The Efficient Frontier of Fat Loss
Run bold experiments to cast a wide net and then narrow it down
tl;dr:
There are very extreme interventions that make you lose a lot of fat
A lot of them would probably work (nearly as well) in much less severe versions
The problem is we don’t know which ones and less severe how
Why would we restrict ourselves more than actually gives us benefits?
Let’s find out where the frontier lies.. by bumping into it
Blatantly stolen from investing: Efficient Frontier
The Efficient Frontier is a neat concept from investing. It describes how, in order to get higher returns, you often have to take on higher risk. Read up on it on Investopedia. Its merits in investing aren’t the point here: I think we can use it in all kinds of fields with uncertainty and not entirely obvious feedback loops. Especially when experimenting in such a confusing and unclear field as fat loss.
As an example you could put all your money into a very safe investment. You probably won’t lose it but you probably won’t make a great return. Nobody pays you to be safe.
On the other hand you could invest all your money into some crazy scheme that is extremely risky. You will likely lose it all. But if it pays off, you could get rich!
Of course this is a simplification. There are different types of risks or returns. Our estimation or perception of the risk or return could be wrong.
Maybe we think housing is super safe and then 2008 happens. Or maybe we think 30-year bonds are super safe and then… now happens.
In addition, different people have different preferences and reactions to certain types of risks. So it’s very conceivable that what seems “perfectly safe” to one person seems extremely unsafe to someone else.
It’s a simplification but a useful one.
Don’t make trade-offs when you don’t have to
The main lesson of the Efficient Frontier is that there are some trade-offs you have to make and some you don’t. Why would you make the ones you don’t have to?
As an example: if you do keto you’re restricting your food intake in a certain way. You limit foods that contain carbohydrates. You could also apply further restrictions and eat only meat or other animal products: that would be carnivore. But I would argue that you better have a reason for the extra restriction. Maybe the reason is just that it sounds badass to only eat meat - I’ll agree that it does - but I’d hope there’s more to it than just that. Maybe there are some plant ingredients that cause auto-immune issues for you and cutting out all plants fixes that.
But you could now be making a trade-off you don’t have to. Say your issues are caused by nightshades. You could just cut out all nightshades and eat other low-carb plants. This way you’d fix both the auto-immune issue and be able to eat certain other plants. (Yea, I get it, some people don’t want to eat plants.)
Similarly for keto: many people don’t actually get amazing benefits from ketosis, at least in the short term or absent certain conditions. If you’re merely doing it for fat loss you could probably achieve the same rate with way fewer restrictions.
For example: “dirty keto” (what the average keto-dieter on Reddit does) is nothing like the original, medicinal ketogenic diet. The original diet is so restrictive that it’s nearly unpalatable to almost anyone and they had trouble getting children to adhere to it even when breaking the diet caused them to go into epileptic seizures again.
For comparison my current diet is 88% calories from fat, consists mostly of heavy cream, only has 30g of protein per day (which people on Twitter are calling me crazy for), and is not fatty enough to qualify for the original “ketogenic” level. It would have to be 90%.
Almost nobody wants to endure restrictions this severe and most people who do “keto” eat ad libitum meat, vegetables, often nuts, some fruits - all things that a strict “ketogenic diet” wouldn’t allow. But most people don’t need to make that trade-off because they’re not epileptic.
And if you can still get decent fat loss, albeit maybe a little slower, without nearly restricting yourself that much, why wouldn’t you?
That’s why I think the new generation of weight loss drugs like Semaglutide and Tirzepatide are so interesting to people: very good weight loss results with almost no restrictions? Who wouldn’t want that? Personally I’d hold off because these seem very new and the data seems split as of yet. But if they work as promised that’s certainly a very good profile on the Efficient Frontier.
Creating the Efficient Frontier: trial & error
If we knew all the mechanisms behind obesity and fat loss we could neatly calculate or simulate the Efficient Frontier precisely and simply pick our favorite intervention or diet from among it.
Unfortunately we don’t know what makes us lose fat, even when we do lose fat. (Incidentally, if you want to feel REALLY stupid, I recommend you do something successfully without knowing how.) Some things work for some people but not for others.
The only method we have is trying things.
Trying things is famously difficult. The different factors (e.g. ingredients) in nutrition are so mixed and tangled up that it’s very hard to attribute an effect to any particular change. The feedback loop is so slow and noisy that it’s difficult to infer any effect at all in less than 30 days and even then it’s very tricky.
And the results don’t seem to reproduce super well in other people. This might of course be a function of diets being so complex they’re hard to describe.
I talked to a friend once who proudly told me he’d already lost 5lbs on a “calorie restricted diet.” When I asked how exactly he’d done it he said he’d cut out all bread.
“So it’s a low-carb diet?” I asked.
“No, I already told you, I’m just cutting calories.”
“But all the calories you’re cutting are carbs?”
“Yes.”
This cognitive dissonance didn’t seem to faze him at all. An example like this is obvious to anyone who’s ever done who’s done a variety of diets.
Was it the calories?
Was it the carbs?
Was it bread specifically?
What sort of bread was he eating before?
Was the bread baked with PUFAs or butter?
Did the bread contain a lot of sugar?
Did he cut out his entire breakfast and accidentally start intermittent fasting?
Depending on what my friend’s intention at the time of the diet was he would describe the reason for his success differently. But we simply don’t know. We’d have to run a 14-30 day trial to test each of those hypotheses. And that’s if his 5lbs weight loss is even a real signal at all and not just natural fluctation of water weight.
It’s very difficult to change exactly one variable in a diet. For example, on ex150, I am doing the following:
Cutting out nearly all carbs
Cutting out nearly all polyunsaturated fat
Eating very low protein
Cutting out nearly all plants
Cutting out nearly all fiber
Eating extremely high saturated fat
Eating very high monounsaturated fat
Eating very high oleic/palmitic acid
Eating somewhat high stearic acid
Eating extremely high dairy
Eating pretty high whey (I think?)
Eating a very low-variety diet
Eating most calories from moderate-at-best palatability food (cream)
Separating macronutrients in meals (no high-carb-high-fat-high-protein meal)
Eating mostly “whole” and “unprocessed” food (for some definition)
Eating mostly “animal based” (except some vegetables and tomato sauce)
Eating a testosterone-boosting diet (my testosterone tested over 1,000 on this diet)
There are probably a lot more hypotheses that this could be attributed to that I just haven’t thought of yet.
The only way I see is eliminating one hypothesis at a time. Unfortunately the scientific method doesn’t allow us to confirm anything. We can only disprove. So it’s going to be a little bit of a slog.
How to place bets on the Diet Landscape
Let’s imagine you’re a fellow diet experimenter looking to quit obesity. You like the idea of trying things and seeing what works. You just finished 30 days of eating less and you did lose a little bit of weight, say 3lbs.
What could your next experiment be? You could expand in a similar direction as you’ve just done. You imagine that cutting out snacks will cut a little more calories than just lowering portion sizes and it might also be less annoying because you won’t have to explain it to the family at the dinner table, which you would with smaller portion sizes.
Now I suggest you don’t do this. Yes, you can take baby steps and slowly move into the direction of the top left (high fat loss/low severity) until you bump into the Efficient Frontier.
But it will probably take you years or decades until you find something truly successful. Meanwhile the 3lbs you lost might come back because you hadn’t even realized your caloric reduction wasn’t maintainable!
Fat Loss favors the Bold
This might not be the advice some people give. Trust me I’d be OK with slow & steady fat loss if we had a known way of achieving it. In fact, if we had such a way, nobody would’ve slowy & steadily gotten fat, right?
But we don’t and we need to experiment and find something that works. That’s why I suggest you try bold experiments and then cut back in severity to see how far off the Efficient Frontier you were.
This way you’re more likely to find a meaningful effect size (rapid fat loss) in a shorter period of time. You can always relax it later and see if it stops working.
Slime/Time Mold Potato Strategy
Consider the fat loss experiment history of Slime Mold Time Mold. Their very first trial was a 100% potato diet - totally wild & out there. No “reasonable” nutritionist would suggest this as the first step you try toward weight loss. But they got a massive effect size with the average person losing over 10lbs during the 30 day intervention.
One hypothesis they had: maybe the high potassium content in potatoes somehow contributes to the massive weight loss effect. It is a very different thing to supplement some potassium every day vs. eating 100% of your meals as potatoes. The severity is extremely low because all you do is sprinkle some slightly different (weird tasting) salt over your regular food. Some people chug a glass of potassium (salt) water in the morning Unfortunately, this had a much smaller effect size than the original potato diet. But now they had two points on the dietary intervention landscape.
The new trial that Slime Mold Time Mold are running is Half-Potato and is starting right now. You can sign up for it! Tell them I sent you, haha.
Half-Potato means about half the severity of full-potato. Potentially less in terms of social awkwardness because you can eat the potato at home and eat normally with friends. Of course we don’t know yet what the effect size will be because the trial hasn’t finished yet. But it’s reasonable to expect (and they seem to expect something like this) that it’ll have about half the effect size of full-on potato or maybe even better.
So you can see how they’re triangulating their fire and walking in the shots:
100% potato works great (for many) but is super severe
Potassium works so-so but is super easy to do
Half-potato will hopefully show them how close to the Efficient Frontier they are
I presume that if half-potato works great they’ll try another “secret ingredient” that potatoes have. Or maybe a quarter/three-quarters potato diet. Or potatoes-with-butter.
ExFatLoss Strategy
My strategy was inspired by Slime Mold and Time Mold, of course. So I also picked an insane first intervention. Low-protein (REEE!), practically-zero-carb, almost all calories from heavy cream! Nuts, right? Except fattier.
But I also got an insane effect size of 20lbs in the first month. Of course much of that was water weight and the real fat loss was probably closer to 10lbs. But since I ran the experiment for 30 days I could easily distinguish the two different rates of change and knew I had a winner.
As you can see in the pretty picture I drew above, my path going forward is somewhat different. Instead of placing a confident shot in the low-severity sector hoping to hit a hidden battleship (I actually LOVE this analogy!) I am slowly walking from my starting blast down the severity scale.
I suspect that this difference in strategy comes from the fact that I still have a ton of fat to lose and they don’t. So I’ll hang out in the known-effect-size quadrant for a little bit, thankyouverymuch.
My first ex150 mod was ex150deli, which varied the type of meat I was eating. Then I did ex150choctruffle in which I’m replacing the whipped heavy cream with chocolate truffle (which is just heavy cream heated and emulsified with melted chocolate). It had a bunch of confounders and I lost some weight but not that much, but I also didn’t gain weight.
It is obvious that I am trying to walk from “huge effect size” to the left, trying to find the Efficient Frontier where whatever ex150 turns into (let’s call it exOptimal) is very permissive while maintaining the biggest effect size possible.
Indicated by the zig-zagging arrow in the picture I don’t exactly know how and where this is going to go but I’ll do my best to just scrape down the Efficient Frontier once I hit it.
What's this testosterone boosting diet you're referring to?