Tuesday, March 27, 2018

Fat tails and fat: Why the government shouldn't be in the business of telling you what to eat


I've ranted before (many times) about how the USDA food pyramid My Plate is wrong and only exists to help the bottom line of the industrial food system. Don't worry I'm not going there again. I'm going to go more fundamental: whether correct or not, the government should not be in the business of telling you what to eat.

I may have mentioned that I've been reading more technical work from Taleb about risk management, and it fits really nicely with the obesity crisis and its correlation with the USDA dietary guidelines.

On a population scale, there was an inflection point right when the first USDA guidelines were released.
Whether the actual guidelines were correct or not is irrelevant here; the key is the magnitude of the cost if they're wrong. When you're affecting the health of the entire world (other countries copy our guidelines), the size of the effect is so large that the risk of being wrong is important even if the risk is small.

Consider this: when you fly in an airplane, there is a very small probability that the plane may crash and everyone on board dies. Over the years, we have managed more and more risk out of air travel. The expectation value for the cost of a crash is very small (probability of crash times cost of crash). And so most people choose to fly because the convenience outweighs the risk.

Using that risk-versus-expectation calculation works for a local system where the probabilities are reasonably well known and the effects are localized: only the people on the plane are at risk.

But consider this scenario: if any plane goes down at any time ever, all people who have ever flown in a plane will die.

Would you fly now?

That's an example of systemic risk or as Taleb likes to put it: the difference between risk and ruin. As time goes, on the probability of some plane going down will approach 1. The expectation value is ruin or death for everyone who ever flew in a plane.

How does this tie in to nutrition guidelines? With the poor state of science (see yesterday's post) and with how guidelines are written by committee, the probability of failure (over the long run) will approach 1. Every school, every hospital, every prison, anyone ever who gets government funding, follows the guidelines. Plus then every doctor through the AMA and every nutritionist through the ADA all prescribe those guidelines. It becomes systemic. The population will follow the government's guidelines. If they are wrong, it will trigger a systemic ruin of the population. Not only is everyone getting sicker, the costs of treating the disease is a significant proportion of our economy - a truly systemic risk.

Even if everyone involved were well meaning and wanted to get the guidelines right, there is only one right answer and infinitely many wrong answers. The probability of getting it right is still vanishingly small. (This is how entropy works too; there's only one way for a glass to be not broken, but an infinite number of ways a glass can be broken - so over time all glasses break.)

Whenever the impacts of being wrong can echo through an entire system, the chance of ruin approaches 100%. It doesn't matter if there is any marginal good, the bad will eventually define the expectation - it's a fundamental property of fat tailed systems.

What I love about the paleo / primal / low carb / slow carb / safe starch / keto / etc. movements is that they are not centralized and making decrees. Everyone is tinkering and learning in a small environment. If anyone is wrong, it has a small effect and vanishes from the scene. If anyone gets something right, knowledge will bubble up. Over time we will gather a set of heuristics that are more right than wrong.

I actually have seen this work in the decade I've played with diet. Everyone had the one true way to eat. Some of them have faded, some of them have merged into a new theory, but every day we are learning more through essentially a random walk of phase space. We learn what's right with only a small downside to being wrong. This is far better than guessing and decreeing an answer the feels right. “Senators don't have the luxury that the research scientist does of waiting until every last shred of evidence is in.” - Sen McGovern, 1976. It turns out that none of us has that luxury.



Today's Workout

clean 5-3-4x3
5 heavy swings between sets

then

superset:
next step in pull-up progression and
deadlift 5x35%, 65%x5, 5x85%, 5x70%

then

medium snatch day:
snatch, bell size +1, 7x(7R+7L)/1:00