Adjusting training to recovery

In 2007, Antti Kiviniemi and colleagues in Oulu, Finland, gave 26 moderately fit men four weeks of 40-minute runs. One group followed a fixed plan of four hard days and two easy days a week. A second group measured heart rate variability every morning and ran hard only when it had held steady or risen. When it fell below their own reference range, or trended down for two days, they ran easy or rested. The fixed group's peak oxygen uptake did not change significantly (54 to 55 ml/kg/min). The HRV-guided group went from 56 to 60, and its gain in maximal running speed was larger too (Kiviniemi, 2007). The difference was timing. The guided group put its hard days on mornings when its HRV said it could absorb them.

That is the whole case for adjusting training to recovery. The plan decides where you are going over weeks and months. The morning data decides how much of today's planned work you can turn into adaptation. This article covers the signals worth reading, how Titan reports them, a decision tree for low-recovery days, a worked week of decisions, and the rules for when and how to deload. The wearable data quality guide covers how accurate each signal is before you act on it. The performance physiology guide explains the endurance physiology under these decisions.

01The plan sets direction, the day sets the dose

Adjustment only works on top of a structure. You still need periodization to sequence training emphases, a mesocycle of a few loading weeks followed by lighter work, and a method of progressive overload that raises the demand over time. Linear, undulating, block and conjugate models all work with the rules below, because each one prescribes a target and each target can be scaled on the day. The choice between them, and the set, rep and load prescriptions that go with them, live in the progressive overload guide and the muscle building guide.

What changes with a recovery-led approach is the form of the prescription. A fixed plan says "squat 120 kg for 4 sets of 5 on Tuesday." An adjustable plan says "squat 4 to 5 sets of 5 at RPE 8 on Tuesday, with the load and the fifth set decided by this morning's recovery and the warm-up." The first version is wrong on any day your capacity differs from the day you wrote it. The second has a range built in, and a set of rules for moving inside it.

Those rules have to be written before the bad morning arrives. An athlete who decides how to respond to low recovery while standing under the bar will either push through or skip the session, and neither is a plan.

02What the morning signals can tell you

Heart rate variability against your own baseline

Heart rate variability reflects parasympathetic activity at rest, and a fall below your normal range is one of the more reliable early signs that training, illness, poor sleep or life stress is outrunning recovery. The raw number carries little meaning across people. A reading of 45 ms is high for one athlete and a warning for another, so every useful rule compares you with yourself.

Titan does this with medians instead of a fixed percentage. Recovery reads two values from the night you just slept. One is overnight HRV. The other is sleeping heart rate, the lowest 30-minute average of your heart rate during that sleep. Each is compared with its baseline, the median of your earlier measured nights in the Recovery window, which is 60 nights by default. Titan scales the gap by your usual night-to-night spread, measured with the median absolute deviation, so a 5 ms drop counts for more if your HRV rarely moves. If you want a rule of thumb for reading your raw numbers outside the app, use this one. 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.

The two-day requirement matters. Single readings move with alcohol, a late meal, a hot room or the position you slept in. Martin Buchheit's review of heart rate monitoring concluded that most contradictory findings in the literature come from inconsistent measurement and misreading of the data rather than from limits of heart rate itself (Buchheit, 2014). Repeated morning readings, compared with your own baseline, remove much of that noise.

Resting heart rate as the cross-check

Resting heart rate is the slower, cruder partner to HRV. A rise above baseline that persists for several mornings often accompanies accumulated fatigue or the start of an illness. On its own it produces false alarms. Paired with falling HRV, it strengthens the case to back off. That is why the rule above asks for both.

When a high reading is not good news

HRV that climbs above baseline usually signals good recovery. There is one exception worth knowing. Yann Le Meur's group put 13 trained triathletes through three weeks of overload training, with 8 more as controls, and found that the athletes who became functionally overreached showed a progressive increase in parasympathetic HRV. The change showed up in daily readings averaged by week and was missed by single weekly recordings (Le Meur, 2013). A high Recovery score paired with falling performance, heavy legs and rising effort at familiar loads belongs with the deload triggers below.

Sleep

Sleep feeds both HRV and resting heart rate, so it usually shows up in the Recovery score before you look at it directly. It also affects performance on its own. When Cheri Mah had 11 Stanford basketball players extend their sleep for five to seven weeks, sprint times fell from 16.2 to 15.5 seconds and free throw accuracy rose 9 percent (Mah, 2011). A short night is a reason to check the other signals more carefully. Titan tracks nightly shortfall against your need as sleep debt.

Perceived effort in the warm-up

The last signal is the one you collect in the gym or on the road. Rating of perceived exertion expressed as reps in reserve, validated for the squat by Michael Zourdos and colleagues (Zourdos, 2016), gives a set-by-set read on capacity. RPE 8 means you could complete about two more reps, RPE 9 means one. If warm-up sets at familiar loads feel a full point heavier than usual, that is information the overnight data cannot give you.

Your own readiness rating

Rate how ready you feel on a 1 to 5 scale before you look at the device. If you see a low score first, you will feel worse, and the rating stops being an independent check. Recorded first, it can confirm or contradict the data. Anna Saw and colleagues reviewed 56 studies and found that athletes' self-reported measures responded to training load with greater sensitivity and consistency than commonly used objective measures, and that the two often did not correlate (Saw, 2016). When your rating and the device disagree, the disagreement is worth investigating.

No one of these signals should drive a decision alone. The 2013 joint consensus statement from the European College of Sport Science and the American College of Sports Medicine on overtraining makes the same point about diagnosis. No single marker separates normal fatigue from non-functional overreaching, and performance decline with persistent fatigue is the defining feature (Meeusen, 2013).

03How Titan reports readiness

The Recovery score

The Recovery score runs from 0 to 100 and describes the night you just slept. Titan compares overnight HRV and sleeping heart rate with their baselines and places the result on a fixed curve. A night with both values at baseline scores 56. HRV one usual night-to-night step above baseline scores 85, and one step below scores 22. Sleeping heart rate counts about half as much as HRV, so a rise in heart rate pulls the score down less than a matching fall in HRV. The detail screen assigns one of three bands.

The baseline covers the measured nights before the one being scored, across the Recovery window. The 60-night default reflects your longer-term baseline, so a rough week shows as a drop instead of pulling the baseline down with it. Titan needs 7 measured nights covering at least a quarter of the window before it scores against a baseline, which is 15 nights on the default window.

Training load and the load ratio

Training load measures accumulated training. Titan weights each workout minute by its heart rate zone, so a minute in Zone 4 counts four, and a workout without heart rate counts by its length and activity type. It then keeps two exponentially weighted averages of daily load. Short-term load uses a 7-day time constant and stands for fatigue. Long-term load uses a 42-day time constant and stands for fitness. Their ratio shows how far recent training has moved from what you are used to. Titan labels a ratio below 0.8 Low, 0.8 up to 1.4 Optimal, 1.4 up to 1.6 High, and 1.6 or above Risk.

Read the ratio as a description of change. Franco Impellizzeri and colleagues showed that the acute-to-chronic ratio has statistical problems of its own, and that no study has shown steering it changes injury rates (Impellizzeri, 2020). A ratio of 1.5 tells you this week has been much heavier than your six-week average. That is useful when you are deciding whether a low Recovery score is a one-night blip or the cost of a big block.

Training balance

The Training balance chart in Trends, described in recovery-based training, compares Exertion with Recovery. Each day Titan multiplies your Exertion score by 10, which puts it on the same 0 to 100 scale as Recovery, subtracts your Recovery score, and plots a 7-day average against a dashed line at zero. Above the line, effort has been outpacing recovery. Below it, recovery is ahead. This is the chart to check at the end of a week, because it shows whether the daily decisions added up to the balance you intended.

The Exertion target

The Exertion target is the one place where Titan turns these inputs into a number for today. It starts from the base range for your training goal.

Recovery then slides the range. At a score of 60 nothing moves. Each point above or below shifts the range by 0.02, up to 0.8 either way. Your last three scored days of Exertion can move it another 0.35. On a Building day with Recovery at 30, the range drops from 5.5 to 7.5 to 4.9 to 6.9. If you have also averaged above 8.0 over the last three days, it drops another 0.35. With Recovery at 80 the range rises to 5.9 to 7.9, and once your Exertion is inside it with Recovery at 65 or higher, a stretch target up to 9.5 opens.

Titan does not plan your session, so the target works best as a limit on low-recovery days and a guide to additional work on high-recovery days. If the target has dropped by half a point or more, your planned hard session should shrink to fit it.

04A decision tree for low-recovery days

Use this sequence when the Recovery score lands in the Low band, or when the HRV rule above has fired, before you change anything else.

1. Check whether the signals agree. Look at resting heart rate against baseline and last night's sleep. A low score with normal resting heart rate after one short night is a weaker signal than a low score with raised resting heart rate on the second consecutive morning.
2. Check the load context. If your load ratio is above 1.4 or Training balance has sat above +10 for several days, the low score is probably the cost of recent work. Treat it as a reason to bring the next deload forward.
3. Cut intensity first. Drop today's planned loads by 5 to 10 percent, or convert interval work to steady aerobic running or riding at conversational pace. Keep the same exercises so you still practice the movements.
4. Cut volume second. Remove one or two sets per exercise, or shorten the main set of an endurance session. Aim to finish inside the Exertion target Titan shows for the day.
5. Use the first working set as the final check. If it comes in two or more RPE points above target, stop loading. Turn the rest of the session into technique work at 50 to 60 percent of your working weight, mobility, or easy Zone 1 and 2 movement.
6. Log the change. Add a custom Journal tag such as "Modified session." Journal Insights counts the days that carry each tag across a 60-day window, so you can see how often low recovery changed your plan within a block.

The same logic runs in reverse on High days. If Recovery is 75 or above, the stretch target is open and warm-up sets feel light, add one set to the main lift or a few minutes to the main interval block. Do not add a new exercise or a new intensity. The goal is a slightly larger dose of the planned stimulus.

Inside the session, rate each working set against its RPE target and adjust.

If RPE is on or above target across two or more sessions at the same loads, stop progressing the load. That trend is also the first deload trigger below.

05One week of decisions

The decision tree handles a single low morning. A full week shows how the HRV rule and small adjustments work together. This one belongs to a recreational runner preparing for a half marathon. The baseline HRV median is 55 ms with a median absolute deviation of 6 ms, so the line for the rule is 49 ms. The resting heart rate median is 54 bpm. The numbers are illustrative.

The runner made small changes on borderline days and saved the large change for the day the rule was met. No single number decided anything, and the readiness ratings agreed with the device all week. Full training resumed once the numbers returned to their normal spread, whatever the calendar said.

06Common decision errors

Reacting to one morning. This is the most frequent error. If you change sessions more than twice a week because of wearable data, you are probably reacting to noise, and the changes break up the planned progression.

Missing a slow trend. Five mornings of drifting HRV with resting heart rate up 4 bpm is a signal even if no single morning looks dramatic.

Ignoring how you feel. A wearable cannot measure joint soreness, motivation, focus or mood. An athlete who feels sharp and ready should not cancel a session because a score reads 45. An athlete who feels flat with a score of 90 should proceed with care.

Treating the model as the athlete. A recovery score is a model of your state. When the model and your performance disagree, give more weight to performance and find out why the model is off.

07Deload rules

A deload is a planned period of lower training stress that lets fatigue clear while fitness holds. Lee Bell and colleagues surveyed expert strength and physique coaches through a three-round Delphi process and found agreement on a working definition and a typical pattern. Most coaches deload every four to six weeks, for about seven days, mainly by cutting volume (Bell, 2023).

When to deload

Planned deloads come at the end of each mesocycle. Three loading weeks and one deload week is the usual starting point. Athletes who handle higher loads can stretch to four or five loading weeks. Athletes who feel flat by week three should shorten the block.

Reactive deloads come earlier, when two or more of these signs appear together.

• RPE runs one to two points above target at the same loads for two or more sessions.
• The HRV rule stays active, or the Recovery score sits in the Low band, for most of a week.
• Training balance stays above +10 for most of a week.
• The load ratio stays above 1.4 without a planned reason.
• Sleep has been short or broken for three or more nights.
• Loads or times regress over two or more sessions.
• Soreness or joint pain does not clear between sessions, or motivation drops sharply.

The list asks for two signs because the Meeusen consensus is clear that single markers mislead. Performance is the tiebreaker. A high Recovery score with falling performance still counts toward a deload, per the Le Meur finding.

How to deload

Cut volume by 40 to 60 percent. Halving the working sets per exercise is the simplest way. Keep the rep ranges. For intensity there are two valid options. Keep loads at 85 to 90 percent of your heaviest week to hold neural qualities, which suits strength athletes. Or drop both load and volume when fatigue is high or the deload was reactive. Laurent Bosquet's meta-analysis of tapering found that cutting volume by 41 to 60 percent while keeping intensity gave the largest performance gains, which is why the volume cut does most of the work in either version (Bosquet, 2007).

A deload usually lasts five to seven days. A light mesocycle may need only three or four. A reactive deload after non-functional overreaching may need seven to ten.

What a deload looks like in Titan

Switch your training goal to Tapering for the week. Titan describes it as "Reducing training load to shed fatigue," and it moves the Exertion target base range to 2.5 to 4.5. Expect the load ratio to fall below 0.8. If you have trained at a steady daily load of 150 and halve it for one week, short-term load drops to about 103 and long-term load to about 139, a ratio near 0.74. Expect Training balance to drift below zero as recovery pulls ahead of effort. For an injury or illness, use Rest Mode, which holds the target at 0.0 to 2.0 whatever your goal.

The deload worked if Recovery climbs back to its usual band, resting heart rate settles at or below baseline, soreness clears and the first heavy session of the next block matches or beats the best session of the last one. If performance does not rebound, either the deload was too short or too light, so extend it by three or four days, or the fatigue is coming from outside training.

Returning to full training

Start the next mesocycle at week-one volume and effort, with loads slightly above the previous week one. Do not jump back to the heaviest week. Switch the training goal back to Building or Peak Training on the first day of the new block so the Exertion target follows the plan again.

08An annotated 16-week macrocycle

A macrocycle is the long cycle that contains the mesocycles. This hypothetical one follows an intermediate lifter who trains four days a week and wants to test his squat, bench press and deadlift at week 16. It shows where the rules above change the plan. The set and load prescriptions for each phase belong in the progressive overload guide.

Weeks 1 to 6, accumulation

The first block builds work capacity with higher volume at moderate loads. Weeks 1 to 3 add one set per main lift each week, week 4 is a planned deload with the goal set to Tapering, and weeks 5 and 6 restart the build slightly heavier.

In week 2 he sleeps five hours before an upper body day. His Recovery score is 34, resting heart rate is up, and his Building target has dropped to 5.0 to 7.0. He keeps the exercises, drops one set from each and finishes inside the target. In week 3 he wakes to a Recovery of 82 with the stretch target open, and warm-up squats move easily. He adds a sixth set at the same load and stops there. In week 5 a weekend hike leaves one knee sore. He swaps back squats for leg press for the week and returns to squats in week 6.

Weeks 7 to 12, strength

The second block raises intensity and cuts accessory volume. A planned deload falls in week 10.

In week 8 a work deadline cuts his sleep to about six hours for five nights. By Wednesday the HRV rule has fired for two mornings, working sets at RPE 8 loads are rating 9.5, and Training balance has crossed +10. Two triggers agree, but the cause is outside training and has an end date. He holds week 8 loads instead of progressing, sets a nine-hour sleep window and agrees with himself that one more low morning means a reactive deload. The deadline passes, Recovery returns to the Moderate band by the following Monday, and he repeats week 8 loads in week 9 on target. In week 11 his deadlift moves faster than expected and the first working set rates RPE 7 against a target of 8, so he adds 5 kg over the plan.

Weeks 13 to 16, peaking

The last block cuts volume hard and keeps loads high, then ends with a test. He sets the goal to Tapering for weeks 15 and 16. His Recovery scores rise through the taper and Training balance drops below zero, which is what a taper should look like. In week 15 an opener single feels heavier than expected while every overnight signal is normal. He reads that as nerves and keeps his planned attempts.

Across 16 weeks he made five changes to the plan. The deloads landed where they were written, and the week 8 slump cost one week of progression.

09Training balance and the Exertion target in Titan

The two screens that carry this method are the Exertion target, which sets today's ceiling from your training goal, Recovery and recent load, and the Training balance chart in Trends, covered in recovery-based training, which shows whether a week of those decisions kept effort and recovery in balance. Tap the Exertion ring on Today for the first. Tap the Trends tile on Today and scroll to Training balance for the second. Set your training goal to match the block you are in, check the Recovery score each morning, and read both against your training load.

10References

Bell L, Strafford BW, Coleman M, et al. Integrating deloading into strength and physique sports training programmes: an international Delphi consensus approach. Sports Medicine Open, 2023, 9:87. https://doi.org/10.1186/s40798-023-00633-0

Bosquet L, Montpetit J, Arvisais D, Mujika I. Effects of tapering on performance: a meta-analysis. Medicine & Science in Sports & Exercise, 2007, 39(8):1358-1365. https://doi.org/10.1249/mss.0b013e31806010e0

Buchheit M. Monitoring training status with HR measures: do all roads lead to Rome? Frontiers in Physiology, 2014, 5:73. https://doi.org/10.3389/fphys.2014.00073

Impellizzeri FM, Tenan MS, Kempton T, Novak A, Coutts AJ. Acute:chronic workload ratio: conceptual issues and fundamental pitfalls. International Journal of Sports Physiology and Performance, 2020, 15(6):907-913. https://doi.org/10.1123/ijspp.2019-0864

Kiviniemi AM, Hautala AJ, Kinnunen H, Tulppo MP. Endurance training guided individually by daily heart rate variability measurements. European Journal of Applied Physiology, 2007, 101(6):743-751. https://doi.org/10.1007/s00421-007-0552-2

Le Meur Y, Pichon A, Schaal K, et al. Evidence of parasympathetic hyperactivity in functionally overreached athletes. Medicine & Science in Sports & Exercise, 2013, 45(11):2061-2071. https://doi.org/10.1249/MSS.0b013e3182980125

Mah CD, Mah KE, Kezirian EJ, Dement WC. The effects of sleep extension on the athletic performance of collegiate basketball players. Sleep, 2011, 34(7):943-950. https://doi.org/10.5665/SLEEP.1132

Meeusen R, Duclos M, Foster C, et al. 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 & Science in Sports & Exercise, 2013, 45(1):186-205. https://doi.org/10.1249/MSS.0b013e318279a10a

Saw AE, Main LC, Gastin PB. Monitoring the athlete training response: subjective self-reported measures trump commonly used objective measures: a systematic review. British Journal of Sports Medicine, 2016, 50(5):281-291. https://doi.org/10.1136/bjsports-2015-094758

Zourdos MC, Klemp A, Dolan C, et al. Novel resistance training-specific rating of perceived exertion scale measuring repetitions in reserve. Journal of Strength and Conditioning Research, 2016, 30(1):267-275. https://doi.org/10.1519/JSC.0000000000001049