The self-quantified, data-to-action human.
On 21 April 2026 I broke my right femur in a low-speed cycling crash and had surgery the same day. A subtrochanteric fracture, fixed with an intramedullary nail. Six weeks non-weight-bearing. The clinical detail, the implant spec and the day-by-day log all live on the recovery hub. What I want to write about here is narrower and more useful to me: I spent four months trying to measure a thing that mostly refused to be measured, and I learned which numbers were worth looking at.
Six weeks of tracking almost nothing
Non-weight-bearing is a strange state for someone whose entire self-image runs on Garmin data. Every metric I normally live by went dark at once. No pace, no power, no training load, no volume. The one number that kept moving was weight, and it moved the wrong way — up. That was deliberate. A fracture heals on a surplus, not a deficit, so I came off the GLP-1 protocol the day of the surgery and let myself eat. Weight went from around 93 kg back to roughly 96 kg over that window, and I would take that trade again. Carrying a couple of extra kilos while you are on two crutches costs you nothing. Under- feeding a healing femur costs you the femur.
What was actually trackable in that period turned out to be a short list: range of motion in degrees, whether I could complete a straight-leg raise unassisted, painkiller use, and sleep. That last one was worth more than it sounds. Nothing else in those first weeks gave a daily read on how the leg was doing.
Everything else was theatre. I tried, in the first fortnight, to build a dashboard for this. It was a way of feeling like I was doing something. The honest version of that period is that the only measurement that mattered happened once, in a radiology room, at Week 6.
The only metric that decided anything
The Week 6 X-ray, on 3 June, showed bridging callus on the medial cortex. On that single image I went from non-weight-bearing straight to 50% partial weight-bearing, skipping the textbook 25% step entirely. Nothing I had tracked for six weeks contributed to that decision. The radiograph did all of it.
That is the cleanest lesson in this whole recovery and it cuts against the instinct that got me into self-quantification in the first place. In rehab, the gates are radiographic and clinical. Your daily metrics do not open them. What the daily metrics are for is the space between gates — making sure you arrive at the next X-ray having done the work, and catching the days you are doing too much.
The Week 12 review, on 13 July, read the healing as on par and cleared full activity: swimming, cycling, longer walks. Light jogging was projected for somewhere around Week 16.
The jog that arrived four weeks early
It arrived the next day. Four minutes on a treadmill, 0.4 km, at Week 12 rather than Week 16. I added a minute a day up to seven minutes, then stopped for two weeks — which in hindsight was the single best decision in the whole rehab, and it came from feel rather than from a number. Running returned properly in August: 1.8 km on the 3rd, then 3.65 km continuous in 25 minutes on the 7th.
Being four weeks ahead of a surgeon's projection sounds like a triumph and I have tried hard not to treat it as one. A projection is a median, my fracture pattern is not the median, and the reason I could jog at Week 12 is that the bone was ready, not that I was clever. The data did not accelerate anything. It just told me when to stop.
Which numbers earned their place
Once training reopened, CTL — chronic training load, the rolling measure of accumulated fitness — became the number I actually used. It bottomed at 14 against a pre-crash peak of 70.7, and by early August it was back to 31. That metric was decision-useful in a way almost nothing else was, because it gave a single honest answer to the question I kept asking badly: how far back am I, really? Not far enough to be racing. Far enough to be building.
The metrics that were noise: daily weight, which swung 2 kg on travel and hydration and told me nothing about a bone. Step count, which measured how much errand-running I had done rather than anything about healing. And any single-session pace or heart-rate figure during the first month back, all of which reflected deconditioning rather than the leg.
The metric I underrated: subjective pain on a fully extended leg. Long after walking looked normal from the outside, a fully extended stance was still uncomfortable, and that lagging symptom tracked the actual state of the leg better than anything I could export from a watch. The same trap as measuring a limb with a tape and calling the muscle intact — see the muscle-preservation piece for that version of it.
What I changed because of it
Two things. I stopped tracking anything during the non-weight-bearing phase that did not feed a decision, which removed a daily ritual that was mostly anxiety management. And I now treat any metric that cannot be acted on within the week as a candidate for the kill list.
The GLP-1 protocol restarted in early June, once the bone was far enough along that a modest deficit was no longer a risk to it — the full timeline of that decision is on the GLP-1 hub. Challenge Roth 2027 is the named comeback target. Back to the Biohacking hub.
Ready to Train Smarter?
Get personalized training zones, race predictions, and performance insights with our free calculators.
