Moving my blogs to SteemIt

Over the next few weeks I will be migrating all my blogs, including this Engineer Diet blog to SteemIt.  I’ll be removing my blog posts here, and those post I feel are still valuable, I will re-post them on SteemIt.

Currently I am blogging on three distinct subjects:

  • Diet and workout from an engineering perspective (this blog)
  • Information security, crypto and computer forensics.
  • Speculative fiction.

You should be able to find my posts for each of these 3 subjects  here in the near future .

The reasons why I am migrating my three blogs are:

  • I’ve been planning to move my WordPress blog to  my C.H.I.P, but WordPress isn’t really that suitable for running on a small ARM board.
  • I’m a big block-chain fan.
  • Maintaining three different blogs started to make me feel like a schizophrenic.
  • As I want to make any book I self-publish, free of charge, it is great that steemit allows me to collect some revenues from my blog posts, that I could use to pay for cover art and editing.

I hope to make my move to SteemIt as seamless as possible. Keeping this site up for a number of months while removing post that I found to not having ripened too well, and replacing other posts by links to SteemIt.  In about six months I’ll be replacing this site with an URL forwarder for my SteemIt blog, hopefully without braking too much. If you followed my blog here, please follow me on SteemIt and if you like my posts, help me by up voting them.



This entry was posted on 19th June 2017.

Using crappy data to help you spot & ignore bull-shit research.

It is interesting to see how different groups, i’dd call them tribes, are ready to embrace research that fits with their preconceptions while rejecting other research. On social media people including myself often get criticized for ignoring evidence. As a data guy I feel strongly that data science is the science of removing noise, that is the science of finding out what should better be ignored. Looking at nutritional science, the problem doesn’t seem to be that (tribal) people choose to ignore compelling evidence, instead the main issue lies in accepting evidence that could likely be spurious and thus should better be ignored. There is a lot of crappy research out there when it comes to nutrition, and a whole lot of crappy underlying data sets that more often than not will be closed to public scrutiny.

While epidemiological research, especially nutritional epidemiological research is often frowned upon as substandard research, and is often grouped together with the likes of rodent studies and the likes, and while from a causality reasoning point of view it will often be close to useless to prove anything conclusively, in this post I want to look at using epi data as bullshit filter for research that is supposed to be of a higher order.

To do this, we pick a piece of potential bullshit, and an open epi data set that should include the variables that are important to us. Note that the research I look at here is just an example of research that is supposed to be stronger and more rigid than the frowned upon epi studies and animal studies, a systematic review and meta-analysis, in this case one investigating the link between dietary fibre intake and risk of cardiovascular disease. We take a good look at the conclusion of this study and see if we can apply our bullshit filter to it. I picked this study because the claim it makes is one that is relatively easy to map to epi data adjustment in order to use that epi data set as bullshit filter, and because the subject of the supposedly massive health benefits of dietary fiber, especially the quantification of these benefits,  seem to most consistently crumble when applying bullshit filter like the one discussed in this blog post. What I’ll be doing here is something that researchers could easily have done themselves and something not uncommon when exploring scenario’s in a forensic setting (much of my own background is in forensics), it is however a check that in nutrition no one seems inclined to applying.

So what does this paper claim? What effect > And how strong is this effect? The paper claims :

“Total dietary fibre intake was inversely associated with risk of cardiovascular disease (risk ratio 0.91 per 7 g/day”

Remember we are talking systematic review here, so we should expect the probability of this effect to be real and show up in any epi set to be rather high.  So lets bring in our bullshit filter. as our bullshit filter we use the 1989 data sets from the China Study.

I’m not talking about the equally named book here, the book is, to state it bluntly, a smoking pile of rhetoric against (animal) protein and saturated fat that lacks the substantial rooting in the actual China Study data set to justify its name. That is, there is surprisingly little china study in the china study book. Don’t get me wrong, it is an amazing book that posed an interesting hypothesis, and I feel strongly we shouldn’t judge pop-science books as if they were scientific papers, it is just that this particular book has rissen to a status of scripture in certain dietary communities, where it should be considered a well written book about an originally compelling hypothesis that like so many good hypothesis’s has contributed to the scientific process but should no longer be considered viable.     But enough about the China Study book for now, we are here to look at poorly done peer reviewed science here, not at pop-sci books advocating superseded theories.

Last November CTSU pulled the China Study data sets, data that had been publicly available for at least 13 years from its website. Tribal conspiracy theorist HFLC might argue that the CTSU pulled the China Study data sets in some kind of evil vegan conspiracy. I’m not assuming any such malice nor do I feel there is any real reason to assume so.  The important thing is that still has the raw data available for us to use as bullshit filter.  The data set isn’t by far as impressive as it is made believe to be. There are many holes and the data has no info on any higher order moments of variables from sub populations. We can’t really be all that picky though. Many interesting data sets are completely unavailable to the public so the people behind the China Study, no matter what one might think about their persisting stand behind an outdated hypothesis, must be lauded for making the data available to the public rather than hiding behind IRB firewalls. It must have been a hell of a job to compile this data set, yes there are nits with its design, but we can still use it.

There are 66 data points that contain both fiber intake and cardiovascular mortality.  Each data point represents a regional Chinese population in 1989.  This is very interesting as this will likely make the data set void of many of the western associational clusters that might dominate our systematic review. So let’s look at our data points.

Pretty much the starry sky type of plot we are used to expect from epi data sets, so lets look at the correlation numbers. Our correlation is +0.06 with an unimpressive p value of 0.64.  Does this mean anything? Well, it means that based on this data we are unable to reject the null hypothesis that this is a random correlation, but that isn’t actually what we are interested here, and if it was, not being able to reject a null hypothesis basicly gives us zero info.  We need a different null hypothesis to work with, and the paper we discussed at the beginning gave us one.

So how do we go about at using this statement as null hypothesis ? :

“Total dietary fibre intake was inversely associated with risk of cardiovascular disease (risk ratio 0.91 per 7 g/day”

Well, let us start off by assigning a standard risk ratio of 1.0 to the mean fibre intake in our data set (12.99). We also look at and take note of the the mean of the cardiovascular mortality figures (3.24).

Now for each data point we have two values, F: the average number of grams of fiber and V the cardiovascular mortality figure for the region. So first we look at how many times 7 grams our F value is removed from the average number of 12.99 grams.

  • n=(F-12.99)/7

The next thing we do is that we calculate an adjustment risk factor based on P:

  • R=0.91ⁿ

Finaly we calculate the residual error for our fit using the adjustment risk ratio and the average mortality number:

  • E=V/(R*3.24)

So now we look at how E looks when plotted against fibre intake:

Well, on first sight just as starry sky as and kind of similar to our original plot, but notice it is kinda tilted. So how about the correlation and p value? Well our correlation went from +0.06 to +0.27, not really massive, and we would expect something like this given that our model predicted a negative r to begin with, but how about the p value? Now here is where it gets interesting and possibly kinda controversial.  Our p value for the correlation in our residual error is 0.026.  I intentionally chose these two studies to search out an edge case such as this. Should we reject our null hypothesis that is based on what is perceived to be higher order evidence, purely because an unrelated and perceived as sub-standard  epi data set suggests we should? I’m saying suggests, because 0.026 isn’t all that strong to begin with, at least, it wouldn’t be if we were looking to reject a regular null hypothesis.

But we should look at what our null hypothesis actually is. In law enforcement and thus by proxy in forensic science, the concept of asymmetry of proof is very important. The evidence needed to convict should be helt to higher standards than exculpatory evidence, and I would argue, especially given the arguably high rate of spurious claims in nutritional science, that this too should be a valid strategy.  So how about epi evidence being substandard evidence? Well one thing to consider is that while the data is arguably confounded, spawning higher amounts of spurious associations on one end, the data points that make up N (used in calculating the p values) represent whole sub populations, thus leading to conceptual inflation of the true p.  If we balance these three aspects, I would argue that a p value threshold of 0.05 would be just about right for determining if a piece of nutritional research should be ignored.

I know that using arguably confounded data sets to determine to ignore evidence that is widely considered superior will be a hard sell, but when considering  the asymmetry of proof as used in law enforcement and forensics to be essential in a field of science plagued by spurious associations and spurious causality claims, we must affirm there is a definite need for bullshit filters. If the epi data has minimum cultural overlap with the research looked at, there is definitely added value to using epi data, and if as in this example we clearly can’t reject the default null hypothesis while we can reject the nul hypothesis that states that the effect+size defined in the paper are real, than  ignoring the paper is an act that helps reduce noise as reducing noise is the real reason to use data science and nutritional science has a whole lot of residual noise as it stands.

As stated, the problem with nutritional science and bias doesn’t lie in bias driving people to ignore compelling evidence, it lies in biased and/or (data-sci wise) under informed people failing to ignore likely spurious evidence. Most research, even ‘golden standard’ RCT research in nutrition has some data science issues with it. Medical and nutritional education seem to only touch the very basics of data science, allowing for a wide range of issues to sneak into nutritional research that should be of the highest standard. The use of epi data as outlined in this post is just a band aid, and one with conceptual flaws, but it is a simple enough band aid for people with any background, including medical or nutritional science to apply, a type of band aid that in forensic science is quite common (though not completely uncontroversial). Use it as a band aid for finding out if you can or rather SHOULD ignore a certain publication.


This entry was posted on 22nd February 2017.

Calculating your path to a better health

in the distance past I wrote about the use of a Generic Body Health index  as replacement of the Body Mass Index in its role as target variable for diet and workout. In the mean time I have been tuning and fitting this concept and have started working on an on-line calculator for helping people setting diet and workout goals. So after the last beta of my calculator I thought it time to revisit the subject in this blog post. In this post I shall try to describe how my calculator can help you to create a path to a better health for yourself by not focusing on weight loss or on getting a six-pack beach body before summer as a goal but by combining different non-intrusive measurements with your strength statistics in a way that accounts for your age and gender and aims for incremental and sustainable progress towards lowering your overall health risks.


Your health is not about your weight. Not your TOTAL body weight anyway. High muscle mass is more healthy than low muscle mass for example and your subcutaneous fat mass might not be helpful in attracting members of whatever sex(es) you are attracted to, we can’t really classify it a risk factor outside of a dating context. Well maybe there are other social risks, but the base line is that there is zero evidence for the hypothesis that subcutaneous fat mass is a serious independent medical risk factor. For you as an individual however, subcutaneous will be strongly correlated to a weight related variable that is very likely to be a risk factor: Visceral Adipose Tissue (VAT). The fat mass in between your organs. Many diets and diet and sport consultants will focus on total weight, expressed commonly bu the Body Mass Index (BMI).  ce9fx35uaamat2z

The BMI in short, is a body length adjusted measure of your TOTAL Body Mass (TBM). Total body weight that includes both the VAT risk factor “AND” the actual protective Lean Body Mass (LBM) that consists of your bones, muscles, organ meat and extracellular water mass. Basically:

  • High VAT is a potential risk factor
  • Low LBM is a potential risk factor
  • SAT in itself is neutral but correlated relatively strongly to your VAT
  • TBM is composed of VAT+SAT+LBM and as a whole quite useless as target variable.
  • BMI is a height corrected measure of TBM and thus useless as a target variable.

If you have a BMI of 40, according to the BMI definition you will be class III obese. As a statistical tool for whole populations, this might be a relatively decent assessment. You as an individual might have a BMI of 40 because you have been doing strong man training six hours a day six days per week for more than a decade with the result that you’ve become a 150 kg mountain of pure muscles who could wrestle a grizzly bear to the ground with one arm behind his back. Your body fat might be as little as 10% of your body weight, but according to the statistics you would be in with the morbidly obese. Given the low percentage of people who actually spent 36 hours a week at the gym lifting the weight equivalent of a small car when  looking at the total population, for an assessment of obesity in whole populations, BMI might indeed be somewhat useful. BMI is completely useless however as a tool in fighting individual obesity or more importantly as a tool for total health risk reduction. If you want to reduce your risks through body composition, then you need to make sure to:

  1. Reduce your VAT
  2. Increase your LBM

So what will be the net result of this for your BMI you may ask. The most direct answer to that question is: ‘it depends’. The more important answer to that question though is: ‘it doesn’t matter, not even a bit!’. Your BMI is irrelevant. it may not be irrelevant as a filter for discovering that you might have a health risk problem, but that is all that it is useful for you as an individual. If you found out your BMI is 35, don’t ignore it, use it as a reason to find out more and if needed take action. Don’t, however think that you need to FIX your BMI, you don’t and if you try to it will only get in the way of sustainability of your progress. The thing that needs fixing is your body composition.

Bio-electrical Impedance Analysis

biaIf you have been reading my blog before, you might know that I’m a real big fan of BIA scales.  Bio-electrical Impedance Analysis uses a small electrical current at different frequencies to measure the frequency-dependent impedance of different paths through your body and then uses the acquired data to estimate your body composition. There are different types of BIA scales of varying accuracy and consistency. You should not buy a consumer grade BIA scale for home usage. You will likely end up paying up to hundreds of euros for equipment with relatively inaccurate, but more importantly often inconsistent results. A better strategy is to seek out a gym that has a professional grade BIA scale, or if you are seeing a doctor or dietitian on a regular basis who has one, discuss using the one at their office with them while you are discussing your health goals. A BIA scale worth its weight will give you multiple useful numbers about your body composition. Some a bit more granular than others ans some more accurate than others.

  • Total Body Mass (TBM or weight)
  • Lean Body Mass (LBM)
  • Total Fat Mass (TFM or TAT)
  • Body Fat Percentage (BFP)
  • Visceral Fat Mass or percentage (VAT)
  • Total Protein Mass (TPM)
  • Total Water Mass
  • InterCellular Water (ICW)
  • ExtraCellular Water (ECW)

It is good to keep track of all of these and keep them neatly in a spreadsheet or journal. We shall be using TBM and BFP in our calculator while keeping an eye on VAT and TPM. VAT and TPM would actually be better measures if we were looking at it from a static risk point of view. The problem though is that due to the smaller numbers involved, the limitations in the accuracy and consistency of the measurements of these variables, these variables are difficult to use effectively in a control feedback setting needed for controlled semi-critically dampened progress. We, however do need to stay on top of these variables. protein mass going down or visceral fat mass going up more than can be accredited to measurement inaccuracy are a sign that something may be going wrong.

Alternatively: skinfold and waist circumference

If for some reason you can’t regularly make use of a BIA-scale to track your body composition, there are a few poor man alternatives to consider. The first one is the use of skinfold measurements. You can use a simple caliper to measure your skinfolds in a number of locations. I won’t go into this too deeply, but this site has a nice explanation and calculator you may use. It is important to note that while both BIA and skinfold have similar accuracy issues, the correlation between the skinfold technique measured BFP and the actual VAT is smaller than that between BIA-measured BFP and VAT. It thus becomes even more important to track something close to VAT apart from BFP. A decent way to do this without BIA is by measuring your waist circumference and your total body mass and then using them to calculate your conicity index (CI).


Consider your CI the equivalent of your protein mass and visceral fat mass measurements. If it goes up you are doing something terribly wrong. It is an important safeguard that is useful even if you are applying BIA measurements, but it is indispensable when  you are resorting to skin-fold BFP measurements.

An age and gender adjusted Body Fat Index (BFI)

What constitutes a healthy level of body fat differs depending on your age and gender.


The formula is simple. If you are a female, use zero for GENDER, if you are a male use one. Fill out your age gender and body fat percentage. The result will be a number that will go down all the way to zero if you are approaching under-fat. An obese person will probably have a BFI in the 3..5 range. Our ultimate goal would be to move the value of our BFI as close to zero as we can get it, but also considering and balancing against an other bodycompositional factor.

Lifting weights

While BFI is a good measure of individual obesity, we know low LBM also can be a possible risk factor. In order to keep this risk factor down we want to increase or at least sustain our lean body mass. While our protein mass is a good static indication of this risk factor, like VAT there are issues with accuracy, consistency and responsiveness of this variable. If we use body strength as a proxy we not only measure LBM improvement, we also get quick feedback when our diet or workout goes all wrong. If your body strength drops quickly as a result of diet that will be a real big sign that something is terribly wrong. Thus lifting weight has two goals. Measuring our body strength as indication of our LBM related health and making sure our low-LBM risk is reduced.

An age and gender adjusted Body Strength Index (BFI)

Just as with body fat, we can use body strength in an index that ideally should move to zero. Just as with the BFI, we define zero as a hard to achieve but still potentially reachable level.

We are less interested in your absolute strength than we are in your relative strength, relative to your body weight. As we care more about body composition than weight, we will see two possible paths to increasing relative strength. You can gain relative strength by getting bigger and stronger muscles and lifting bigger weights or you can gain relative strength by managing to maintain your strength and muscle mass while losing body fat. For some people, the first comes more natural and for others the second.  We define our relative strength as the sum of our three rap maximum for the big three strength exercises: deadlift, squat and bench press.


The reason that we choose three rep max rather than one rep max is that training with a three rep max is more likely to improve your lean body mass figures than a one rep max approach.

The first thing to realise is that starting at middle age, you will have to adjust your goals if you want to keep them realistically. After having looked through multiple sources, the following correction factor appears to be a decent fit for 39 to 65 year old people. :


Given that the correction factor only applies to middle aged people and younger and seniors, we define:


Or for those who prefer computer programming notation, we can write the same as:


We now define the top achievable strength for a given age and gender as:


So with this top strength metric we can finally determine our Body Strength Index (BSI)


The Body Health vector

Now about using our new BSI and BFI. Consider we put BSI on the X axis and BFI on the Y axis and define our Body health vector as a point on this graph. Ideally, BFI will be low ass will BSI but chances are one or both aren’t as low as we would like them and we need to look for a good path for our BHI vector to gradually move closer to the origin. We propose that the best and ideal route from any starting point balances the goal of reducing both components with the idea that balancing out the two risk factors is important. Realizing that many people will have to work with at most high-school level mathematics, we choose to use a circle fragment as ideal path through body health space.  There is no reason to make it any more complex than that. We want our vector to move over time from its current position to the origin, where the ideal angle for reaching the origin would be 45 degrees.  So we start by looking for a circle that fits the bill. A bit of simple geometry gives us a formula for first finding the radius of our circle:


And with this radius, we look at finding the center of our circle.


The path of 5% decrements milestones

With the center of the circle we just found ,we can look at plotting a first step to progress. A first major milestone. As progress will become harder as we come closer to the origin, we go for relative incremental milestones. A good target for such is a 5% target. Let us look at how we calculate this 5% target. It is important to realize that the center of our circle can either be in the second or the fourth quadrant depending on if BSI is bigger than BFI or if BFI is bigger than BSI.  When we look at the angle for the origin as viewed from the circle, we get:


Now for the angle of our BHI vector. Depending on the quadrant of the circle center that angle will be one of:


for fourth quadrant circle centers or


for circle centers in the second quadrant. Now with those two angles we can easily determine the angle for our next milestone:

phit That in turn gives us the milestone target for BSI and BFI



Now I hear many of you thinking, I know it’s easy high school level math, and it is not hard to understand, but it’s quite a bit of work when compared to the good old BMI formula, and you would be right. For that reason, I’ve created a simple web-based calculator for you. The web page looks a bit archaic, I know. I am not a web designer so I don’t even try. Give the calculator a spin and see for yourself what suggestions it makes using the above algorithm.

Important: I’m not a doctor and neither is my calculator!

As many of you will know or will have guessed from the above, I’ m neither a doctor not a nutritionist but my calculator above will give you very specific advice for health targets. While I feel strongly that the advice is the result of good engineering, there might be reasons why for some people the advice might go against an individual’s medical best interest. Discuss usage of this tool with your GP or dietitian, especially if you suffer from medical issues, and if using control feedback loops build using this tool yield very high or very low levels of specific foods or nutrients, discuss those with your GP or nutritionist as well. In that sense, it should not be seen any different from BMI except for the fact that this BHI approach, when practiced under professional and qualified supervision should yield better and more sustainable results than those that could potentially flow from the usage of BMI as intervention target.

Measure what goes into your cabinet/fridge/bin, deduse what goes into your mouth.

So far we only discussed our target variables and target milestones. What we did not address was our input variables such as workout timing, length, exercises, macro split, meal timing, micro nutrients, etc. While we will discuss these and I have discussed these in other posts, one thing is important to note. It is notoriously hard to keep track of your food and drink intake when you try to keep up with what goes into your mouth. A simple solution for this is to simply not do that, at least not in the early stages of your efforts. Start off by only making note of everything you buy. Keep your reseeds and keep track of spoiled items you throw away. That is, keep track of what goes into your fridge and what disappears into the bin. The difference will mostly be what went into your mouth or that of the people living with you. Surprisingly, even with a household with many family members, a fridge-bin estimate is likely to be more accurate than any attempt to keep up with what actually goes into your mouth.

The error vector and (under-)critical dampening

I’ll close this post with lifting a tip of a subject for an other day. The result of the calculator is a 5% milestone, but chances are you will miss that target. You might undershoot or overshoot your BFI target while reaching your BSI target or vice versa. What you do then is that you take the two vectors, one from your old BHI to your target BHI and one from your old BHI to the BHI you achieved, and you look at the angle between the two. This angle is your error signal, and your task in designing a personalized control feedback for your diet and workout is to find the proper variables for translating these errors into effective feedback. Don’t worry if you don’t understand this last part. You can use the target milestone even without understanding control feedback theory. The result will not quickly move towards something close to critical dampening, but that is OK for a beginning, as long as you are making progress.

This entry was posted on 8th November 2016.

An outline of Control-feedback based diet&workout tuning.


I’ve talked about the use of Control-Feedback (CF) theory a few times before, but what I truly haven’t discussed yet before is a relatively clear outline describing how to approximately go about actually using it. I’ve been running CF based experiments on myself for quite some years now, sometimes with great results, sometimes with poor or even bad results. There are two things that are essential that I’ve learned the hard way and that I need to share so others don’t fall into the same traps:

  • The external CF loop needs to have a significantly slower feedback than your internal feedback loops, at least for the big change inducing input vars.
  • Multi variable feedback only after switching single var rounds is found to result in input-var oscillation.

So basically we focus (first) on using single var feedback loops to find decent levels for the big impact input vars. But where to start.

Step 1: establishing a baseline (3 months)

In our first step, we make a couple of relatively big ‘timing’  changes to our diet and lifestyle, but without actually changing much about the things we actually eat over the span of a day and without actually changing much about our workout routine. We do this in order to help us separate a number of variables that can greatly influence each-other when they are close together in time. We also do this because this reordering in time seems to actually have positive effects for many people all by itself. So lets look at what we change to get at our baseline:

  • Determine your current macro split (meals + snacks +drinks)
  • Move “all” your macros to two or three bigger meals a day, no more snacking or drinks with in any way significant number of calories in between.
  • An extra (small) high-protein post-workout meal is both OK and suggested.
  • Divide your day in a low carbohydrate and a low fat half, that could be.
    • Low fat breakfast & lunch;  low carb diner.
    • Low fat breakfast; low carb lunch & dinner.
    • Low fat breakfast; no lunch; low carb diner.
    • No breakfast; Low fat lunch; low carb diner.
    • Low fat breakfast;Low carb lunch; no dinner .
    • Low carb breakfast & lunch;  low fat diner.
    • Low carb breakfast; low fat lunch & dinner.
    • Low carb breakfast; no lunch; low fat diner.
    • No breakfast; Low carb lunch; low fat diner.

Low carb breakfast;Low fat lunch; no dinner.

  • On workout days, make sure your workout falls in the low fat part of the day.

Now that is our diet baseline. You eat exactly as much of every macro as you are used to eating, the only real change we really made was making sure there is a couple of hours between fatty meals and high carb meals. Now for the workout baseline. Bodily strength is a major and relatively quick responding variable that gives you information about things going well or going horribly wrong. A high relative strength indicates a high lean body mass and quickly decreasing bodily strength indicates major issues with your diet and workout program. We need a way to measure your strength and we choose to do so in a way that actually helps improve your strength and overall health. So for our workout baseline we make the following changes:

  • Replace (part of) your workout with strength geared resistance training.
  • Incorporate the “big three” into your resistance training regime:
    • Squats
    • Bench press
    • Dead lifts
  • If your current workout routine is less than 5 hours of workout per week: increase is and increase your caloric intake accordingly.
  • If your current protein intake is less than one gram per kg of lean body mass, swap some carb and fat calories for protein calories in such a way that your carb/fat macro split remains the same.

That’s it. You eat basically the same but time it differently. Only if you were truly protein deficient to begin with do you change your macro split. You work out the same number of hours per week but shift (part of) your workout from aerobic and or fitness oriented workout to a more anaerobic and strength oriented workout. Only if you were really inactive do you increase your workout hours. Meanwhile you measure. You simply keep track of a number of important variables related to body composition and strength, but keep this baseline stable for about 12 to 13 weeks. It may sound strange, especially when you are used to counting calories, but this baseline will lead to body compositional changes all by itself. You probably won’t lose weight or anything, but chances are you will improve your lean body mass and visceral fat measurements already without actually working out more or eating less. This is also the reason we need to keep this small change up for a couple of months and measure it. If we were to change many things at once we would not be able to differentiate what variable was responsible for what change. It might even be possible that different changes counteract each other in such a way that a negative change might get interpreted as positive. One variable at a time at least until we have established the impact and optimal ranges for the big impact variables. But now for our measurements. Our primary tool for measurements will be a Bio-Electrical Impedance scale. Most gyms and the more informed dietitians and general practitioners will have a BIA scale. If your gym, GP or dietitian doesn’t have one, try to switch gym/GP/dietitian if you can. A professional grade BIA-scale is an indispensable tool for a reliable CF setup. If you happen to be a dietitian of GP, get yourself one of these scales, your patients are worth it. And no, I don’t have any shares in companies who make these things for those who might be wondering. So what do we measure?:

  • Your body strength: (Squat / Bench / Dead-Lift)
  • Your total body weight.
  • Your lean body mass.
  • Your protein mass.
  • Your total body fat percentage
  • Your visceral fat percentage.
  • Your waist circumvent.

Measure these things every workout day if you can, or every dietitian/GP visit if you must, but measure them as often as you can and keep them in a spreadsheet for later usage. As this blog post is just about an outline, I won’t go into how to use these measurements in composite variables best suitable for CF usage, I’ll come back to that in later posts. Chances are you will see improvements to these measurements in this 3 month run-in phase, and rate of chance will have now been reduced to a relatively low level. So now it is time for our first real change in terms of macro usage.

First round: protein (2..4 months)

We have established our baseline and we know how our measurements have been changing up until now. Time to make changes and apply CF theory. So what do we go for in our first round? Protein! Protein within our macro split that is. We take care to keep our fat/carbohydrate split the same as it was before. We take care to keep our total calories the same as it was before. All we do is make changes to the protein vs carbs+fat part of the macro split. Chances are your protein intake has been relatively low so far and swapping non protein calories for protein calories could have a relatively high and positive impact on your stats. I’ll get to the technical details in future posts, for now just consider finding a good protein starting point to be our first goal.


Second round: allowing snacks (1/2) ? (2x 6 weeks)

By now you’ve gone something of half a year without snacks. As many people will see no snacking as highly restrictive and unmaintainable long term. We start with the one least likely to be problematic, snacks within a same-macro window. Half our day is low fat, the other half is low carb. In this phase we keep the no snacking gap between the last low-X and the first low-Y meal intact, but apart from that we allow for snacking within. Pick one macro, fat or carbs and move some calories within that part of the day to allow for snacking. We are not looking for improvements here but for absence of deterioration. How much snacking can we get away with basically. How non-restrictive can we make our day halves in other words.

Third round : resting day carbs (2..4 months)

As far as macros are concerned, workout days and resting days can be something quite distinct. Chances are you are eating a way to much carb centric diet on days without any anaerobic activity. In this round we start swapping carb calories to fat calories. Chances are you will find yourself at the low carb side of the spectrum and you will find yourself fasting half the day at the end of this round, and your results will have improved from where they were in our first round.

Fourth round: minimizing the inter-macro gap. (6..12 weeks)

For those who find a many hours gap between fat and carb to be restrictive, we need to look at minimizing the gap as best we can.  Carbs and fat combined in high quantities in a single meal spells disaster, but how much time needs to go between these two macros differs from person to person and may even differ over time. In fact, a decreasing minimum for your inter-macro gap is an important marker of progress. It seems the minimum inter-macro gap might be a relatively decent no-needles solution for acknowledging improvements in your insulin response, but that is a subject for an other day.  I’m suggesting looking at your minimum inter-macro gap once a year to see how you are doing. That is unless the following rounds have you at one of the extremes in the carb/fat spectrum.

Fifth round: workout day carbs

While it is very likely your resting days will end up low carb, for workout days this might or might not be the case. In this round we see what happens to our body strength when we move carbs to fat on workout days and try to establish a minimum number of workout-fueling-carbs to consume on workout days.

Sixth round: calories finally

So far we hadn’t changed anything in our caloric consumption. But now that we are finally here, we are in a position where watching calories has actually become useful. Using our measurements we can plot an ideal-progress circle-fragment in a plot that puts a bodily-weakness index against a bodily-over-fatness index, a subject I discussed in other posts before and that I shall soon revisit. From here on calories will be a major factor in balancing muscle mass and strength improvement goals against body fat related concerns. If we had targeted calories before, chances are there would not have even been a caloric level allowing improvement along our ideal-progress circle-fragment, that is, loosing body fat might have been implicitly linked to unacceptable high loss of strength and/or lean body mass. Gaining substantial strength might have been implicitly linked to unacceptable high body factor even visceral fat mass increase. It is only now that we can start doing the really interesting control feedback stuff. remember the important thing from here on is our ideal-progress circle-fragment. We want to move as close as possible to that arc at the fastest rate achievable. From this point on we start to make binary changes for at least 6 week periods while using calories in a CD feedback loop in order to move across the ideal line.  Always give at least 6 weeks for binary changes before writing them off. Give your bodies internal CF system the time to adjust and add more weeks if you notice a reversal there.

Where to go from there?

Basically we now have calories, timing and top level macro split covered, so now it is time to look at swapping out macro sub types and sources. Swapping out weight training exercises, changing the sizes of your sets, adding or removing cardio from your workout program, looking at micro nutrients supplements, etc. Too many things basically so you need to pick those you think are most promising. Read papers and wild claims about nutrition and body composition. Read strength sports magazine articles and wild nutritional and workout claims made in those. The science behind these claims might be totally based on educated or even uneducated guesses and anecdotes or inconclusive possibly spurious correlations and causation, we are looking at YOUR body using ENGINEERING here, not nutritional science. Further, once you are reasonably on track, multivariate control feedback finally becomes achievable. You can design a personalized CF system that allows you CD on multiple input variables at the same time. Something we definitely could not have done at the start.

Zero dogma

Before you are at the stage where multivariate control feedback loop system design becomes a serious option you will probably be a few years down the line, but your diet and workout will be more dynamic and customized to your body than could be achieved in any other way. It has been a hard pill to swallow for me to have to come to the conclusion that either this long run-in is inevitable for nutrition and workout or my engineering skills are somehow insufficient to find the proper balance between system stability and speed. It might be possible to create a shortcut to some of this by extrapolating some of my current data from me and less than a hand full of my gym buddies that I managed to give all of this a try, but currently that would be just guesswork. When I write about this stuff I tend to get lost in technical details and in the urge to want to explain and justify every step along the way. In this blog post I’ve intentionally omitted any such detail in order to show a multi year outline to using engineering in general and CF theory in particular to slowly finding your own personal sustainable diet and workout program. I realize the missing details and justification, while making it more accessible possibly make it less convincing and I’ll try to get back on multiple things mentioned here in later posts as to fix that, and I also realize that many people would rather opt for the quick results that many diets offer rather than using a slow CF system to try and find your own personal optimum. If you do however, I encourage you to go for it until you hit a wall in your progress as you doubtlessly will, and than start your baseline from there. The great thing about this approach is that while being slow, it is completely agnostic with respect to different dietary dogmas. If your HFLC diet is too low in pre-workout carbs, you will find out, If it isn’t you will see acknowledged you were right to start with. If your low fat plant based diet is too low in protein or too high in sugary fruit, you will find out, and again if it isn’t you will see acknowledged you were right to start with. Good results or lack of return of bad results aren’t necessarily optimum results, and even what might be best for a population as a whole might not be particularly suitable for you as an individual. CF theory does not lie. In the case of workout and especially in the case of nutrition it may require patience due to interaction with your bodies own internal control feedback loops. We are aiming for coming relatively close to critical dampening and avoid any kind of oscilative behaviour.  Unfortunately that means we need our feedback to be relatively slow and accept it will take a long time for an optimum diet and workout routine to reveal itself.

Some wild speculations for speeding things up

If you don’t like wild speculations, stop reading here. If you are willing to take a big chance in order to potentially speed things up a whole lot and potentially get much quicker results, then I would suggest doing the following things in the run-in period, deviating from the play it safe minimum change paradigm. A warning, this is stuff that works for me and at least more than one of my gym buddies or not for me but for more than two of my gym buddies. No guarantees any of this works for you:

  • On your resting day only eat a big high fat high protein low carbohydrate breakfast, fast the rest of the day.
  • Start of with an average of 1g of protein for each lbs of lean body mass.
  • If you can manage, don’t return to snacking.
  • If it fits into your time schedule, start your workout day with nuts and berries as breakfast, then hit the gym early, have a small high protein low fat moderate carb post-workout than don’t eat till lunch. Rest of the day low carb high fat.
  • Stack up on low caloric veggies (both whole and juice)
  • Take BCCA, L-Carnitine , L-Leucine  and CLA supplements.
  • Eat organ meat twice a week.

At this point in time I don’t know of anyone who started out this way, but of multiple who ended out this way, so it might be a shortcut to start out this way in the run-in period, but then it might not. At this point in time it is just an educated guess. A guess though that when correct might save you possibly a year or more of CF induced slow changes, so it might well be worth the gamble.

This entry was posted on 28th October 2016.

A high-protein Ovo-Ento Vegetarian diet for body and planet health.


Are you concerned with the environment? Are you concerned with your health? Like most people, you are most probably concerned with both. The two are probably connected anyway, but in much more complex ways than some people want us to believe. In this blog post I am going to challenge the ethics of both the vegan community and of a group of doctors with strong ties to militant animal rights groups who want you to believe that a whole food plant based diet is supposed to be both good for your health and the environment when objectively it is an ill defined diet that in many forms introduces major health risks, while doing relatively little for the environment. I am also going to challenge, and I know this is going to be a hard one to crack, the cultural aversion of many of you with respect to a certain food sources that you may not only not think of as food, but might even be completely repulsing to you at this point in time.

But first things first, lets look at decomposing the problem with our diets that we are trying to solve. Our primary goal is health. Our own health and that of all other humans. In order to be healthy, we strive for a healthy body AND a healthy environment for us and all others to live in.

Thus the first part of the equation:

Health = Healthy body + Healthy environment.

Lets start of with the first part of the equation. I’ve been looking at many studies, and more importantly, many data-sets for many years now that all make the wildest claims about nutrition and health. When looking at the data though through the eyes of a data-engineer, very few of these claims seem to be very solid. In fact, in most other fields of science the results of most nutritional science studies would be simply considered as inconclusive. Not surprising thus that we get so many conflicting messages about different foods and supplements being super foods one day and being causal in disease pathways the next. For many things, the environment (that we will look at next) might be an actual bigger factor for overall health than something like eating fibrous foods or eating meat. One piece of data that does however stand out above all is a cluster of variables that are directly or indirectly related to obesity. If we zoom into the problem of obesity, the first thing we notice however is that the main measure of obesity as used today, the Body Mass index, is actually one of the weakest variables from that cluster with respect to health impact. If we zoom into this cluster, we get variables like waist to height ration, conicity index, waist circumference, etc as having massive correlations with a wide range of health issues. On the other side, then looking at things like body weight and BMI that are linked much weaker to these health issues, we see that body muscle mass and lean body mass not only mitigate the effects of our harmful cluster in a way that would be expected if the size of the belly was all that mattered and muscle mass was irrelevant. There seems to be quite convincing data showing there is independent risk reduction tied to higher lean body mass levels.  When we look at some excellent studies with respect to the variables in our main cluster, we can see convincing facts pointing to Visceral Adipose Tissue (VAT), the fat that is in between our organs to be the main thriving risk factor within the cluster, and indeed if we apply the hypothesis that VAT is our primary risk driver and add to that the concept that high LBM reduces VAT driven health risk, we get at the second part of the equation:

Healthy body = Low VAT + High LBM

So we want our VAT levels to be low, but we don’t actually care about subcutaneous fat. Have massive love handles? Don’t mind those, they are NOT a health risk. Have a BMI of 35 or even 40? Don’t mind that, it’s not an independent risk factor. We care about the visceral fat time-bomb. If you aren’t a power lifter and have a 150 cm waist and a BMI of 40, chances are your VAT will be high though, but remember, when we are going to try and fix that, we won’t be aiming at lowering your body weight, in fact, your body weight might actually increase quite a bit first while working on the low LBM problem.

It is important to truly grasp the implications of this part of the equation. We want VAT to go down, but NOT at the expense of LBM. We want LBM to go up, but not at the expense of VAT. But trying to fix both can be hard. If there is an imbalance, as there often is, between the two components of good health, it is good to focus on the most urgent one first and aim for balancing the problem first. This may mean that if your LBM is quite decent and your VAT is quite high, focusing on weight loss first might be a good idea, even if it means losing some of your precious LBM. Surprisingly though, and contrary to the way most nutrition experts approach it, for many so called over-weight people, especially those who went through the dieting circle several times around, a low LBM is actually the dominant problem and gaining lbm should take precedence, even at the expense of possibly gaining some more VAT initially. This part can be a big mental issue for many who are conditioned to think of themselves as being over-weight. But now, lets look at how to address the VAT issue at any given caloric level. Its about both what you eat and when you eat it. The next part of the equation:

Low VAT = Low bad carbs + Timed good carbs

So what are bad carbs? In short, its added sugar and refined grains. There are sufficient sources available showing why these two should be avoided and how they are linked to obesity, so no need to rehash that here.

Low bad carbs = low added sugar + low refined grains

So how about timed carbs? We will get to the resistance training part in the section below about LBM, but basically you will want to time most of your good carbs around your workout, dead-center inside of a fat-free time window. So before your carb window start, you should not eat any fat for a period of time. After it closes, you should not eat any fat for an other period of time. Your workout should probably fall inside of your carb window. The reason for this window is that carbs and fat together truly spell growing body fat deposits.

Now for the tricky part, and the part leading to the dietary advise that contradicts that of much of the vegan experts on the subject, increasing our LBM. In order to increase our Lean Body Mass, we want to grow our muscles, most importantly the bigger muscles. Just as with our love handles, this is not about getting a beach body, it’s about health. Do resistance training. If you have time for it do things like shoulders and biceps, but for the bigest part of it focus on the bigger muscles, these are the ones where you can gain the most weight the quickest. Legs, core, triceps maybe. As for this part of the equation, its quite simple really:

High LBM = RT + Protein

You need to really hit those weights, and you really need to eat proteins and not at the levels that your vegan doctors will have you believe are sufficient, well maybe at first if your LBM is really low, but as your LBM goes up, so should your protein intake. Normally I use the metric system, but this one is one where pounds are easier to reason with:

Total Protein = LBM g/lb

Taking your current lean body mass in pounds as a measure of how much total protein in grams you should eat as an active person who does resistance training should be a reasonable level for most of us. You need protein as source of calories, but more importantly as building blocks to repair and grow your muscle mass in the recovery phase of your workout schedule. You can tweak it a bit to see if you get better results, but more than marginally more or less than the 1 gram per lb of LBM is something that is testable sub-optimal for most of us. As for what protein to eat. There is the issue of amino-acid completeness. You don’t want to get a one sided amino acid profile for multiple reasons. There is however also the environmental concerns with many of the broad spectrum protein sources. Given that the total amount of protein we take is relatively high and plant based proteins from nuts, legumes and mixed source protein powder has quite a diverse profile already, we can get most of our protein from plant based sources. Looking at epi for protein source split all cause mortality, especially when looking on epi data from countries/times with low processing levels of common sources of protein such as meat, we see there is a real concern that the lower extreme of animal based protein is not a place without potential health concerns. To steer clear of this extreme, yet also consider environmental concerns, and as we will later see, concerns with respect to cultural acceptability of certain foods,  a safe split of protein sources could be:

Total Protein = ⅔ Plant protein + ⅓ Animal protein.

Lets zoom in a bit. Let us first look at look at the animal sourced part.

Animal protein = ⅓ x LBM g/lb

So for example, if you are a 200 pound man with a body fat percentage of 25%, who has started going to the gym at least 4x/wk. Your lean body mass will be 150 pounds. That means for your protein, 100 grams of plant based protein and 50 grams of animal based protein would be close to your ideal intake. What do you need to eat to get 50 grams of animal protein? Some single source examples:

  • A 200 gram steak
  • Four large eggs
  • 200 grams of Gouda cheese
  • 250 gram of chicken wings
  • 300 gram escargot
  • 175 grams of tuna
  • 250 gram meal-worms
  • 250 gram shrimp
  • 400 gram calamari
  • 75 gram whey protein powder.

It is relatively easy to consume 50 grams of protein a day from these sources if we don’t consider the environmental part of things. A subject we look at later. .But then we need to look at the 100 grams of plant based. We see that getting 100 grams of plant protein without restoring to supplements might actually be quite challenging to many people:

  • 5 kg potatoes
  • 3.7 kg rice
  • 3.5 kg brocoli
  • 2.5 kg kale
  • 1.1 kg of lentils
  • 1.1 kg black beans
  • 800 grams of brown bread (about 30 slices)
  • 700 gram walnuts
  • 500 gram almonds
  • 400 gram peanuts
  • 275 grams hemp seed
  • 175 gram spirulina (sea weed)
  • 125 gram vegan blend protein powder.

If we now look at this list from an ecological footprint point of view,  there are some rather surprising facts. One is: some of our plant based protein sources actually have a worse ecological footprint than some of our animal based protein sources. For fun, as an exercise to the reader, look at the amount water needed to produce a single gram of protein for each of the sources listed here. How much land is needed? How much water? How much pesticides? herbicides? How much does production contribute to eutrophication?  And last but not least, mow much fossil fuel is burned for production and transport? If we add these all up and look at the total ecological footprint PER GRAM OF PROTEIN, we get the outlines of an image where the idea that a plant based protein would implicitly when compared to an animal sourced diet would be implicitly good for the environment can be quite easily dismissed.   The baseline here is, looking at these sources, both plant and animal based, we should look for those sources of protein that are:

  • Relatively low in eccological footprint per gram of protein.
  • Relatively low in number of grams needed to achieve daily targets.
  • If possible provide us with valuable micro nutrients
  • If possible don’t mess with our carb/fat split.
  • Combines a wide spectrum of sources.

Well, for our top list, the steak is a no-brainer. It needs to go. So does the chicken and the fish if we look at the ecological impact. Broccoli, other than what some sources may want you to believe simply isn’t viable as a source of sufficient protein, so eat it, sure but don’t expect too much from it in terms of protein. So what sources should we add to our pallet that allow for a realistic daily protein intake program that fit with our concerns for health and our ecological concerns. Well, everybody will need to weigh these concerns themselves, but for me, weighing them, the following list came to be my list of protein sources:

  • Eggs
  • Mollusks
  • Insects
  • Crustaceans
  • Legumes
  • Nuts
  • Spirulina
  • Vegan blend protein powder

At the moment, Eggs and mollusks make up a major part of my animal based protein intake, due mostly to financial concerns and the fact that my wife doesn’t like me breeding my own insects in our home with little kids running around. These I think, at this moment in time are concerns everyone must look at locally: How ecological sound are your local sources of eggs? Can you breed your own insects ? Can you step over the cultural bias against eating insects that may or may not be strong in your community and may or may not have social aspects that might make eating or breeding insects socially awkward. If however you are truly concerned with the ethical aspects of food, than in many ways eating insects should be a massive part of your considerations. Not only are insects a source of protein that have an ecological footprint lower than that of many plant based sources of protein, considering insects as a source of food, rather than just a pest that needs pest control opens up possibilities for agriculture that might even make many plant based foods more ecologically friendly. While introducing foreign species is never a good idea, there are possibilities in many regions of edible natural pest control. If you look closely at the local ecosystem, chances are you can pair plant based food production with insect based food production in a way that reduces or eliminates the need for pest control. Thus, eating insects as part of a healthy high protein ovo-entho vegetarian diet or an entho vegetarian diet, should not only be superior to a vegan diet from a health perspective, and this is something my vegan friends are likely to want to contest, it is also potentially more ecologically sustainable if you can match up pest control with insects as protein source in your region.

And remember, our ancestors ate insects. Many cultures in the world eat insects. Even our so called vegan cousins, the great apes all eat insects. Insects are one of our natural food sources. A food source that is to our ancestors would commonly have been more abundantly and more effortlessly have been available than things like meat or fruit. If you can’t stand the idea of eating insects, ask yourself why. If you want to start with small steps in looking if an ovo-entho vegetarian diet is something for you, start of with a buffalo-worm omelet. Buffalo worms are not as crunchy as some other insects and their taste is really neutral. Once you are used to the idea, you can move on to meal-worms and crickets and maybe to whatever pest-control insects might be local to your region if you happen to be growing your own food.

This entry was posted on 3rd August 2016.

data-science versus system-engineering for nutrition.


Let me start this blog post with a big warning. On a scale from zero to one hundred, zero being vulkan style emotional detachment and 100 being a full out rant, this post is a fat ninety five.

Since I got interested in nutrition 6 years ago, after the disastrous effects that dietary advice and medication, given to me by medical and nutritional professionals, had on my personal health, and since I’ve been trying to apply my knowledge and experience with regards to data engineering in a forensics setting to my growing knowledge about nutrition, there have been some aspects of nutritional science that have led to a growing discontent with the whole field of science. To start of with I thought, well, I’m new to the field, so maybe I don’t get why things are done like this, but as my knowledge of nutrition, biochemistry and epidemiology has grown over the years, so has the seed grown that started of as the simple thought that there was something fundamentally wrong with the way nutritional science applies the tools handed to it by the field of data science. There are some great studies out there by people who clearly have a good grasp of data-sci, but many of the studies that come out, and many of the pop-science books about nutritional science, including those that are widely lauded as game changing seem to be using definitions of what constitutes solid proof that, as far as I can judge, align what the rest of science would agree only constitutes a possibly interesting link that may warrant further research. I will not be getting too technical as to not alienate the non technical readers, but I still need to touch on a few technical and data-science concepts.

Obsession with linearity

For some strange unexplained reason, nutritional science seems to have an obsession with linear associations. When I knew little about the subject and read nutritional science papers, after a while I was starting to assume that something fundamental about biochemistry made linear associations a fundamental truth of the field. After getting deeper into the subject however, there appears to be absolutely nothing to justify the obsession of large groups of nutritional scientists with the idea that associations need to somehow be linear. So what is a lineair relationship? Lets say there is a suspected association between the intake of for example a specific amino acid and mortality from a specific ailment, a study could look at intake of this amino acid in a number of regions and then look at mortality numbers and recorded cause of death in these regions. If you plot the per region mortality from our ailment against the average intake of our amino acid in given region, you get a scatterplot representing the relationship within our data set between the two variables. Well its a bit more involved than that as we will later see, but it is the base idea. Now if, as nutritional scientists seem to always do we define a straight line defined by:

Y=aX + b

In such a way as to minimize a the average of a simple function of the error between the line and the points on our scatter plot, this line and the errors as defined by our scatter plot define the hypothesis of  a linear relationship between X and Y. You can use other functions to fit against that are not linear, and you can, as other fields of science tend to do, fit Y against multiple variables at once using a random subset of the data and validate that fitting using the remaining data points, but for some reason these common techniques seem quite rare in both nutritional science and epidemiological science papers. Why is this? I really couldn’t tell. All I can guess is that there is some strange schism between these fields and the rest of science.


It isn’t just observational data where many scientists assume linearity, it is randomized controlled trails as well. This “golden standard” of research trails in nutritional science often leads to trail design that implicitly assumes linearity.






There are two problems with nutritional studies that aggravate each other. The first is the assumption that everything is lineair. The second are truly low standards and a wide spread misconception of the concept of p-values. The p-values issues seem to allow scatter plots that to visual inspection are clearly not best fit for ‘linear’ regression, to not only be falsely assumed to be linear, but even pass the arbitrary p-value threshold used to signify significance of that linear association.   So what are p-values anyway?

“The p-value is defined as the probability, under the assumption of the null hypothesis H, of obtaining a result equal to or more extreme than what was actually observed.”

While this definition is correct it is also the source of a whole lot of confusion that I wont go into for the sake of not going to technical.  The most important thing we shall need to focus on is the true p value versus the sample calculated p value. I am not a data scientist in the strict sense of the word. I’m not doing raw math most of the time. Instead most of my  data work revolves around writing and running simulations, many times over most of the time, from a data engineering rather than a data-science perspective. One of the things anyone who has run such simulations will know is that sample p-values can have rather erratic behaviour across runs. It came as no surprise thus when  Nassim N Taleb came with his meta distribution for p-values. I know a whole lot of people will probably stop reading here. Sure Taleb has a bit of an inflated ego and lacks the social sense to not refer to people below his intellectual level as imbeciles, but his work on the p-value meta distribution brings to data science what data-engineers have known through experience from simulations. The arbitrary threshold for the p-value as used by nutritional scientists is way to high if we account for the properties of the p-value meta distribution.  I’m probably already too technical here for most readers. The most important thing to take from this is that with a sample p-value marginally below the nut-science p-value threshold will have a very sizeable probability of having a true p-value well above the threshold value below what significance is denoted.  This apart from the fact that there is a general over reliance on p-values that isn’t just a nut-science issue though, and the widespread malpractice of p-hacking that appears to be unusually commonplace in nut-science.

missing variables


When looking into biochemistry and when studying actual data sets, there is one big chemistry variable that is omitted in most nutritional science data set: heat exposure. You don’t need more than a few months worth of chemistry lessons to know that exposure to heat can make a whole lot of difference. Studies looking at things like macronutrient consumption though hardly seem to record heat exposure or even differentiate between something as crucial as consumption of unheated PUFA versus consumption of PUFA in cooking oil used for baking stuff. Another often omitted variable is lean body mass. While loss of body fat, especially of the visceral type,  is beneficial to most people in western countries, loss of lean body mass is really bad news. Too many studies that record weight change however fail to also record body compositional values.

sub-population variance and skew

skewness-and-kurtosisFor epi studies there is quite a different type of missing variable problem. If we accept that an association is often nonlinear, and knowing that the average for a sub-population is just that, a sample average, we really need to pull some more advanced techniques out of the hat than just work with sample averages. The point is, we can’t just assume all sub populations to only differ in sample mean while sharing the exact same variance. Next, we can’t assume the sub-populations distribution to be skew-less.  Again these are a bit technical terms so lets look at them a bit. If a true association between a nutrient or food and mortality from a specific ailment is shaped like a bathtub, then many of the deaths in a population might arise from one of the extremes of the sub-population distributions. The volume of people contained in these extremes depends on distribution properties like variance, skew and possibly kurtosis that together define the shape of the distribution. If these moments differ between sub-populations, then any conclusion drawn from looking at the population through just the sample mean becomes basically something that is bound to yield misleading results. When recorded, differing variance across sub populations, or sub distributions with significant skew or kurtosis should be reason to call in the big guns.

Static versus differential markers.

Now let’s assume nutritional scientists found a solid association and got the shape of the association right. Let’s say higher levels of marker X are strongly associated with higher levels of mortality number Y. Does that imply that a negative ΔX for an individual will lead to a negative Δ in P(Y) ? Well, maybe, but then there might even be a possibility that it will lead to a positive P(Y). Point is: a marker, even a good marker isn’t necessarily a suitable variable to use as controlling variable for Y. For an engineer these things are blindingly obvious. For many nutritional scientists though, the assumption that all that is needed for a good controlling variable is a  strong correlation seems to be the default assumption.

surrogate endpoints and other die-hard misleading variables

In nutritional science there is one step beyond the wrongful use of  markers as control variable: surrogate endpoints. Nutritional science has an abundance of questionable surrogate endpoints that are stand ins for real endpoints (mortality) for different diseases. Things like LDL-c, BMI and high blood pressure. I’ ve talked about LDL-C and BMI in earlier blog posts and won’t rehash them here now. The big problem with surrogate endpoints is that they persist and contribute to poor follow-up studies. For example, using BMI as surrogate, one might conclude that an increase in lean body mass (LBM) would increase the risk of CVD-death. This while all data shows the actual reverse to be a very viable hypothesis.

Adding it all up


If we look at all the issues discussed above, we could ask if nutritional science is fixable. Looking at the actual data sets from many studies with a pair of data-engineering spectacles, we must come to the conclusion that from a data-sci perspective, most data is very much inconclusive with respect to the claims being made. If we fix nut-science, the field should become quite boring with very few claims being made either by observational or by RCTs. It seems the influence of individual components of nutritional profiles on mortality numbers is either very small or non existent. Possibly when missing variables such as heat exposure are added there might be interesting claims that could stand up to modern data-sci scrutiny and that is something that should definitely be looked at. But who is going to fund studies when nine out of ten studies results in inconclusiveness. So what other options are there?  While scientists tent to look down at the idea of N=1 stuff, engineering thrives on it. Using engineering practices, especially multivariate control-feedback theory, an individual might find his/her personal ideal nutrition and exercise program all by him/her-self. Do we still need nut-science then? Yes we do, but only at the extremes basically. An N-1 control-feedback loop can’t obviously factor in mortality numbers into the feedback loop, yet individual progress might lead to extremes (100% meat diet, 100% bananas diet) that in turn might result in sudden death of superficially healthy individuals.  If we accept that nutritional science hasn’t got that much useful + conclusive to say about the mid-range of most foods and nutrients, and if we accept the non-lineair nature of many things, including nutrition/mortality associations, the field could focus on the extremes as an important way of supplementing what should be a primarily N=1 engineering effort aimed at individuals.

So that rounds up what I already warned was a bit of a rant. I hope I’m being way to pessimistic with all this, but the more nut-sci papers I read and the more data-sets I actually end up being allowed to look at, the more pessimistic I’m growing about this field of science that as it seems can only excist and can only make funding-worthy claims as long as it keeps ignoring what data-science has to offer. So as it is my conviction that we truly need to develop the field of nutritional engineering to fill in the gaps that nutritional science is failing to fill, sadly just keeps growing with (almost) every paper I read.

This entry was posted on 1st August 2016.

An engineers vegan food pyramid.

A while back I wrote a blog post describing my engineers food pyramid that came to be using a combination of some base data engineering on raw data from epidemiological studies, end the extensive use of N=1 control-feedback-loop experiments. Whenever both data engineering and control theory are suitable, I went for control theory first. Whenever nutritional science has different theories with conflicting resulting advise, I tried to see what theory the epi data showed to be least viable. Each control-feedback-loop experiment takes me at 6 to 18 weeks, and some experiments have come at the expense of my own health. The last experiment was a 12 week run of a 15% carb 15% protein 70% fat diet without carb-cycling, that my health didn’t agree with. I both lost some strength and gained a substantial amount of body fat. As I have stressed many times, everyone is different, many people are doing amazing on a high fat moderate protein keto diet, and control-feedback-loop based dieting by definition is an N=1 solution aimed to find the best diet tailored on you as a unique individual. My main approach: trust the data. There are many wild nutritional theories out there, but many can be easily shown either wrong or inappropriate for you as an individual with some basic control theory and data engineering effort.  As I have described in multiple other posts, I use two health markers (weight relative body under-strength and age/gender adjusted over-fatness) that I combine into a vector that in turn I use to calculate a desired delta. By using the vector difference between desired delta and actual delta, we get an error vector that can be used as error in a control-feedback loop. While most dietary theories come with their own shortcomings, as I outlined in my previous post, the most disastrously wrong theories all seem to be linked to veganism sub-cultures. People become vegan for different reasons, and some of these reasons are absolutely important to consider.

  • serious ecological sustainability concerns.
  • Ethical concerns regarding cruelty to animals.
  • Ethical issues with killing animals.
  • Health concerns.

For those of us that, like me, are concerned about the environment, animal cruelty and health concerns, but don’t have ethical issues with killing animals, especially if these animals are just insects, replacing most or all animal products with insects, if you can get passed the western aversion to eating bugs, should be a serious option to consider. Personaly, my favourite nutritional pattern could probably be described as ovo-ento-vegetarianism. However much I would like to promote ovo-ento-vegetarianism, as I have done a few times before, in this blog post I want to share my dietary findings with those of you who share some of the above concerns and either have ethical issues with or an uncontrollable aversion against eating bugs.  So here it is:


If you have been a vegan for a while, chances are you have been exposed to books and publications that will disagree strongly with some of the things I’m presenting here. Some sub cultures of vegans have put their trust in gurus, questionable scientific theories and scientific organisations with strong ties to rather extreme animal rights groups. If you are a big fan of Dr Campbell, the PCRM or Freelee the banana girl, than what I propose here may against many of your current preconceptions. But please bare with me while I try to explain the ideas behind this alternative food pyramid. Everything if data based first and theory based second.

One important thing to note about this pyramid is that it assumes you are working out; lifting weights, and doing so at least three times a week.

layer 1: Protein, Micro-nutrients, fluids and feedback-loops 

At the basis of our pyramid we a set of important concerns that should be considered pivotal:

  • The engineers approach to diet and workout dictates we use a control feedback loop approach to our diet. No two persons are the same. Measure your results as objective as possible (in our case using the Generic Body Health Index  and its differential.
  • The most static part of our diet is protein. You should stay within a daily protein intake bandwidth ranging from 2.0g to 2.6g per kg of lean body mass. You should do this in a way that maximizes the diversity of amino-acids.
  • An other relatively static part of our diet is micro-nutrient diversity.
  • Where possible, we should try to get at least part of our protein using protein sources that also provide  relatively high levels Omega-3 fatty acids.
  • Where possible, our micro nutrient rich foods should also be sources of Omega-3 fatty acids.
  • Finally, sufficient fluids are an other foundational ingredient of our diet.

So how do we achieve these levels? We will combine supplements with whole-food plant sources for our protein intake. Nuts are both a source of healthy fats and of proteins. The ratio’s of fat versus protein in nuts and diary are higher than what is suitable for our protein requirements and total caloric needs, so we need  to turn to add lean low-carb sources of protein. Unfortunately the choice for whole-food low-carb vegan protein sources is close to non-existent.  Yes I know. Protein-powder could be considered processed food, and processed foods in general are a bad choice. The problem though is that plant based sources of protein tend to both have a to low protein/calorie ratio to satisfy our protein needs, and also lack beneficial diversity in amino-acid content. Fortunately the quality of mixed-source vegan protein powders is absolutely excellent on all accounts.  The foods discussed so far aren’t just sources of protein. If we choose them carefully and consume the whole spectrum, they should provide for a wide range of amino acids, part of our Omega-3 needs and part of our micro-nutrient needs. We should absolutely add crushed flax seeds to the diet as they are high in Omega 3 and compensate for not eating fish or fish based oils.There are other essential micro-nutrients missing though. To supplement these, and to address our fluids requirements, we add veggie juice and water to our foundational layer. Take it easy on the starchy tubers though and add berries for taste and nutrient spectrum. We want to maximize the micro nutrient/carb ratio of our juice, at least at this level. Finally there are two more essential suplements that every vegan should take: Carnitine and vitamin B12.

Layer 2: Fat & Veggies

Once we have our foundation secured, we can look at adding (mostly) fat and veggies up to a level slightly above our bodies base resting metabolism. We add veggies, spices and dark chocolate for additional micro nutrient diversity and additional dietary fibre. The fat part should mostly be mono-unsaturated fat, but saturated fats are a decent supplement when we don’t overdo it. When you cook, don’t use unsaturated fats, use coconut oil instead. The saturated fats in coconut oil may be worse than mono saturated fat when unheated, heating fat changes the rules of the game. Oxidation of unsaturated fats makes them rather unhealthy, using coconut oil should help there. We already covered (most of our) Omega 3 needs in our foundation. Stay away from processed oils, oils made from GMO crops and fat sources high in Omega-6. No Canola oil or anything like that. Being a vegan, chances are high that your cholesterol levels are pretty low. Good for you! Or is it. There are actually real and serious health risks that are linked with LOW cholesterol. If your cholesterol is anywhere below 150, this should be a serious concern and eating more coconut oil, that is high in healthy stable saturated fatty acids could quite probably reduce many health risks. So don’t be afraid of saturated fats. Don’t be afraid of coconut oil. Use it liberally when cooking. It may do little for non vegans who get plenty of SFA from animal sources already, and causality may not have been confirmed yet for many correlated health issues, but why artificially keep your cholesterol at extreme low levels when we know there are strong links? Why take the risk? The research simply isn’t there yet, but that doesn’t imply it will be safe until the research is finally there. There is a sizeable probability that low cholesterol levels are causal for many health issues and mortality numbers. Don’t put yourself in dangers way and eat plenty of coconut oil prepared dishes.

Layer 3: Timed carbs

You may have noticed that up to now everything has been quite in line with low-carb discipline. Fat, protein and lots of micro-nutrients with relatively little carbohydrates. The point to this has been: You truly don’t need carbohydrates if you are sedentary. When you want to do a high intensity strength oriented workout, like the one my ebook will be advocating, than you really really need some well timed carbs. But when you do take carbs, its a good idea to time these carbs perfectly relative to the time of your workout.  Much of the carbs from our veggie juice can be included in our carb timing on our workout day, but chances are these won’t be enough.  We add more sources of carbohydrates to the timed layer of our diet. We do this in three parts:

  • Starchy foods in our pre-workout meal
  • Fruits as pre-workout snack
  • Coconut juice mixed with water and possibly part of our veggie juice as sports-drink for consuming during our workout.

It is possible that your personal control feedback loop will add some post-workout carbs also, but the core concept of this layer of the pyramid is that most of your carbohydrate intake is concentrated around your workout. On resting days your diet should most probably be a low-carb diet. There seems to be absolutely no benefit in high carbohydrate levels for sedentary days. If you do consume high carbohydrate foods on resting days, do so in limited quantities and prefer starchy foods over sugary foods. No sweat fruits or fruit juice and certainly no dried fruits.

Layer 4:  Use sparingly

A food pyramid wouldn’t be complete without a ‘use sparingly’ section. While processed foods of most types should be considered off-limit, a diet meant as a way of life instead of a short period of suffering, can only be sustained if you allow yourself to indulge yourself in a bit of the bad stuff. In this diet the bad stuff that is acceptable for occasional indulging are sugar, grains and non-processed oils rich in Omega-6. You shouldn’t make consuming these foods to much of a habit, but banning them completely won’t make you popular at parties. So if you are invited to a lunch, have that bun of bread, or at that party that piece of pie. Unless you have a really big family and end up eating pie twice a week that is 😉

This entry was posted on 1st April 2016.