Sixty squats logged. Forty left. The meeting ended at 9:18pm.

Marcus had been running an early-stage startup for fourteen months — a small team, a closing seed round, the particular flavor of founder schedule where any hour of the day could be eaten by something with a forty-eight-hour fuse. He had been an athlete in college, not a serious one but a consistent one, and the body he had carried into his late twenties had quietly drifted by his early thirties into something that sat through Zoom calls for ten hours a day and walked the four blocks between the office and the BART station and called it activity. He was not out of shape in any dramatic sense. He was out of the specific kind of shape that comes from asking your legs to do hard work on a regular schedule, and he could feel the absence in the way his lower back complained after long working sessions and the way stairs registered as effort instead of as nothing. He had read about squats as the highest-yield bodyweight movement for someone who could not commit to a gym routine — full posterior chain, no equipment, scalable by reps rather than load. A hundred a day was a number he could say out loud without flinching. Distributed across the day in sets of ten or twenty, it was the kind of target that fit between calls instead of competing with them. He had tried to build the habit twice in the previous year without external structure. Both attempts had collapsed inside ten days, not because he forgot but because there was always one more meeting that ate the window and no cost to letting the window close.

The first fourteen days ran cleanly. Marcus distributed the squats across natural breaks: 20 reps before his 9am standup, 20 after his 11am one-on-one, 20 in the bathroom stall on the third floor of the WeWork at 1:30pm, 20 when he got home around 7pm, the final 20 before dinner. The distribution had slack in it — on days when a meeting ran long he could compress to four sets of 25, and on one Saturday with no calls he finished by 2pm and walked the rest of the day knowing the contract was already settled. The pose detection became something he stopped thinking about. The phone went on the floor. The counter ticked. He moved to the next thing. By day fourteen he had logged fourteen consecutive days at 100 reps or higher, and the practice had stopped feeling like a contract and started feeling like a thing he did, which was the transition he had been waiting for through both of the previous failed attempts. Day fifteen was a Tuesday. The morning had a 5:30pm scheduled call with a prospective lead investor — a partner at a fund that had been circling the round for three weeks without committing. The call was on the calendar for forty-five minutes. By 5pm Marcus had logged 60 reps: 20 before the standup, 20 after the one-on-one, 20 in the bathroom at 2pm. He had planned the remaining 40 around the 5:30pm call: 20 immediately after, when he expected to be home by 6:30pm, and the final 20 before dinner. 5:30pm. The call started on time. The partner had brought two associates Marcus had not been told about. The conversation went deeper into the unit economics than any prior meeting had — a sign the firm was getting serious, which was good, and a sign the meeting would not end at 6:15pm, which was less good. By 7pm the partner was still asking questions. By 8pm Marcus's co-founder had joined the call from his apartment to walk through the technical roadmap. By 8:30pm Marcus was aware, in the back of his mind, that he had ninety minutes until the deadline and was on a call he could not leave. By 8:45pm the partner said she had ten more questions. The questions took until 9:18pm. The call ended. Marcus closed the laptop. The Lockin notification was already on his phone: deadline passed. 60 reps logged. Forfeit registered. He had not done a single squat between 2pm and 9:18pm — seven hours and eighteen minutes — and the deadline had closed during a conversation he was not allowed to walk out of for any reason involving knee angles.

Marcus reset the contract on Wednesday with a single structural change. On any day with a meeting scheduled after 4pm — investor calls, board prep, customer demos that had a tendency to overrun — the first 80 reps had to be logged before the meeting started, not distributed around it. The remaining 20 could float. The rule was front-load on uncertain days, not float on uncertain days, and uncertainty was defined operationally as any external party on the calendar past 4pm whose meeting end time he did not control. The rule held. On Thursday he had a 4:30pm customer call and logged 80 reps by 4:15pm — three sets of 20 in the morning and a set of 20 at 3:50pm in the office bathroom, which he registered as slightly absurd and did anyway. The call ran to 6:10pm. He did the final 20 reps at 6:30pm before dinner. The deadline was not a factor. On Friday a board prep slot at 5pm slipped to 6:30pm and ran to 8:15pm. He had logged 100 by 4:45pm. He did not look at the deadline once that evening because the deadline was no longer a question. The distributed-across-the-day method worked exactly as designed when the day was predictable. It failed when an external meeting could metastasize into the deadline window without warning, and the failure was not the meeting's fault. It was a planning error rooted in treating the morning as a normal morning when the calendar already said the evening was at risk. The deadline was the only variable the method could not protect itself against, and the morning was the only place the deadline could be neutralized in advance.