I have more nerdy follow-up on how to lose 30 lbs the not-impossibly-hard way.

I've been keeping a log of what I ate and how much I weighed every day for more than 4 months now. And I already mentioned that I like data and charts and understanding how things work. I'm going to see if I can figure out when calories turn into weight.

The Data

I have weigh-in data for 105 days. More importantly, I have 91 pairs of weigh-ins collected 24 hours apart. So I want to see if there's a lag between calorie intake and weight loss (or gain).

So I looked at 1, 2, and 3 days of lag time to see which is the best predictor of weight, here's what I got:

The 1-day lag has the strongest predictive power for my weight. That's pretty intuitive: the weight comes from the calories I eat between weigh-ins. 2-day lag has about 1/3 the predictive power, and 3-day lag is actually slightly negative.

So that seems clear enough: You gain the weight the same day you eat the calories.

But Wait...

Let's try something else. What if we look at which days of the week I lose the most weight. Here's a plot of average weight gained (or lost) by day of week throughout the 4 months.

Check this out: I only lose weight Tuesday through Saturday. Sunday and Monday, I tend to gain a little weight.

What's going on there? Is it just calories?

Here's a plot of the calories eaten between weigh-ins, grouped by day of week.

Kind of a mess, right?

But there's something there... Check out the graph plotted along with the 2-day lag (the one with less predictive power):

When I look at it this way, I get a different answer: calories take about 2 days to turn into weight. Calories I ate on Friday seem to show up Sunday.

So What's Going On?

It looks like there's something else in my weekly routine that drives weight loss/gain, and weight lags behind it.

My best theory: Bulleit Rye Whiskey and Stone IPA.

Alcohol is loaded with calories (like 200 per Stone IPA). If I had alcohol, it was most likely on a Friday or Saturday. The extra calories slowed my weight loss, but alcohol dehydrates me, so it looks like I'm losing even more weight. That accounts for the poor fit between Saturday's weigh-in and Friday's calories. By Sunday and Monday, I'm back to normal hydration and it shows up as gaining a little weight.

It's also worth mentioning that we're dealing with small amounts of weight and a pretty narrow range of calories. About 2/3 of these days I ate between 1,200 and 1,500 calories, and the range of weights is similarly small: 2/3 of the time, weight loss was between -1.6 and +0.4 lbs per day.

So that's my theory: I gain the weight the day I eat the calories, but if the calories were from alcohol, then the weight is hidden for a day.