Week 4 is the fourth active Sleepgenic Weekly Report, and it should be read directly against Week 3. The Week 3 report ended with an explicit watch item: REM sleep had declined for two consecutive weeks and deep sleep had fallen below one hour, while the score layer stayed stable. The question Week 4 was set up to answer was simple — would the next week convert that thinning architecture back into recovery, or would the scores follow the architecture downward?
Week 4 answered it. The scores followed.
The training week kept the same structure as the three weeks before it: Rest, Summit, two Recovery days, Rucking, Run, Walk. The stimulus rhythm did not change across the entire four-week run. That is what makes the Week 4 result meaningful — the body was responding to a familiar load, not a heavier one, and it responded worse than in any prior week.
At the score layer, Week 4 was the weakest of the run. The average Overall Score was 52.1, down from 68.1 in Week 3 and 70.7 in Week 2 — a drop of 16 points week over week, and the lowest weekly average recorded, below even Week 1's 60.2. Quality Score fell to 59.9 and Duration Score to 64.7. The decline was not evenly spread. Two nights drove most of it: May 9, after the Rest day, scored 37, and May 14, after the Run, scored 35 with a Recovery Score of just 21 — the single lowest sub-score of the entire run. Both were tagged "Negative Short And Poor Quality." The other five nights held in the 50s and 60s. This was not a week where every night collapsed. It was a week with two genuine failure nights and five mediocre ones, and no strong night at all — nothing above 67.
The physiological layer is where Week 4 is most stark, and it confirms the Week 3 warning directly. REM sleep averaged 0.17 hours. That is not a decline; it is a near-disappearance. REM had already fallen from 1.10 hours in Week 2 to 0.67 in Week 3 — and in Week 4 it effectively collapsed. Three nights recorded REM at or below 0.05 hours, and two further nights returned no REM reading at all. For practical purposes, restorative REM sleep was absent for most of the week. Every single night carried a feedback tag referencing insufficient or poorly structured REM. This is the clearest single-metric signal in four weeks of data.
Sleep Stress moved the same direction. It averaged 23.6, up sharply from 14.9 in Week 3 and 15.8 in Week 2 — the highest weekly Sleep Stress of the run, above even Week 1's 20.4. The two failure nights carried Sleep Stress readings of 33.5 and 41.3, far outside anything seen in the prior three weeks. HRV, which went largely unrecorded in Week 3, was captured this week and averaged 37.6 ms — down from Week 3's (sparsely measured) readings and back to Week 1 levels. Resting heart rate rose to an average of 62.9 bpm, up from 58.9 in Week 3, with two nights at 66 and one at 71. The autonomic markers, the architecture markers, and the stress markers all moved together, in the same direction, for the first time in the run.
What kept Week 4 from reading as a complete breakdown was the Recovery Score and total sleep. Recovery Score averaged 63.4 — lower than Weeks 2 and 3, but not collapsed, and three nights still posted Recovery Scores above 84. Total sleep held at 6.40 hours, nearly identical to Week 3. The body was still spending adequate time asleep, and on its better nights still registering recovery. Deep sleep actually rose, averaging 1.47 hours, the highest of the run. That last point matters: deep sleep was not the casualty this week. REM was. The body was prioritizing deep sleep and sacrificing REM — a recognizable signature of a system under load, protecting the most metabolically essential sleep stage while the more fragile one thins out.
The interpretation is that Week 4 is the surfacing of the strain Week 3 identified. Week 3 showed stable scores over a weakening architecture and named the divergence as latent strain. Week 4 is that strain becoming visible — the score layer dropping into line with the physiological layer it had been masking. The REM collapse is the throughline: it declined in Week 3 beneath stable scores, and in Week 4 it bottomed out while the scores finally fell with it.
Sleepgenic classifies Week 4 as strain surfacing. It is not a catastrophic week — Recovery Score held, deep sleep held, total sleep held, and most nights were mediocre rather than failed. But it is unambiguously the low point of the four-week run, and it is the predicted low point. The Week 3 report said this was the week the strain would either reverse or become visible. It became visible.
The watch item going forward is REM recovery, and it is the only watch item that matters right now. REM has declined for three consecutive weeks — 1.10, 0.67, 0.17 hours — against an unchanged training load. Deep sleep is intact and resting heart rate is not alarming, so the system has capacity. The question for Week 5 is whether removing or reducing load allows REM to rebuild. If REM returns toward the 1.0-hour range, Week 4 will read as the trough of a recoverable strain cycle. If REM stays suppressed into a fourth week, the interpretation changes from strain to something more persistent, and the methodology's next job becomes identifying what is blocking REM specifically.