Bellenger and colleagues (2016) pooled 24 studies of endurance athletes and found that training blocks which made athletes slower produced a small rise in resting RMSSD, the standard vagal index of heart rate variability. Blocks that made athletes faster raised it too. Their conclusion was that resting HRV is largely unaffected by overreaching, and that other measures are needed to tell good training stress from bad.
Anyone who trains with a wearable meets this problem every morning. The watch reports one number, the number looks low, and you have to decide whether to run the intervals. This article gives you a routine for that decision. It takes about two minutes each morning and five minutes at the end of the week, and it reads four signals together. The first is a Recovery score built from overnight HRV and sleeping heart rate against your own baseline. The second is last night's sleep debt. The third is sleep consistency. The fourth is the seven-day balance between the effort you put in and the recovery your body showed. One bad morning changes the plan only when the pattern holds across days.
01One bad morning is a data point
The joint consensus statement from the European College of Sport Science and the American College of Sports Medicine (Meeusen et al., 2013) separates three states that look the same on any single day. Functional overreaching is a short drop in performance after hard training that resolves within days to weeks and ends with you fitter than before. Nonfunctional overreaching takes weeks to months to resolve and brings no gain. Overtraining syndrome takes months. The panel reviewed hormonal, biochemical, immune, psychological and performance markers and concluded that none of them met the criteria for general use as a diagnostic test.
Two points follow for daily training. A depressed reading the morning after a hard session is what functional overreaching looks like, and a hard block is supposed to produce it. The difference between the productive state and the harmful one is time, so you only learn which one you are in by watching several days.
Plews and colleagues (2013) showed why single HRV readings mislead even elite athletes. In their review, both rises and falls in HRV accompanied negative adaptation, and some athletes gained fitness while their HRV fell. They described HRV saturation, where very fit athletes show lower HRV alongside a lower resting heart rate, a pattern that can accompany high fitness with no fatigue at all. Their fix was averaging and reading each athlete against their own long-run history. Addleman and colleagues (2024) reached the same position in a review for strength and conditioning coaches. They recommend rolling seven-day averages over daily values and treat "higher" and "lower" as meaningful only relative to the individual's baseline.
This blog uses one rule of thumb for HRV outside the app. A morning HRV more than one median absolute deviation below your baseline median for two or more consecutive days, with resting heart rate above baseline, is a signal to reduce training intensity. Each part of the rule answers one of the problems above. The median absolute deviation is your own normal day-to-day spread, so the threshold scales to you. Two consecutive days filters out a single late meal or bad night. The resting heart rate condition screens out the saturation case Plews described, where HRV falls while resting heart rate falls too. There is no fixed millisecond or percentage cutoff, because a 5 ms drop means little for someone whose HRV swings 15 ms a day and a lot for someone whose HRV barely moves.
02Four signals on four time scales
Each signal answers a different question, and they move at different speeds. The timelines below are illustrative.
| Signal | What it compares | Window | What one bad value tells you |
|---|---|---|---|
| Recovery score | Last night's HRV and sleeping heart rate against your baseline | 60-night baseline | Last night sat low against your history |
| Sleep debt | Last night's time asleep against your sleep need | One night | You slept short once |
| Sleep consistency | Last night's time asleep against your 7-day median | 7 nights | Last night's length was unusual for you |
| Effort and recovery balance | Daily Exertion against daily Recovery, averaged | 7 days | Effort has outrun recovery across the week |
Recovery reacts overnight. Sleep debt describes one night. Consistency describes how much your sleep duration varies across the week. The balance is the slowest and the hardest to fool. The morning routine reads them in that order and lets the slower signals decide what the faster ones mean.
03Recovery measures overnight HRV and heart rate against your own history
Titan's Recovery score scores the night you just slept from two inputs. Overnight HRV is the average of every HRV reading recorded during your main sleep, taken on a log scale. Sleeping heart rate is the lowest 30-minute rolling average of your heart rate during that sleep. When no sleep session was saved, Titan reads 10 PM to 10 AM and marks the score Partial night. Wear the watch to bed, since both inputs come from the night.
Your baseline for each input is the median of your measured nights before the one being scored, across a 60-night window by default. Titan measures how far last night sits from that median in units of your median absolute deviation, which is your usual night-to-night spread. HRV counts about twice as much as sleeping heart rate. A night exactly at both baselines scores 56. With heart rate at baseline, HRV one step of your usual spread above its median scores 85, and one step below scores 22.
A 60-night baseline reflects your usual readings over a long period, so a rough week shows as a drop instead of pulling the reference down with it. That suits the pattern-based routine in this article. Personal baselines explains how the baseline builds.
A baseline needs at least 7 measured nights covering a quarter of the window, which is 15 nights on the default 60. Until then Recovery has no score to show, so treat your first few weeks of scores as provisional. A night with heart rate but no HRV gets an Estimated score, and the Recovery screen names the reason. Each night's score becomes final at noon on the day you wake.
Reading the three recovery bands
| Band | Score | Titan's guidance | How to use it in the routine |
|---|---|---|---|
| Low | 0 to 39 | Favor easy work, sleep and nutrition | Check whether yesterday also scored low and resting HR is up |
| Moderate | 40 to 74 | Train, but keep an eye on intensity | Run the planned session and stop adding reps if it feels flat |
| High | 75 to 100 | Ready for quality training | Put the key session of the week here when the schedule allows |
The band describes one night only. A Low score the morning after the hardest session of the block is the expected cost of that session. A Low score on the third morning in a row, with resting heart rate above baseline, meets the HRV rule and calls for a different response.
04Sleep numbers that matter for recovery
Sleep is the recovery input you control most directly, and the research gives specific numbers.
| Finding | Number | Source |
|---|---|---|
| Adequate sleep for adults in IOC and NCAA guidance | At least 7 hours | Charest and Grandner (2020) |
| Athletes' measured sleep compared with non-athletes | 6.55 h against 7.11 h | Leeder (2012), via Charest and Grandner |
| Athletes who report sleeping under 7 hours | 39.1% | Mah (2018), via Charest and Grandner |
| Injury history in adolescent students sleeping under 8 hours | 1.7 times as likely to report an injury | Milewski (2014), via Charest and Grandner |
| Cyclists after 4 hours of sleep restriction | 15 W lower work rate in a 30 minute effort | Charest and Grandner (2020) |
| Basketball players after 5 to 7 weeks of 10 hours in bed | Sleep rose from 6.7 h to 8.5 h | Mah et al. (2011) |
The Mah study is the clearest case for sleep as a performance lever. Eleven collegiate basketball players spent 2 to 4 weeks on their normal schedule, then aimed for 10 hours in bed each night for 5 to 7 weeks. Actigraphy showed their actual sleep rise from 6.7 to 8.5 hours. Their 282 foot sprint time fell from 16.2 to 15.5 seconds. Free throws went from 7.9 to 8.8 made out of 10, and three-point shots from 10.2 to 11.6 made out of 15.
Two details in that study shape how to measure sleep. The players needed 10 hours in bed to get about 8.5 hours of sleep, so time in bed overstates what you get. Their baseline of 6.7 hours also sat close to the average Leeder measured in athletes, which suggests many athletes carry a nightly shortfall they never notice.
05Sleep debt counts one night at a time
Titan's sleep debt is the number of hours you slept short of your sleep need on a single night. It counts only time asleep, so awake time in bed adds nothing. Your need is one setting, the Sleep Debt Threshold, which defaults to 7.5 hours and runs from 5 to 12 hours in half-hour steps.
Each night stands alone. A sleeper at the Mah baseline of 6.7 hours with a 7.5 hour need shows 0.8 hours of debt every night. Titan draws seven bars of 0.8 hours and never adds them into a total of 5.6. A 9 hour Saturday shows 0 debt and pays nothing back, so the short nights keep their bars. That design stops a single long night from hiding a week of short ones, which is exactly what the weekend catch-up plan tries to do.
For the morning routine, look at last night's bar and at how many of the last seven nights carried debt. One short night before an easy day needs no response. Four or five short nights across a hard week explain a sliding Recovery score better than any training variable.
Two settings deserve a check. Titan's Training Goal does not change your sleep need, so if you move into a hard block and want a higher need, raise the threshold yourself at Settings > Recovery & Sleep. The Mah results suggest 8 hours is a reasonable threshold during heavy training. Your Sleep goal, which the Sleep score uses, is a separate profile setting with an 8 hour default. Set the two to the same number so every sleep figure points at one target.
06Sleep consistency is a separate signal
Windred and colleagues (2024) followed 60,977 UK Biobank participants for a mean of 6.3 years after seven days of wrist accelerometry. They scored each person with the Sleep Regularity Index, which measures how often you are in the same sleep or wake state at the same clock time 24 hours apart. After adjustment, the most regular fifth had a hazard ratio of 0.70 for all-cause mortality compared with the least regular fifth, and regularity predicted mortality better than sleep duration did. The cohort averaged 63 years old and the outcome was death, so the study says nothing direct about next week's intervals. It does show that the regularity of sleep carries information that total hours miss.
Titan's sleep consistency measures a related thing, the regularity of how long you sleep. For each night, Titan takes the median time asleep across the 7 days ending that night, measures how far the night sits from it in either direction, divides the gap by 2 hours and subtracts the result from 100%.
| Difference from 7-day median | Consistency |
|---|---|
| 0 h | 100% |
| 0.5 h | 75% |
| 1 h | 50% |
| 1.5 h | 25% |
| 2 h or more | 0% |
A night an hour longer than your median scores the same as a night an hour shorter. Titan needs at least 3 nights with sleep data in the window to score a night.
The reason to track consistency next to debt shows up in two ordinary weeks, both with a 7.5 hour need.
| Week pattern | Sleep debt | Consistency |
|---|---|---|
| 6.5 hours every night | 1.0 h on every night | 100% on every night |
| Six nights at 6.5 hours, then a 9 hour night | 1.0 h on the short nights, 0 h on the long one | 0% on the long night |
The steady short sleeper looks perfect on consistency and short on debt every night. The weekend sleeper's long night clears the debt bar and scores 0% consistency, because 9 hours sits 2.5 hours from a 6.5 hour median. Neither number alone describes either week. Read together they do, and they point to the same fix, which is a longer and steadier night on training days.
Hold the wake time steady
Titan also shows a Target Wake Time on the Sleep screen, the median of the times your sleep ended across the nights recorded in the last 14 days. Wake time is the clock anchor the Sleep Regularity Index rewards, and it is the one time you control on most days. Keep it within 30 minutes of the target and move bedtime earlier to protect duration. To keep consistency at 75% or better, stay within 30 minutes of your usual time asleep.
07Effort against recovery over a rolling week
The last signal answers the question the consensus statement says no single marker can, which is whether the load has run ahead of your capacity to absorb it. Titan's Training balance chart puts Exertion, a 0 to 10 daily effort score from heart rate zone minutes and steps, on the same scale as Recovery by multiplying it by 10. It subtracts that day's Recovery score to get a daily value, then averages the daily values over the 7 days ending on each date.
| Exertion | Exertion × 10 | Recovery | Daily value |
|---|---|---|---|
| 6.5 | 65 | 58 | +7 |
| 8.0 | 80 | 45 | +35 |
| 2.0 | 20 | 80 | -60 |
The chart draws a single reference line at zero. Above it, effort is ahead of recovery. Below it, recovery is ahead of effort. Titan labels no zones on the chart, so the thresholds this routine uses are this article's recommendation. Treat a seven-day average above +10 as effort outrunning recovery, and one below −10 as recovery outrunning effort.
The seven-day average is where a single bad morning loses its weight. A hard day at +35 followed by six days at 0 averages +5, well short of +10. To push the line above +10, effort has to beat recovery on several days, or recovery has to sag while effort holds. The second case is easy to miss, because a falling Recovery score raises the daily value even when your training stays the same.
A planned hard block will push the line above +10 for a stretch. Meeusen's functional overreaching looks like this, and the line should drop back within days of the block ending. While it sits above +10, consider tapering intensity or adding a lighter day. The reading to take seriously is a line that stays above +10 after the planned hard days are over, while the Recovery history keeps sliding.
A reading below −10 depends on your training goal. In a building or peak phase, it can mean you have room to add load gradually if you feel fresh. During a taper or a rest week, it is the expected result.
08Rest day or active recovery
Light movement has good evidence within a session. Nalbandian and colleagues (2017) had 17 trained men ride three 30 second bouts at 90% of peak power with 4 minutes between them, either pedaling at 60% of lactate threshold or sitting still. Active recovery kept blood lactate lower, and the Wingate test that followed produced 715 W of peak power against 671 W after passive rest, a 6.6% gain, with mean power 5.3% higher. Those are minutes-scale effects. The case for an easy day between hard days is simpler. Active recovery keeps aerobic volume in the week at an Exertion cost low enough that the balance line barely moves.
Full rest is for repeated signs of poor recovery. Here is how the four signals combine into a decision.
| What the morning shows | What to do today |
|---|---|
| Recovery Moderate or High, little sleep debt, balance within ±10 | Train as planned |
| One Low morning with a clear cause such as a short night, balance within ±10 | Easy aerobic work, and move the quality session back a day |
| HRV rule met, balance within ±10 | Keep intensity out until HRV is back within one deviation of baseline |
| HRV rule met, balance above +10, several nights of debt | Take a full rest day, follow it with easy days, and plan a lighter week |
| Recovery High, balance below −10, goal is Peak Training or Building | Add load gradually, one session at a time |
The middle rows catch two opposite mistakes. A single Low morning with an obvious cause gets treated like a crisis and costs a good session. Three Low mornings with rising resting heart rate get ignored because the athlete feels fine. The routine removes both mistakes by asking whether the pattern holds before it asks how you feel.
09A five-minute weekly review
At the end of each training week, answer four questions in order.
- How many of the last seven nights carried sleep debt, and did any exceed an hour? Open the Sleep screen and set the Debt tab to 7 Days.
- Did consistency stay at 75% or better on most nights? Switch the same chart to Consistency.
- Did Recovery stay steady compared with the previous week? The Recovery history section compares your last 7 days with the 7 before, once each week has 4 scored days.
- Where does the Training balance line sit, and which way is it heading?
The answers set next week's load. A balance within ±10 with steady Recovery and few debt nights means a normal week. A balance above +10 at the end of the week, or a week in which the HRV rule fired, means a lighter week with the intensity cut first. If balance stays above +10 and the Recovery comparison falls, check your training log and performance, sleep, and other stressors. If performance drops or familiar sessions also feel harder, consider bringing forward a planned deload week. If performance keeps falling and the signals do not recover after a lighter week or two, you may be past functional overreaching into the state Meeusen and colleagues describe as taking weeks to months to resolve, and that belongs with a sports medicine clinician. For how to rebuild load after a lighter week, see progressive overload planning, and for more on reading HRV trends, see the HRV training readiness guide.
10Training balance in Titan Trends
The weekly half of this routine lives on one chart. Tap the Trends tile on Today and scroll to Training balance, just below Recovery vs exertion. Trends opens on the Year period, and the period menu at the top switches every chart to 30 Days, 60 Days or 6 Months. The chart plots the seven-day average of Exertion against Recovery described above, with a reference line at zero, and it needs at least 3 of the 7 days to carry both scores before a point appears. Sleep debt lives on the Sleep screen, under the Debt tab of Sleep history. The full calculation is in the Recovery-Based Training help article.
11References
Addleman JS, et al. (2024). Heart rate variability applications in strength and conditioning: a narrative review. Journal of Functional Morphology and Kinesiology. https://pmc.ncbi.nlm.nih.gov/articles/PMC11204851/
Bellenger CR, et al. (2016). Monitoring athletic training status through autonomic heart rate regulation: a systematic review and meta-analysis. Sports Medicine. https://doi.org/10.1007/s40279-016-0484-2
Charest J, et al. (2020). Sleep and athletic performance: impacts on physical performance, mental performance, injury risk and recovery, and mental health. Sleep Medicine Clinics. https://pmc.ncbi.nlm.nih.gov/articles/PMC9960533/
Mah CD, et al. (2011). The effects of sleep extension on the athletic performance of collegiate basketball players. Sleep. https://pmc.ncbi.nlm.nih.gov/articles/PMC3119836/
Meeusen R, et al. (2013). Prevention, diagnosis, and treatment of the overtraining syndrome: joint consensus statement of the European College of Sport Science and the American College of Sports Medicine. Medicine and Science in Sports and Exercise. https://doi.org/10.1249/MSS.0b013e318279a10a
Nalbandian HM, et al. (2017). Active recovery between interval bouts reduces blood lactate while improving subsequent exercise performance in trained men. Sports. https://pmc.ncbi.nlm.nih.gov/articles/PMC5968977/
Plews DJ, et al. (2013). Training adaptation and heart rate variability in elite endurance athletes: opening the door to effective monitoring. Sports Medicine. https://doi.org/10.1007/s40279-013-0071-8
Windred DP, et al. (2024). Sleep regularity is a stronger predictor of mortality risk than sleep duration: a prospective cohort study. Sleep. https://doi.org/10.1093/sleep/zsad253
