Lab Rat: Lose It! and the Bodyweight/Calorie/Exercise Intersection

One year later, we're revisiting some of the aggregated data we've generated via the Lose It! app. See the 2014 write up here. Consider this a 7-year test of whether or not the Lose It app functions as a weight loss instrument for someone who exercises regularly. 
I'll ruin any suspense early: it does.

What we'll be doing differently here is ignoring the type of exercise done (which, Lose It the app doesn't really care about anyway; converting all work into calories expended). If we were doing hypertrophy work, we'd expect some weight gain. If we were doing more endurance work, we'd expect muscle loss, and therefore, weight loss. Maybe. But we're also not going to have any composition data to compare in this set up. Bodyfat up or down? We'll not know.
Optimally, we want to compare Bodyweight/Bodyfat to total food intake, with a macros breakdown, and an additional layer comparing total work volume done; and that, by work type - perhaps broken down by energy system.

Lose It! assumes that the user wants to lose bodyweight, and simply assumes that the less you eat, the less you'll weigh. Which is true; if you eat nothing, soon you'll weigh as much as worm shite since you'll be dead. Lose It does compute work calories by subtracting them from your daily total ingested. So if you work more, you get to eat more that day without going over your daily "limit." So we can examine those three factors while using the app - Bodyweight, Calories ingested, and Calories expended through exercise.

We've got 7 years of tracking, with a two-year hiatus hole in the middle where I didn't do any logging. Let's just look at the graphed data overlays and see what they suggest. First, we compare Bodyweight to Calories ingested:

...and see no real correllation. Late in the graph, we see that where the purple calories are up, weight goes down. Not what one might expect at all, right?
There's those two peaks on the right, where calories dipped and bodyweight dipped correspondingly by about 8-9 pounds, but generally, there's 3 years worth of plateau-ed intake with inexplicable BW change.

Compare the same Bodyweight chart, but now to an overlay of Calories expended during exercise. 

Before the hiatus, we see low BW when calorie expenditure via exercise is high, and a general BW climb when exercise calories are lower. After the hiatus, exercise climbs and BW drops - still what the general consensus would presume. Especially since we aren't comparing intake as well.

So, let's add that. The three factors compared:

 Which is messy. Cut to the chase. Here's where Lose It shines. The app takes the Calories In for the day minus Calories Out and ka-smushes them, into total Calories over or under your daily limit (which is the red line below, each data dot is a monthly average either above or below the limit line).

Very nearly a direct relationship between calories over showing as bodyweight gain, and calories under showing as bodyweight loss. Even with exercise type, and any macros breakdown, removed from the equation, the Eat-to-Work-to-Bodyweight math works within the limits of the app.

If you'll notice, there's also a month of lag-time along the rise and fall of the graph - maybe I have to undereat for a couple of months before my body decides it's going to be losing weight. The same happens when gaining. A visual reminder to me: whether cutting or building, have patience in the programming.