An error of measurement is always a possibility. Unfortunately, I can't travel back in time to measure my weight from before two weeks again.

To add more context, at this moment I am quite overweight (most that I've been at least in the recent ten years), because in recent months I had a lot of stress and almost no opportunity to exercise, plus an access to a lot of great food (I am really bad at resisting temptation: my only strategy for not overeating is "don't buy it"). Generally, my behavior depends a lot on my environment, and most of my self-control and self-improvement is based on strategically modifying my environment; and I had almost zero control over my environment until recently when we moved to a new flat. Now every morning I do sit-ups and five different exercises with a 10-kg dumbbell, most of them until I am too tired to continue. And half of the days my diet consisted of some joylent (about 10-20% of daily intake) and vegetable soup, and maybe a slice of bread.

I attribute this change to the "beginner's luck", which means that if you haven't been doing anything and you suddenly start working hard, you usually get some huge progress at the beginning and then it slows down. I expect that doing exactly the same thing for the next two weeks will have much smaller effects. (Unfortunately, this also makes the "error in measurement" hypothesis unfalsifiable.)

Anyway, I congratulate myself for the willpower to exercise regularly, regardless of whether the outcomes were measured correctly. That was an important hypothesis I needed to test experimentally -- whether my behavior will really change in the new environment, or whether I am just rationalizing my laziness.