactic acid
- 3Deload weeksDeload weeks
- 4Active vs passive recoveryActive vs passive recovery
- 5Sleep — the master switchSleep — the master switch
- 6Hot & cold — ice baths and saunaHot & cold — ice baths and sauna
- 7Training hydrationTraining hydration
Chapter 1
Supercompensation
Supercompensation
Training itself doesn't make you stronger — it only issues an order: 'this area needs to be stronger.' The part that actually grows happens during recovery after training, and this curve is the spine of recovery science.
Hans Selye's 1956 'General Adaptation Syndrome' proposed that any stressor (training, cold, psychological pressure) triggers a three-phase response: alarm, resistance, exhaustion. Training science refines it into a four-phase supercompensation curve:
During training to +24h: training plus fatigue accumulation, performance temporarily drops+24-72h: the trough, repair starts, protein synthesis rises, performance still depressed+72-120h: the supercompensation peak, performance above baseline — this is the adaptation windowBeyond 120h with no new stimulus: adaptation gradually fades, the gains disappear over 3-7 days
The Issurin 2010 review notes that different systems have different recovery windows: the neural system 24-48 hours, muscle fibers 48-72 hours, connective tissue (tendons) 5-7 days, central nervous fatigue 5-10 days. This is why training a single muscle 2-3 times per week is optimal — a 48-72 hour interval lands right on the peak, stacking new stimulus on a higher baseline. Too low a frequency misses the peak; too high keeps you stuck in the trough.
Hans Selye's 1956 'General Adaptation Syndrome' proposed that any stressor (training, cold, psychological pressure) triggers a three-phase response: alarm, resistance, exhaustion. Training science refines it into a four-phase supercompensation curve:
During training to +24h: training plus fatigue accumulation, performance temporarily drops+24-72h: the trough, repair starts, protein synthesis rises, performance still depressed+72-120h: the supercompensation peak, performance above baseline — this is the adaptation windowBeyond 120h with no new stimulus: adaptation gradually fades, the gains disappear over 3-7 days
The Issurin 2010 review notes that different systems have different recovery windows: the neural system 24-48 hours, muscle fibers 48-72 hours, connective tissue (tendons) 5-7 days, central nervous fatigue 5-10 days. This is why training a single muscle 2-3 times per week is optimal — a 48-72 hour interval lands right on the peak, stacking new stimulus on a higher baseline. Too low a frequency misses the peak; too high keeps you stuck in the trough.
Where are you on the curve?
Without lab equipment, here's how an ordinary person can tell where they are on the curve:HRV (heart rate variability): tracked automatically by Garmin / Whoop / Apple Watch; your own baseline ±10% is normal; three straight low days means you're in the trough — deload that dayResting heart rate: 5 bpm above your baseline for three days, same signalSubjective recovery (1-10): 8-10 is the peak, train hard; 5-7 is mid-range, moderate training OK; below 5 is the trough, active recovery or restStrength performance: the same lift at the same weight dropping 5%+ means you're in the trough — don't push
This also helps distinguish 'functional overreaching' from 'pathological overtraining': the former is a short-term performance dip after 2-3 weeks of overload that rebounds with supercompensation after a week or two of reduced load — a planned training tool; the latter is 5+ weeks of escalating load with sustained performance decline and prolonged HRV / resting-HR deviation, requiring 1-3 months to recover — rare but serious, mostly in elites and anxiety-prone amateur runners.
Chapter 2
DOMS is not lactic acid
DOMS is not lactic acid
The soreness that arrives two or three days after training is called delayed-onset muscle soreness (DOMS). 'Post-training soreness = lactic acid buildup' is an error from early-20th-century textbooks; the truth has been established for 50 years but the slogan hasn't kept up — lactate is essentially fully cleared 30-60 minutes after training, so the timeline doesn't match DOMS's 24-72 hour window at all.
The Nosaka 2004 and Cheung 2003 reviews give the modern understanding — DOMS is a three-phase process:
Training to +24h, eccentric-induced microdamage: eccentric contractions (muscle bearing load while lengthening — descending in a squat, lowering a barbell) load individual sarcomeres unevenly, Z-discs develop microscopic tears, and damage markers (creatine kinase, myoglobin) are already detectable in plasma+24h to +72h, inflammation and neural sensitization: damage activates TLR4 / NLRP3 pathways; interleukin-6: A pro-inflammatory signal molecule (cytokine) released by immune cells during inflammation., tumor necrosis factor alpha: A strong pro-inflammatory signal molecule that runs high in chronic inflammation., and prostaglandins rise, and these inflammatory mediators sensitize the C-fiber pain neurons, lowering the pain threshold. So the sore, tight, aching feeling isn't the muscle itself crying out — it's a sensitized nervous system raising the alarm+72h to +5d, repair and adaptation: satellite cells activate, fuse into damaged fibers, protein synthesis is upregulated, and the repaired fiber is more robust than before — the starting point of training adaptation
So DOMS isn't a bad thing — it's an adaptation signal. But there's no need to chase DOMS for its own sake.
The Nosaka 2004 and Cheung 2003 reviews give the modern understanding — DOMS is a three-phase process:
Training to +24h, eccentric-induced microdamage: eccentric contractions (muscle bearing load while lengthening — descending in a squat, lowering a barbell) load individual sarcomeres unevenly, Z-discs develop microscopic tears, and damage markers (creatine kinase, myoglobin) are already detectable in plasma+24h to +72h, inflammation and neural sensitization: damage activates TLR4 / NLRP3 pathways; interleukin-6: A pro-inflammatory signal molecule (cytokine) released by immune cells during inflammation., tumor necrosis factor alpha: A strong pro-inflammatory signal molecule that runs high in chronic inflammation., and prostaglandins rise, and these inflammatory mediators sensitize the C-fiber pain neurons, lowering the pain threshold. So the sore, tight, aching feeling isn't the muscle itself crying out — it's a sensitized nervous system raising the alarm+72h to +5d, repair and adaptation: satellite cells activate, fuse into damaged fibers, protein synthesis is upregulated, and the repaired fiber is more robust than before — the starting point of training adaptation
So DOMS isn't a bad thing — it's an adaptation signal. But there's no need to chase DOMS for its own sake.
Do vs don't
Helpful (modest evidence): light activity or low-intensity aerobic raises blood flow and improves subjective comfort; foam rolling cuts DOMS intensity 5-10% short-term (Wiewelhove 2019); massage has a similar effect; continuing to train the sore muscle doesn't worsen DOMS and actually accelerates adaptation via the 'repeated bout effect'.Not helpful or counterproductive: ice baths give acute pain relief but chronic use blunts training adaptation (Roberts 2015: the ice-bath group gained 23% less strength over 12 weeks vs control); long-term high-dose ibuprofen suppresses muscle protein synthesis and satellite cell activation (Trappe 2002); stretching to relieve DOMS showed no significant effect across 12 Cochrane RCTs (Herbert 2011); 'lactate-detox massage' is marketing language.
A frequently misread heuristic: 'more DOMS means better training' is wrong. DOMS is driven mainly by movement novelty and eccentric ratio, not by volume or intensity. An experienced trainee piling on volume with familiar movements gets almost no DOMS while hypertrophy and strength keep climbing. Measure progress by strength, circumference, and state — not DOMS. One more important distinction: you can 'be sore but still train' (DOMS is local and improves after warm-up), but you can't 'be fatigued and still train' (true fatigue is systemic with a clear strength drop and elevated resting HR — that needs rest).
Chapter 3
Deload weeks
Deload weeks
A deload is a planned reduction week: after 4-8 weeks of accumulated training, you intentionally lower volume (sets or reps) or intensity (loads) to give the body a chance to clear fatigue.
Why planned rather than reactive:
Training fatigue accumulates latently — you feel 'fine' while HRV and strength may already be slowly droppingProactive deload is far better than reactive post-injury deload — two weeks of forced rest after an injury costs ten times more than a planned one-week deloadSupercompensation requires a trough; without the fatigue-clearing of a deload, the peak never arrives
Three deload styles all work: volume deload (same load, halve sets or weekly sessions); intensity deload (same sets and reps, drop loads or pace to 60-70%); mode change (switch the activity entirely — lifters go walk, cycle, or swim; runners go lift — with high psychological repair value). Beginners and intermediates do one intensity-style deload every 6-8 weeks, advanced lifters combine volume and intensity reduction every 4-6 weeks, and for race prep you taper 1-2 weeks before competition (large volume cut while preserving intensity).
Bell 2020 plus real coaching experience both note that a deload week's biggest value is actually the psychological reset. About 90% of the early signals of pathological overtraining are psychological and behavioral (irritability, dread of training, poor sleep, lost appetite), not strength numbers. So don't treat a deload as 'a lost week' — it's the price of being able to train for five years; and don't use a deload week to 'catch up on missed sessions' or 'test a 1RM', which is the ego talking, not a real deload.
Why planned rather than reactive:
Training fatigue accumulates latently — you feel 'fine' while HRV and strength may already be slowly droppingProactive deload is far better than reactive post-injury deload — two weeks of forced rest after an injury costs ten times more than a planned one-week deloadSupercompensation requires a trough; without the fatigue-clearing of a deload, the peak never arrives
Three deload styles all work: volume deload (same load, halve sets or weekly sessions); intensity deload (same sets and reps, drop loads or pace to 60-70%); mode change (switch the activity entirely — lifters go walk, cycle, or swim; runners go lift — with high psychological repair value). Beginners and intermediates do one intensity-style deload every 6-8 weeks, advanced lifters combine volume and intensity reduction every 4-6 weeks, and for race prep you taper 1-2 weeks before competition (large volume cut while preserving intensity).
Bell 2020 plus real coaching experience both note that a deload week's biggest value is actually the psychological reset. About 90% of the early signals of pathological overtraining are psychological and behavioral (irritability, dread of training, poor sleep, lost appetite), not strength numbers. So don't treat a deload as 'a lost week' — it's the price of being able to train for five years; and don't use a deload week to 'catch up on missed sessions' or 'test a 1RM', which is the ego talking, not a real deload.
Chapter 4
Active vs passive recovery
Active vs passive recovery
'Recovery' operates on two scales: the few minutes between sets, and the day or two between sessions.
Between sets, active or passive? Greenwood 2008 and Toubekis 2008 found that active recovery (slow walk, easy cycle, shake-out) clears lactate 30-50% faster than passive (sitting still), because muscle blood flow is maintained and the lactate shuttle keeps moving it into other tissues. The effect on next-set performance is small (3-5%) but more pronounced with short rest (60-90 second hypertrophy training). For long rest in strength training (3-5 min), passive is fine — lactate has cleared naturally. For HIIT intervals, active recovery is recommended — stopping fully makes heart rate harder to bring down.
Between sessions (24-48 hours), light activity beats complete bedrest: walking, easy cycling, yoga, or swimming lowers subjective pain 10-20% (Cheung 2003), improves next-day performance, and sustains motivation; complete bedrest slows circulation, worsens DOMS, and tends to dip mood.
But be clear about what counts as active recovery (HR below 60% max HR): walking, easy cycling you can chat through, slow swimming, yoga or tai chi, and light-load training far from failure all count. 'Easy' running (most recreational easy runs actually sit in the Zone 3 grey area), any interval work, and group sweat classes don't — that's another session. The practical rule: active recovery should leave you feeling better than when you started; if you're more tired afterward, it wasn't recovery. Still keep 1-2 days a week of truly no training.
Between sets, active or passive? Greenwood 2008 and Toubekis 2008 found that active recovery (slow walk, easy cycle, shake-out) clears lactate 30-50% faster than passive (sitting still), because muscle blood flow is maintained and the lactate shuttle keeps moving it into other tissues. The effect on next-set performance is small (3-5%) but more pronounced with short rest (60-90 second hypertrophy training). For long rest in strength training (3-5 min), passive is fine — lactate has cleared naturally. For HIIT intervals, active recovery is recommended — stopping fully makes heart rate harder to bring down.
Between sessions (24-48 hours), light activity beats complete bedrest: walking, easy cycling, yoga, or swimming lowers subjective pain 10-20% (Cheung 2003), improves next-day performance, and sustains motivation; complete bedrest slows circulation, worsens DOMS, and tends to dip mood.
But be clear about what counts as active recovery (HR below 60% max HR): walking, easy cycling you can chat through, slow swimming, yoga or tai chi, and light-load training far from failure all count. 'Easy' running (most recreational easy runs actually sit in the Zone 3 grey area), any interval work, and group sweat classes don't — that's another session. The practical rule: active recovery should leave you feeling better than when you started; if you're more tired afterward, it wasn't recovery. Still keep 1-2 days a week of truly no training.
Chapter 5
Sleep — the master switch
Sleep — the master switch
No matter how perfect the training plan, if sleep is 5-6 hours, adaptation is only 50-70% of what it should be. Sleep isn't 'extra' — it's a core variable of the training plan.
Reilly 1994's classic experiment showed that after 24-hour sleep deprivation, bench-press 1RM dropped 7% and reaction time slowed 15%. Mah 2011's Stanford basketball RCT showed the reverse: players who extended sleep to 10 hours/night for 5-7 weeks improved both 3-point and free-throw accuracy by 9%, with faster sprint times. The Knowles 2018 review summarizes the effects of chronic insufficient sleep (under 7 hours/night): blunted training adaptation, strength gain down ~30%, sleep deprivation directly suppressing mechanistic target of rapamycin: The cell's master 'grow / build' switch — turned on by enough protein and resistance training. (the master switch of protein synthesis), higher injury risk, appetite dysregulation (ghrelin up, leptin down), and lower immunity.
What the body does during sleep:
N3 deep sleep (15-20% of the night, mostly in the first half): about 70% of daily growth hormone (GH) release happens here, directly driving muscle repair and bone remodeling; most post-training muscle glycogen is also replenished in N3; and brain waste clearance (glymphatic) happens mostly hereREM sleep (20-25% of the night, mostly in the second half): new movements and coordination learned during the day's training are consolidated in REM (Walker 2017); insufficient REM raises chronic stress cortisol
So staying up late sacrifices GH and muscle repair, while waking early (4-5 AM) sacrifices REM and skill learning. Sleep on the night after training is especially critical — it determines your next session's state. The bottom line is 7-9 hours, prioritizing duration and consistency (same bedtime, same wake time); weekend 'catch-up sleep' has limited returns.
Reilly 1994's classic experiment showed that after 24-hour sleep deprivation, bench-press 1RM dropped 7% and reaction time slowed 15%. Mah 2011's Stanford basketball RCT showed the reverse: players who extended sleep to 10 hours/night for 5-7 weeks improved both 3-point and free-throw accuracy by 9%, with faster sprint times. The Knowles 2018 review summarizes the effects of chronic insufficient sleep (under 7 hours/night): blunted training adaptation, strength gain down ~30%, sleep deprivation directly suppressing mechanistic target of rapamycin: The cell's master 'grow / build' switch — turned on by enough protein and resistance training. (the master switch of protein synthesis), higher injury risk, appetite dysregulation (ghrelin up, leptin down), and lower immunity.
What the body does during sleep:
N3 deep sleep (15-20% of the night, mostly in the first half): about 70% of daily growth hormone (GH) release happens here, directly driving muscle repair and bone remodeling; most post-training muscle glycogen is also replenished in N3; and brain waste clearance (glymphatic) happens mostly hereREM sleep (20-25% of the night, mostly in the second half): new movements and coordination learned during the day's training are consolidated in REM (Walker 2017); insufficient REM raises chronic stress cortisol
So staying up late sacrifices GH and muscle repair, while waking early (4-5 AM) sacrifices REM and skill learning. Sleep on the night after training is especially critical — it determines your next session's state. The bottom line is 7-9 hours, prioritizing duration and consistency (same bedtime, same wake time); weekend 'catch-up sleep' has limited returns.
Chapter 6
Hot & cold — ice baths and sauna
Hot & cold — ice baths and sauna
Ice baths and sauna are social media's hottest 'recovery' interventions, but their evidence strength differs a lot — and their effects on training adaptation point in opposite directions.
The real trade-off of ice baths (10-15°C immersion for 10-15 min): acutely useful — subjective pain drops in the 1-2 hours after training, and recovery is sped subjectively in multi-day events (Bleakley 2012). But chronic use carries an adaptation cost: Roberts 2015 (J Physiol) RCT showed that over 12 weeks of resistance training, the group icing after every session gained 23% less strength and 18% less lean mass. The mechanism is that cold water suppresses post-training inflammatory signaling (nuclear factor kappa B: The cell's inflammation master switch (a transcription factor) — when flipped, it turns inflammation on., mechanistic target of rapamycin: The cell's master 'grow / build' switch — turned on by enough protein and resistance training.), and that inflammation is itself the trigger for adaptation — suppress it and the adaptation goes too. So during a hypertrophy or strength block, don't ice immediately after training — wait at least 4-6 hours; when race prep requires competing again within 24 hours, ice baths are fine, accepting that adaptation cost.
The evidence for sauna is much stronger. The Laukkanen group's Finnish KIHD cohort (2,315 middle-aged men, 20-year follow-up) shows a clear dose-response: 4-7 sauna sessions per week, each over 15 minutes, with cardiovascular mortality down ~40% and all-cause mortality down ~40%. The mechanism is heat-shock-protein induction (a cellular adaptation resembling exercise), nitric-oxide release improving endothelial function, and a peak heart rate of 100-150 bpm that approaches the circulatory load of moderate aerobic exercise but without mechanical wear. Unlike ice baths, a sauna 30 minutes to 2 hours after training doesn't suppress hypertrophy — and the heat-shock-protein induction may even enhance adaptation.
The real trade-off of ice baths (10-15°C immersion for 10-15 min): acutely useful — subjective pain drops in the 1-2 hours after training, and recovery is sped subjectively in multi-day events (Bleakley 2012). But chronic use carries an adaptation cost: Roberts 2015 (J Physiol) RCT showed that over 12 weeks of resistance training, the group icing after every session gained 23% less strength and 18% less lean mass. The mechanism is that cold water suppresses post-training inflammatory signaling (nuclear factor kappa B: The cell's inflammation master switch (a transcription factor) — when flipped, it turns inflammation on., mechanistic target of rapamycin: The cell's master 'grow / build' switch — turned on by enough protein and resistance training.), and that inflammation is itself the trigger for adaptation — suppress it and the adaptation goes too. So during a hypertrophy or strength block, don't ice immediately after training — wait at least 4-6 hours; when race prep requires competing again within 24 hours, ice baths are fine, accepting that adaptation cost.
The evidence for sauna is much stronger. The Laukkanen group's Finnish KIHD cohort (2,315 middle-aged men, 20-year follow-up) shows a clear dose-response: 4-7 sauna sessions per week, each over 15 minutes, with cardiovascular mortality down ~40% and all-cause mortality down ~40%. The mechanism is heat-shock-protein induction (a cellular adaptation resembling exercise), nitric-oxide release improving endothelial function, and a peak heart rate of 100-150 bpm that approaches the circulatory load of moderate aerobic exercise but without mechanical wear. Unlike ice baths, a sauna 30 minutes to 2 hours after training doesn't suppress hypertrophy — and the heat-shock-protein induction may even enhance adaptation.
Hormesis — the shared meta-concept
Sauna, cold exposure, exercise, and fasting share an underlying mechanism called hormesis: mild, brief stress triggers a systemic overshoot in recovery that makes the body healthier long-term, while excessive, chronic stress causes damage. The overall shape is a U-curve, with both an optimal dose and an over-dose risk.This curve explains many seemingly contradictory phenomena: short-term exercise carries oxidative stress and muscle microdamage, yet long-term adaptation brings mitochondrial biogenesis and a strengthened antioxidant system — which is also why taking high-dose vitamin C / E right after exercise can weaken mitochondrial adaptation (Paulsen 2014), because reactive oxygen species (ROS) are themselves the hormesis signal. Likewise, short-term heat or cold stress produces adaptation, while chronic sleep deprivation is chronic stress on the damage side, not hormesis.
The operating principle is one line: regular, controlled, mild stress (exercise, sauna, cold shower, intermittent fasting) tilts healthy; extreme, uncontrolled, chronic stress (life stress, sleep deprivation, chronic fatigue) tilts damaging. 'More is better' is a wrong simplification.
A few influencer claims worth flagging: cold exposure's weight-loss effect is minor — it isn't a real 'cold-burns-fat' route; 'infrared detox' is marketing — sweat mainly clears water and electrolytes, not significant chemicals; practicing Wim Hof breathing in water carries a real risk of shallow-water blackout and drowning — don't. Cold exposure is fine in mild doses (a 30-60 second cold shower is safe), but don't treat it as a core health intervention — sauna is the one with A-grade cardiovascular evidence.
Chapter 7
Training hydration
Training hydration
When dehydration reaches a 2% body-weight drop, endurance performance falls ~20%, strength ~5%, heart rate rises 5-10 bpm at the same intensity, and core temperature climbs 0.3-0.5°C (Sawka 2007 ACSM guidelines). Converted, a 70 kg person at 2% is about 1.4 L of water; moderate training sweat rate is around 0.8-1.5 L per hour, so an hour of moderate training with no fluid already approaches 2% dehydration.
But conversely, most recreational athletes' 'hydration problem' is actually over-hydration, not under. Sweat is over 99% water, with sodium averaging 700-1500 mg per liter (not the 5000 marketing claims), so 1 L of sweat loses about 1 g of sodium — while daily diet already provides 3-5 g. So for training under an hour this sodium loss is nearly negligible, and a cup of plain water is enough.
The scenarios that genuinely need electrolytes are: exercise over 2 hours (especially marathon, triathlon), heat exposure, and the salty-sweater type (white salt rings on clothes after training). 'Daily hydration also needs electrolytes' is basically a marketing trap.
Over-hydration carries a real risk: exercise-associated hyponatremia (EAH, blood sodium below 135 mmol/L). The mechanism is 'drink before you're thirsty' plus prolonged heavy plain-water intake diluting blood sodium. Almond 2005 (NEJM) studied 488 Boston Marathon runners — 13% developed hyponatremia, 0.6% severe, with neurological symptoms cascading from headache and nausea all the way to coma and even death (cerebral edema). So the current consensus (USATF + IMMDA 2020) is to drink to thirst rather than chug on a schedule; for long exercise switch to a sodium-containing sports drink rather than plain water; and if you finish heavier than you started, you drank too much. In one line: trust your thirst.
But conversely, most recreational athletes' 'hydration problem' is actually over-hydration, not under. Sweat is over 99% water, with sodium averaging 700-1500 mg per liter (not the 5000 marketing claims), so 1 L of sweat loses about 1 g of sodium — while daily diet already provides 3-5 g. So for training under an hour this sodium loss is nearly negligible, and a cup of plain water is enough.
The scenarios that genuinely need electrolytes are: exercise over 2 hours (especially marathon, triathlon), heat exposure, and the salty-sweater type (white salt rings on clothes after training). 'Daily hydration also needs electrolytes' is basically a marketing trap.
Over-hydration carries a real risk: exercise-associated hyponatremia (EAH, blood sodium below 135 mmol/L). The mechanism is 'drink before you're thirsty' plus prolonged heavy plain-water intake diluting blood sodium. Almond 2005 (NEJM) studied 488 Boston Marathon runners — 13% developed hyponatremia, 0.6% severe, with neurological symptoms cascading from headache and nausea all the way to coma and even death (cerebral edema). So the current consensus (USATF + IMMDA 2020) is to drink to thirst rather than chug on a schedule; for long exercise switch to a sodium-containing sports drink rather than plain water; and if you finish heavier than you started, you drank too much. In one line: trust your thirst.