I held under 2,000 on five of seven days, and the trend turned back down — adherence was the missing variable, not the estimate.

Last week I said I would stop treating 2,000 as a suggestion, and this week I mostly did. Intake averaged 1,957 kcal/day, under target for the first time in the series, and I held below 2,000 on five of the seven days — not a clean sweep, but the closest I have come. The trend answered almost immediately: it reads 168.4 lb, down 0.7 lb on the week and now 1.0 lb below the 169.4 I started at — which erases the half-pound week four had handed back and sets a new low for the series. After a month of the line drifting sideways, one honest week of staying under the target was enough to bend it down again.

That settles the question week four left open. I could not tell, from a flat trend and a string of weeks over target, whether my 2,000 was too generous or whether I simply was not hitting it. The answer turns out to be the second one: when the intake actually lands below the number, the weight moves the way the arithmetic says it should. The estimate was fine. Adherence was the variable, and it was the one entirely within my control.

I am not going to over-read a single week — a 0.7 lb move on an EWMA filter is real but small, and one week of discipline is a data point, not a habit. But it is the first week the method and the result have agreed without me having to explain away a contradiction, and that is worth more than the number itself. The plan for next week is simply to do it again.