Final Guest-Blog From Seth Roberts

Here is the final post from our guest blogger, Seth Roberts. If you need to get up to speed on Seth’s unorthodox research with weight-loss, mood, acne, and sleep, click here (our N.Y. Times article about him), here (research extras and pix), here (the first round of reader comments), and here, here, here, here, and here for his earlier blog postings.

“Thank You”
Sun., Sept. 18

The highlight for me was the acne discussion. So much interest! And all of a sudden, there it was: considerable support for the idea that acne is caused by certain foods. Plus at least one person was able to figure out what those foods were: “For nearly a year, I diligently experimented with different foods/exposures to test their effect on my face [and have] identified about 8-10 seemingly unrelated food triggers … As long as I avoid these triggers, my face is clear. It’s the damnedest list that includes such seemingly innocuous foods as spinach and oatmeal.”

I am happy to help others figure out exactly how he generated his list. His method, which would surely be free, safe, and available to anyone, would help millions, if not tens of millions. I hope to write a book about the use of self-experimentation to solve everyday problems other than obesity, such as acne. The value and underlying logic of such a book could hardly be better put than what one commentator wrote:

“Your approach, OTOH, is setting up experiments (which may derive from the physician- promulgated information) in the universe over which you have control, and to which you have access. The results are highly relevant to that universe. At that point, the information is released so we ‘other universes’ can experiment toward replication to test not the validity of your experiments, but the size and nature of the universe to which such results apply. Now that’s science, in a new democratic variation of the paradigm.”

The acne discussion convinced me that there is room on the web for a site devoted to the use of self-experimentation to solve everyday problems. Maybe there is already an acne site — but a) people interested in doing self-experimentation — perhaps a tiny minority of all acne sufferers because, at least in the beginning, self-experimentation to find non-drug solutions is surely long and complex with uncertain outcome — might benefit from a site more devoted to them; and b) those using self-experimentation to alleviate acne might benefit from contact with people using self-experimentation to solve other problems. Leonard Syme, a Berkeley public health professor, once told me that NIH was poorly organized: Division into institutes by organ and disease (Eye Institute, Cancer Institute, Institute of Mental Health) made little sense if, as was surely the case, different diseases and diseases in different organs had common causes. I hope there is eventually such a site and when there is you will be able to find it by going to Dubner was surprised I did not have a web page. Why? he asked. No good reason, I said.

You will be able to learn more about my weight-loss ideas and their application by going to So far, people in publishing seem to like the title The Shangri-La Diet so that may actually be the title of the weight-loss book I hope to write. I am especially interested in hearing from people who have had trouble finding success with the diet because there is surely room for improvement; and, being human, I would like to hear from people whom it has helped.

Finally, I want to thank Andrew Gelman and Alex Tabarrok who together brought my long eccentric paper to Dubner’s attention. I know Andrew from Berkeley, where we taught a freshman seminar on left-handedness. A reporter for the San Francisco Chronicle visited our class one day, presumably to write about it; and then wrote nothing. I have never met Alex. After Dubner read my paper, he told Levitt about it. “I thought he wouldn’t like it,” Dubner said, “but he did.” Maybe that’s why they’re friends: They like the same things.


Hello friends:
I'm discouraged that though I've been trying the diet, I'm still very often hungry between (regular sized)meals. I believe I'm doing the right thing (with the "dosages" and the 1-hour wait) but whether I do the oil or the sugar water, the hunger does not go away. In fact, this morning I did BOTH the water and the oil out of desparation and still felt probably as hungry as I would have been if I ate nothing.

I'll keep trying for a bit though. I've been tracking food and appetite and may notice something. And, if nothing else, I'll learn something about myself.

Any one else hungry for decreased appetite?


seth roberts

you can send weight-loss questions to shangri la diet @ (which should be spelled without spaces--added to avoid spam). I am happy to try to answer "more serious questions about my methods" here. If anyone is still reading this.

Dennis C. During

I thought that the method Prof. Roberts is recommending is that people systematically test for themselves what works for themselves. Many commentators here seem to expect to be able to simply take the conclusions of a series of self-experiments by somoeone else and directly apply them to themselves.

Doctors and psychologists routinely recommend some kind of vague self-monitoring, without offering any specific suggestions or support for the patient's self-monitoring efforts. Prof. Roberts' approach points in the direction of providing such suggestions and support.

If we wait for big-science studies of most products and services, that means we are waiting for studies that are provided by almost entirely by enterprises with a patent- or trademark-protected monopoly on what is being tested.



I tried fructose water for 3-4 days and GAINED THREE POUNDS!!! 45-60 minutes after drinking the sugar water I WAS RAVENOUS!!!

I'm now using oil supplements in gelcaps and fiber supplements for the non-flavor calories. Two days so far, 1200-1500 net calories (counting the oil) and no weight loss. I am not actively hungry anymore after the oil, but nor do I find my appetite suppressed in the least when I do officially eat. So for this plan to work, it had better be affecting my set point behind the scenes.

(Also, BTW I can barely drag myself out of bed in the mornings, since I began this experiment.)

I am willing to try suffering the hunger pangs and waiting longer after the sugar before eating-- because it's been reported elsewhere that (little known fact) hunger pangs disappear after about an hour. If I can get beyond that point, and not have to eat while feeling hungry I may be able to do better. But let's see if the oil helps any.

BTW based on "Seth's friend Tim" it took him a very long time to lose the weight and much of it was made possible by essentially giving up the idea of getting any pleasure at all from food. Maybe some of us have to do this, like alcoholics can never touch a drop. But as pathological as some people's relationships to food can be, isn't the idea of choking down mush just as dysfunctional? Is taste celibacy worth it? Can such an approach be a long-term strategy?



To all the self-righteous contributors who admonish us with "just experiment on yourself" instead of asking for a formula...

It is a valid point. It took much trial and error for me to find out that sunscreen chemicals are very bad for my skin and that mineral sunscreens are fine.

But... Roberts spent 15 years experimenting on himself. I don't have 15 years -- I'm 38 with two failed IVFs (ending in miscarriage) behind me and under a stern warning to reduce my BMI to a specific number before trying again. I'll spare you the math, but I have to find a way to live on very few calories -- and increasingly fewer calories for the rest of my life IF I lose weight. (Surgery isn't an option due to the waiting period before attempting pregnancy.)

None of this, or course, is Seth Roberts' lookout. But it's cruel to tease people like he has done.

Also during his 15 years, (acc. to the NYT piece) Seth found several strategies that DIDN'T work before zeroing in on the sugar. Wouldn't it be nice to know what those failures were? Especially if there are mistakes one can make that will DOOM the entire approach.

The point of self-experimentation is of course to help yourself but hey, it was Seth who decided to put himself out there.

And he's made it clear that he INTENDS to provide specific help in the form of a diet book, so he isn't about making everyone spend the next 15 years repeating his experiments. He's about dollar signs. Which is beautiful and all-American, but WHEN WILL THE BOOK BE AVAILABLE?!?

Those of us eager to start in the meantime need something to go on.



your site is short on knowledge and long on links to advertising (and i suspect spyware). You are obviously trying to make $$$. No problem, but i wish you would have more content. When the number of hits to your site decease, you won't need a pseudo-scientific study to figure ou why. I just told you.


I've lost about 3 pounds in ten days. Not earth shattering, but slow and steady is good I think. I tried fructose, sucrose, canola oil, and light olive oil. The oils seem to work better for me. The first few days I barely felt any hunger, but now I get hungry. The hour waiting period after ingesting the oil does slow me down. I find I'm more aware of the difference between real hunger and cravings or boredom eating. And once I start eating, I do seem to fill up faster. So I'm eating smaller portions. Still, I'm taking in 360 calories per day in oil, and at my target of 1200 calories per day (I'm a woman btw)there's not much room for tasty satisfying foods. Speaking of gender, I'm wondering if Dr. Roberts has had any women try his diet. I think women's set points might be sturdier and harder to budge. Our bodies seem to want to hang onto fat.


>>Still, I'm taking in 360 calories per day in oil, and at my target of 1200 calories per day (I'm a woman btw)there's not much room for tasty satisfying foods.<<

I think that's the entire point. If it tastes good, it's fattening. If you're so unfortunate as to be predisposed to obesity, you just can't fix that and continue to enjoy food on a regular basis.

Note that any heavily marketed diet does the opposite. It tries to find a way for you to somehow stop ingesting all these calories without suffering. Ornish: eat more, weigh less. Just no fat. Atkins: enjoy rich satsifying foods. Just no carbs. Agatston (South Beach): No need to cut carbs or fat, just make sure you only eat the correct ones.

The truth of this diet came to light when "Tim" disclosed on the previous page just how he lost the weight. Food is basically over for him. If he could, he'd get a feeding tube inserted.

This would be refreshingly honest if Roberts had been, well, honest about it.

BTW 360 calories in oil seems like twice as much as the other people seem to be talking about. What happened to 140 calories per day? How did you arrive at 360?



I've been taking 3 tablespoons canola oil one hour before meals @ 120 calories per. That's more than originally suggested, but I increased the amount from 2 to 3 when my appetite increased about a week into the diet. I'm going to eliminate the pre-breakfast "dose" today and see how I do.

I found another site online ( that interviewed Roberts and here's what he eats per day on "maintenance":

One normal meal (about 900 calories)
Two pieces of fruit (about 75 calories apiece)
Sugar water (150 calories)

He drank more sugar water when he was actively losing weight.

Sounds like he fasts all day, controlling his hunger pangs with sugar water, and then eats a satisfying dinner. I could live with that. If I could stay with it. And there's the rub.



So where, in your understanding, does the set point come in?

The NYTimes piece implies that ingesting unflavored calories CHANGES the set point. My prior idea of the set point was that the body tries to hang onto the weight by changing the metabolism itself in response to changes in the food supply. That's why only the most radical diet will work: when you cut calories, the body burns fewer calories due to the "famine." You can somewhat counteract this by building muscle mass (which burns more calories) and paying attention to the glycemic index/load. But essentially if you're already obese you're doomed to an unnaturally small caloric intake for the rest of your life.

But basically in this interpretation the set point is controlling the *metabolism*, not the appetite. I always viewed the appetite as a separate thing, either caused by genetic issues (like Leptin receptivity problems) or eating disorder.

But what I seem to be gleaning from Seth's paper and the blogged comments is the idea that sugar/oil is suppressing the appetite. In this case, either the set point is irrelevant, or the set point controls the appetite, not the metabolism.

What do people think about this?

The good news would be that, if only we can get ourselves to stick to the caloric allowance, the weight loss WILL be foolproof (in the absence of a metabolic issue such as thyroid disease). The fact that people DO lose the weight when FORECED (via surgery) to do so lends weight to this interpretation.

The bad side to this is: what if the desire to eat is not a function of physical appetite but a symptom of a psychological issue (IOW an eating disorder)? In this case, these physical techniques to suppress the appetite will be no match for the eating disorder and will do more harm than good by adding 300 more empty calories to the person's intake. To truly suppress the appetite of someone with an eating disorder you have to actively make eating unpleasant.



As I understand it, eating tasteless calories signals our caveperson brains that the food supply is dwindling and food is scarce. In response, the brain signals the appetite, and tells it to slow down, so we won't be ravenous now that there's very little food. So we eat less. Makes sense. That's the appetite suppression part.

The set point apparently isn't (so much) affected by weight or metabolism, but rather by flavor- calorie associations. The higher the association, the higher the set point. Which means that the set point isn't the “writ in stone” number we usually think of, but rather, it fluctuates, depending on what we eat. And eating foods with low flavor-calorie associations, like sugar water and olive oil, drops the set point. I'm not sure why this happens – my eyes start to glaze over around this part of Roberts' academic paper – but I think that's the gist of it.

My personal dieting experience hewed to the set point being immutable. I could diet away 15 pounds with a lot of discipline, but beyond that, forget it. Getting below that would require superhuman effort. And there's no way I could sustain that. I'm REALLY hoping that following this diet will adjust that number downward. By another 15 pounds I hope. Won't know until I get there.

I'm going to switch from this blog and start frequenting Ann Hendricks blog (see above) for postings about the diet. Hope to see you there.



"But as a person who has had rosacea for the past 8 years or so, I have to tell you: Rosacea is a kind of acne. 'Acne Rosacea' they call it."

We know that. But it's a condition known to be sensitive to personal triggers, and standard advice is for a patient to self-experiment to see what triggers rosacea for him/her.


I also am confused as to whether the diet influences metabolism or only appetite. Set point normally is used to describe the weight one stablizes at, with implications for metabolism. Roberts appears to be using it to describe the weight one stablizes at, with implications for appetite (and reduced metabolism, and reduced number of calories one can eat, once one's lowered the set point with the diet). The presentation is confusing on this score.


I'm it sugar water or oil, or both? How much and when?


Experimentation is, of course, the foundation of all science (knowledge), so there is a reasonable tradition here. What is useful is to run experiments on groups of unbiased people to confirm results. I did that with Frutels, a new vitamin/mineral product which I created for myself. Then tested in about 100 kids in Manhattan. Then had independent testing in another 110 or so. Got better than 75% positive results. Why? Because I am a trained epidemiologist who is a follower of Linus Pauling, and understand that acne is an internal problem. Frutels(www.frutels.com0 are in the form of a gummy (no sugar), taste pretty powerful - because they are. Right now available on the web only. Our findings are that there are a few general reasons why people get acne and their bodies need additional support in those areas, e.g.; adolescents get it in part because of hormonal change. So Frutels includes a patent pending formula to address those causes. 1% of Frutels profits are donated to the Linus Pauling Institute to support its research in the use of vitamins and minerals to address disease. Perhaps this is the website Seth is looking for.



This post has been removed by a blog administrator.


This post has been removed by a blog administrator.


This post has been removed by a blog administrator.


A disappointing series that seems aimed at self-promotion. Roberts comes off as an expert teaser, accomplishing little more than feeding off his 15 minutes of fame.

Andrew Gelman

I wouldn't call Seth a "teaser." There's tons of information on his self-experiments in his Behavioral and Brain Sciences article (see here). This represents something like 15 years of research, so I think "15 minutes of fame" is taking the short view of things. There are some large issues about how to conduct exploratory scientific studies (and there's room for debate on these issues). My impression is that Seth was aiming for a conversational, impressionistic tone in his blog entries. You can look to his scientific papers for more precision.