The Math Diet

This page discusses the method for feedback weight control that I have described in my IFAC Blog article.

My method has received substantial media attention and I have received a huge number of questions. Below, I will try to describe the method step-by-step and to answer some of the many questions.

Here is an overview of the method:

  1. Start by recording your body weight an evening after the last meal of the day. This weight is your initial point, e.g. 83.0 kg.
  2. Select an end target and a number of days to reach that target, e.g. 80.0 kg and 30 days.
  3. Compute the corresponding daily weight loss, e.g. 3 kg = 3000 g in 30 days corresponds to 100 g per day.
  4. Compute for every day the predefined daily target after your weight loss, e.g.
    • 1st evening: 83.0 kg (initial measurement)
    • 2nd evening: 82.9 kg (100 g less)
    • 3rd evening: 82.8 kg
    • 4th evening: 82.7 kg
    • 30th evening: 80.1 kg
    • 31st evening: 80.0 kg (end target)
  5. Now, the method commences! You have to measure your weight in the morning, e.g. 81.8 kg in the morning of Day 2. That means that you can eat at least 82.9-81.8 = 1.1 kg during Day 2.
  6. You now have to weigh your meals. As an example, you can plan to eat 1.1 kg food as 300 g breakfast, 300 g lunch and 500 g dinner.
  7. During the day, however, you will lose more weight. Therefore, you have to measure your weight again prior to the last meal of the day. Assuming that your weight at that time is 82.2 kg at Day 2. That means that you have to eat exactly 82.9-82.2 kg = 700 g for the last meal of the day (instead of the 500 g estimated in the previous step). Right after this meal your weight have to equal the daily target of the day precisely, e.g. 82.9 kg on Day 2.
  8. The following days you have to do exactly the same:
    • Measure your weight in the morning
    • Subtract the morning weight from the day’s target
    • Plan the meals of the day based on the difference
    • Measure yourself again in the evening just before the last meal
    • Subtract the evening weight from the day’s target
    • Eat exactly this difference during the last meal
    • Now, you should have reached the daily target precisely!
  9. After the scheduled number of days, your weight should be exactly equal to your end target weight! It is important to reach your daily target each day, i.e. to be neither below nor above the predefined daily target.
  10. When you have reached your end target, the daily target stays constant, e.g. 80.0 kg. The method proceeds exactly like before, but now with this constant daily target.
The figure shows the results from my pilot experiment. The first days I lost a little too much. Later (Day 12), I once had a little too much. Otherwise, I hit the daily target precisely by weighing my meals and following the steps described above.


Several readers have alerted me that a method, which in several ways is similar to mine, has been published in:

John Walker: The Hackers’s Diet: How to lose weight and hair through stress and poor nutrition, 1991.

This is a very well-written reference, that I would strongly recommend!

What about liquids?

I get a lot of questions concerning liquids. In principle, all liquids have to be weighed in, and that is what I did during my pilot experiment. In practice, however, this is not crucial, as you quickly secrete roughly the same amount of liquid as you drink. In fact, you secrete a little more liquid than you drink, when you are losing weight. In short, you do not necessarily have to include liquids in the meal weights, as long as you make sure to hit your daily target spot on every day.

What about coke?

Liquid consumption is in principle the same e.g. for water and for coke. If you drink 2l coke a day, you also secrete approx. 2l of water, so that balances out. A coke contains approximately 27 g of sugar, so in principle that should be included, but in terms of weight that is really a marginal number. However, if you drink 2l of cola today, your weight loss tonight is going to be considerable lower than it would have otherwise been, so tomorrow, you can eat much less than if you had drunk 2l of water.

Surely it matters whether you eat fatty or lean food?

First, let me emphasize  that it is always recommendable to eat healthy – also when following my method. If you eat energy rich food, the body extracts less from its reserves and more in direct conversion, than if you eat lean food. That means that next morning, you will weigh less if you had lean food, and more if you had fatty food. Thus, with an energy rich diet you get a smaller ‘margin’ next day to your daily target, and can therefore eat considerably less. With my method, in principle you can eat whatever you like, but if you do not eat in a healthy way, on the long term you won’t get to eat a lot. This is actually motivating to eat in a healthy way. For the sake of science, I tried both during my pilot experiment. The days following an evening with lots of red meat + wine + snacks were quite tough to get through…

How accurate does my bathroom/kitchen scale have to be?

It is not highly important to have an accurate scale. Most bathroom scales have a resolution of +/- 100 g, and that works fine. On a given day this can of course give some variation in the amount you can eat, but that balances out in a few days. I had chosen a daily weight loss of 239 g a day. This loss can of course not be met accurately on a single day unless you have an unusually high resolution on your bathroom scale. I could, however, hit 2.4 kg in ten days, etc. The same applies to the kitchen scale. Potential measurement errors balances out in a couple of days.

For the nerds: in technical science we distinguish between accuracy and precision – please, see figure below. In this context, the important concept is precision. It doesn’t matter a lot that the scale doesn’t show the correct number, as long as it doesn’t vary too much from time to time – otherwise you get unpleasant variations in your amount of food.