Dec 28

The Big Reveal is Tomorrow at Noon

Well, here we are. My doctor has the final labs in hand for my longest, most ambitious experiment yet. I’ve instructed him to not reveal it in any way — not even to hint at what it might say.

Over the last few months, I’ve told many doctors and researchers of my planned experiment for December and some of them got to see me doing it up close (such as Drs Jeffry Gerber and Steven Horvitz) when meeting me for dinner. I’ll give the longer details on it in a later post, but needless to say, there were some unexpected curveballs that weren’t entirely under my control. All in all, it was about a month of work including the washout period.

Now I have confirmation all labs are final, although the NMR didn’t make it (long story). But that’s okay as I made sure to get a standard lipid panel, which will give me the metrics I was most focused on demonstrating change: Total Cholesterol, LDL-C, HDL-C, and Triglycerides. Thus, as per my original plan, I’ll be revealing this data to myself and everyone else LIVE.

I’m setting the reveal for tomorrow, December 29th at 12pm PST (3pm EST). I’ll be opening it on a Facebook Live feed from my page.

(I’ll also separately film the event and post that to my YouTube channel at a later point.)

I’ll concede this is the only test I didn’t want to immediately see and add to my spreadsheet. I’m happy enough waiting until tomorrow as it could very well mean I’ll have another month-long experiment or two to do if the data doesn’t turn out as I’d expect. This one experiment was pretty challenging, yet achingly monotonous and tedious at the same time. And I might be completely right on my Phase II hypothesis, just wrong on the presumption of top-off glycogen points or things associated with timing (there’s really a lot of variables to control for). Anyway, any data collected is still better than no data at all, so whatever happens, I’ll hopefully have some valuable new clues in hand as I move forward.

Dec 26

“We regret to inform you…”

Today is not a good day.

I just got a call from the lab to let me know that one of the tests didn’t work out. As she opened with this statement, my mind immediately raced to the one test above all others that I cared about, “please don’t say the NMR, please don’t say the NMR, please don’t say the NMR!!…” I was getting CMP, CBC, fasting insulin, hsCRP — any of those I can live with, just not the NMR!

It was the NMR.

I genuinely felt a rush of adrenaline hit me. This was terrible news. I began pacing as I talked to her on the phone. I could tell my fear and frustration coming through my voice, but I tried to contain it.

“Just come back this week and we can retest.” she offered. I’m sure from her perspective, I was probably just annoyed I’d have to travel there and get stuck one more time by the needle. But of course, it was something much more substantial than that. The reason the month of December was so relevant is that this was the finale test of the longest and most controlled experiment I’ve done to date. And once I got that blood in the tube Saturday morning at the lab, I made good on my promise to the family that I was done with the experiments and I could eat whatever I wanted, which made their lives easier as well.

Including the washout period, this was close to four weeks of work. And while I won’t reveal what I was eating yet (that will be a very big post), I can assure you my food logs look like an Andy Warhol painting, which is to say, very boring and repetitive to isolate the variables. Lots of time, money, and exhaustion setting up these tests.

I know it was an honest mistake and I truly have a lot of respect for phlebotomists and the lab that turns around these tests. But there’s no question it just sucks. Probably some of the worst timing I could imagine.

The silver lining? I wisely also included a basic lipid panel, so at least I’ll still have the toplines of TC, LDL-C, HDL-C, and TG. While not the particle subfractions I need for the deeper data, I’ll still get plenty of value from these, regardless.

Dec 09

December Ninth, Twenty Fifteen


On December 9th, exactly two years ago, I received a document that would change my life forever. It was an advanced cholesterol test known as an NMR (Nuclear Magnetic Resonance).

I had only taken one before that, about two weeks earlier at the end of November. That first test had put me into a depression. I saw that my LDL cholesterol (LDL-C) had skyrocketed from my average of 130 to 239, and my LDL particle count (LDL-P) was a whopping 2,705. I had never measured my LDL-P before then, but I knew the “reference range” for it was below 1,000.

I decided to schedule the December 9th test, hoping it might suggest this first one was a lab error. And for 15 long days, I dropped everything I was doing to read and learn all that I could about cholesterol and the lipid system that trafficks it. Surprisingly, I found this system had quite a lot in common with networks in my own field of software engineering. In fact, I started to wonder if I was simply projecting my own beliefs where they didn’t belong, as a way to cope with this miserable time.

When my results came in for the December 9th test, there was a resounding “click” in my head!

You see, while I was experiencing this terrible depression, my appetite waned. I resolved I wouldn’t eat one more bite of saturated fat than I was truly hungry for. This led to my eating about 1/3rd of the calories I normally do, even while still being on a ketogenic ratio. So overall, my dietary fat dropped significantly. Thus, if my dietary fat was way, way down — then my LDL-C and LDL-P would have presumably dropped, right?

Instead, my total and LDL cholesterol had gone even higher!

Now I realize this won’t make sense to a lot of you, but that was very relieving. As an engineer, I immediately understood two things:

  1. This was indeed an energy distribution system above all else. Cholesterol is a passenger, not a driver.
  2. I had a lot of work ahead of me!

Here I am a couple years later, and I can remember that moment like it was yesterday. I was alone and determined when this all began, but now I’m proud to share this journey with so many others who are helping to change the paradigm.


Nov 29

Lowering Cholesterol as a LMHR


As our resident Lean Mass Hyper-Responder (LMHR) on the site, I have been an outlier so far with my Lipids. Despite an additional 200g/day of fat, my LDL was effectively unchanged in the Ketofest experiment. I knew exercise was potentially a confounder given Dave’s experiments with distance running, but I had a full training schedule this year leading up to my first marathon earlier this month.  Now, with a fat-fueled, goo-free, finish in the books, I decided to overlap some science with a forced period of rest.


The Inversion Pattern is a theory about the impact of short-term diet on blood lipids (TC, LDL, HDL, Triglycerides). For LDL cholesterol, the supposed “bad” kind, it can be simply stated as:

All else being equal, LDL-C is inversely related to the dietary fat consumed in the 3-days prior to the blood draw

The implications of this pattern are profound. If blood lipids are not static, but instead highly dynamic from short-term factors, all conclusions based on a single sample are suspect. It would be like concluding you lived in a desert from a single sunny day.

Here’s the complete set of relationships from the Inversion Pattern:

  • Total Cholesterol tracks with inverse of dietary fat for the 1-3 days before the blood draw. (87% inverted correlation)
  • LDL-C tracks with the inverse of dietary fat for the 1-3 days before the blood draw. (90% inverted correlation)
  • LDL-P tracks with the inverse of dietary fat for the 3-5 days before the blood draw. (80% correlation)
  • HDL-C tracks with dietary fat for the 1-3 days before the blood draw. (74% correlation)
  • HDL-P tracks with dietary fat for the 3-5 days before the blood draw. (correlation not calculated)
  • TG tracks with the inverse of dietary fat for the 1-3 days before the blood draw. (61% inverted correlation)

Note: Correlation numbers are based on Dave’s data through 4/8/16 and don’t include subsequent experiments where additional variables were introduced.


The goal of this experiment was to replicate the Inversion Pattern using the Feldman protocol.  If that held, I would use this data as a baseline to compare against my earlier tests that included exercise. And, just for fun, I would be able to see the short-term variability of a number of other metrics in a controlled setting.

Experiment Design

I adapted the 4-test version into a 2-week schedule following my race. The first two days were a washout period with ad lib food. That was followed by a 5-day low-calorie phase and a 5-day high-calorie phase. Throughout, I refrained from all exercise. I even switched to riding the subway instead of my normal bike commute.

For measuring cholesterol, I got both a Nuclear Magnetic Resonance test (NMR) for particle counts, and the standard lipid panel (for redundancy).  I followed my own advice on testing and also got hsCRP, Insulin, Metabolic panel, and WBC.


To evaluate the performance of the Inversion Pattern, we’ll score it based on how it does predicting which test shows the highest and lowest values for each of the metrics.

Here are the predictions:

Lipid Score Highest Test Lowest Test
Total Cholesterol (TC) Test 2

From days 5-7 (Low Cal x2, Fast)

Test 3

From days 8-10 (High Cal x3)

LDL-C Test 2

From days 5-7 (Low Cal x2, Fast)

Test 3

From days 8-10 (High Cal x3)

LDL-P Test 2

From days 3-5 (Low Cal x3)

Test 4

From days 8-10 (High Cal x3)

HDL-C Test 2

From days 5-7 (Low Cal x2, Fast)

Test 3

From days 8-10 (High Cal x3)

HDL-P Test 4

From days 8-10 (High Cal x3)

Test 2

From days 3-5 (Low Cal x3)

TG Test 2

From days 5-7 (Low Cal x2, Fast)

Test 3

From days 8-10 (High Cal x3)


The Experiment

Food Journal

Overall, I managed to stick to the schedule pretty well.  I managed a 40-hour fast on Day 7 during the low-calorie phase (calories: a coffee with cream and a spoonful of fish oil). On Day 11, I had some unexpected family travel, so it ended up being a Medium Calorie day. Fortunately, there were enough blood draws to still get a sample after three High Calorie days in a row.



My go-to meals during the experiment, easy to vary the quantity.

  • Breakfast: Bacon and Eggs (with HWC, butter)
  • Lunch: Bunless Burger salad (with Olive Oil)


Total Cholesterol (TC)

First, let’s look at Total Cholesterol.

Finally, Feldman drop achieved! As predicted, TC peaked on Test 2 after the 40-hour fast.  Three days later, TC dropped 112 points to its minimum value. Interestingly, the Medium Calorie day allowed a difference in dietary fat between Test 3 and 4, which showed a small increase of TC, as predicted.  

Prediction Score: 2/2


Next, let’s look at LDL-C, the “bad” cholesterol that everyone is worried about.

Similar to TC, LDL-C peaked on Test 2 after the 40-hour fast.  Three days later, LDL-C dropped 100 points to its minimum value for Test 3, as predicted.

For comparison, my normal LDL-C with regular exercise was in the 210-230 range. For those keeping score at home, that’s a 200 point difference from the peak of this experiment.

Prediction Score: 4/4


Next, I was eager to see LDL-P. In Dave’s results, LDL-P has both a higher correlation and a trickier formula.

Boom. Exactly as predicted. From Test 2 to 4, LDL-P dropped by an impressive 1200 in 5 days. I found it particularly interesting how large the change was between Test 3 and 4 in LDL-P, even though the LDL-C went in the other direction in that interval. Clearly, LDL-P is a lagging indicator from LDL-C.

For comparison, my normal LDL-P (from one NMR in August) was 1597.  In this case, the total net difference was 1700. Without medications.

Prediction Score: 6/6


On to HDL-C the “good” cholesterol.

At first glance, HDL-C doesn’t seem to move much. However, since going low-carb, my HDL-C has been in a pretty tight 110-115 range, so 102 was actually a low-outlier, as predicted.  The two highest fell within my normal range and didn’t appear in Test 3 as expected. Maybe it takes a little bit more time to spin up my reverse cholesterol transport to above 110.

Prediction Score: 7/8


After the first prediction miss on HDL-C, my expectations were lower for HDL-P.

Wow. HDL-P moved from 31.3 to 36.9 exactly as predicted by the Inversion Pattern. Even if the cholesterol cargo in HDL has some additional variability, the particle counts (and thus emission and/or clearance rates) clearly have short-term diet as a driver.

For comparison, my one previous NMR in August showed HDL-P of 34.7.

Prediction Score: 9/10

Triglycerides (TG)

Triglycerides also played along and followed the 3-day fat intake as predicted.  At first I thought the higher TG in Test 1 and 2 might have been due to the lingering expectation of energy needed for exercise, due to my sudden drop off in activity. However, their correlation with 3-day fat matches the Inversion Pattern and implies a more responsive dynamic.

For comparison, my TG tests have typically ranged from 50-70, going back to before my low-carb diet.

Prediction Score: 11/12


Insulin is a key hormone that I was interesting in following during the protocol.

Here I plotted against 3-Day Fat, but protein was similar. Despite keeping Carbs generally below 40g (except days of 73g and 56g), I still managed to have a 6x change from min to max. Clearly, Insulin is responsive to non-carb nutrients. I wonder how much of the later results were driven by late-night dairy (e.g., 4-oz of half-and-half at 10pm).

For comparison, my Insulin from August was 4.3.


I tracked High Sensitivity C-Reactive Protein to see how inflammation would track as I recovered from my first marathon.

I was extremely sore for several days following the race. I was feeling better by Test 1 on day 6, but still had somewhat higher inflammation. As expected, inflammation declined as I rested and healed.  My previous CRP reading in August was 1.01, so it’s possible that the 0.89 minimum was because I excluding some inflammatory food from my diet. Dairy is the most likely culprit, and it would be interesting to exclude it and check my CRP again.


This experiment was a clear confirmation of the Inversion Pattern for all lipid scores except HDL-C. Short-term diet was demonstrated as an input factor was able to move LDL-C by 100 points. In previous tests, exercise during the 3-days prior was sufficient to confound a similar change in diet. In addition, infection, injury, and other short-term inputs have known or theoretical impact on these numbers.

One of my takeaways is that the lipid system is much more dynamic than conventionally understood.  For those who are getting advice to change their lifestyle or take medication from someone ignorant of this fact, I would encourage you to learn more first. For those who need a low LDL or TC score to get cheaper insurance or to get their doctor off their back, I would encourage you consider using the protocol yourself.

Another takeaway is that my blood lipids are quickly reacting to serve essential bodily functions. Some of the missing LDL-C from my exercising tests were likely used for muscle repair via endocytosis, reducing my recovery time. In addition, the four hours of marathon running without goo was only possible because of VLDL distribution (and possibly even HDL to some narrower vascular channels after CETP). I’m grateful I have a “hyper-responding” lipid system to get the job done.

In the future, this data should prove valuable if I decide to continue with further experiments. They might have to study the impact of exercise, though, since the prolonged rest was difficult for me.

Nov 25

Permission Pathways – How We Justify Cheating

For two and a half years I’ve been eating LCHF, save the times I’ve left it for experimental purposes. Generally, I’m pretty staunch on the macronutrient requirements, but I also try to further “do better” on things such as reducing processed foods and items high in polyunsaturated fatty acids (PUFAs).

To give an example, I got hooked on Quest Chocolate Chip Cookie Dough bars. Their ingredients weren’t great, given how processed they were (as is the case with most “bars”). And while it seemed acceptable as a Sometimes Treat, I certainly wasn’t eating them infrequently. I was having one or two of them daily in the week I discovered them, and I noticed my anticipation of eating them was very high throughout the day. So I resolved to quit them cold turkey and haven’t had one since April.

The Breach

While visiting family last week, I found myself falling asleep late one evening. I had left my toothpaste and floss back home, so I fretted about not having either when winding down for bed. As it happened, there was a 24-hour convenience store across the street from where I was staying, so I decided to jaunt over to get what I needed.

I came back with a Quest Chocolate Chunk bar and a Diet Mountain Dew, both of which I had no intention of buying in the first place. Worse, I decided to have them right then and watch lipid YouTube videos for the next few hours instead of the original plan to get some healthy sleep.

So what happened?

If you read this blog, you know I’m pretty disciplined, particularly with my experiments. But while my methodical discipline has me recording 99.999% of everything I ingest, I often point out my diet can be 1/3rd restaurant and fast food, and that I can still struggle with commitments such as these.

Permission Pathways

Fortunately, my overactive mind likes to break down how and why these motivations change and what was involved to make it happen. I call these Permission Pathways because they are often identifiable and usually involve more than one.

In short, Permission Pathways are the series of excuses as to why one would break a commitment made — particularly to a specific diet or lifestyle.

All of these have two important things in common:

  1. They are considered a justification for breaking the original commitment.
  2. They are often assumed to be a temporary — an “exception” that is usually circumstantial.

Sensory Permission

  • “I was passing by the donut shop and the glaze just smelled so good. I need to quit walking that route!”
  • “I heard someone eating nachos at the table behind us and it sounded good. I decided to order one.”

Description: Obviously, our five senses of Sight, Smell, Sound, Touch, and Taste come into play as powerful activators. And naturally, this is one of the most common excuses for breaching a commitment.

However: While the most common, this is the least compelling reason to break the commitment. You’ll likely be in proximity to foods that tempt you the more common they are, and thus your senses will be engaged. But if there are certain foods that are especially activating, you may need to avoid them altogether.

Amended Quantity Permission

  • “Well, I’ll let myself have just three of these…”

Description: A limited, seemingly small number is declared in advance of the action. This makes it seem as though the cheating is more acceptable given there is still some limit in place.

However: The justification is that there is still “a limit” being imposed by this new declaration, but this new limit is really exactly the opposite — it is to ease the existing limit you had in the first place. This is especially bad as it is the easiest to apply in any given situation.

Event Permission

  • “It’s Christmas. I’m only going to be bad until it’s over.”
  • “It’s a concert, I should be drinking anyway.”
  • “I can’t turn down Girl Scout cookies. Good thing they only sell them at this time of year.”

Description: An event or circumstance of some kind leads you to consider it a “special occasion” and worthy of breaking the commitment.

However: This is problematic for two reasons. First, it is often tightly defined what events warrant the break, but gets looser over time (i.e. “Originally I allowed it for just Holloween, but I figured it might as well be all holidays…”). Second, it is often tempting to extend the period of time in question (i.e. “At first I meant Christmas day. But now it’s pretty much Christmas season.”)

Influence Permission

  • “I only get eat this way around Ted.”

Description: This is really more about excusing behavior due to being in contact or hanging around a “bad influence.” The actual statements that influencer makes can be any of the Permission Pathways listed here.

However: You’re acknowledging one or more specific people (or groups) influence you to break your commitment. This can be intentionally or unintentionally on their part. But regardless, you acknowledge this influence as counter to your commitment, suggesting you should either limit where you go with them or contact with them altogether.

Fortunate Circumstance Permission

  • “Oh, look at that! They never make peanut butter cookies here, I should take advantage of it.”
  • “Okay, there’s no way a gas station like this would have a Vanilla Coke. But if it actually does, I’m going to have to get one.”

Description: Some situation or location appears to have an unusual opportunity, or at least, that’s how it is justified.

However: You must recognize these are often attempting to turn cheating into a game, with the “lucky chance” encounter as equivalent to winning. In a roundabout way, you’re making this chance outcome its own reward, which it certainly isn’t.

Earned Commitment Breaking Permission

  • “I think I deserve a break from the diet given how many greens I ate over the last week.”
  • “I’ve been perfect for six months. I think I’ve more than earned a night of unlimited cake.”

Description: The commitment can be broken temporarily as a “payment” for a certain amount of good behavior leading up to it.

However: This is typically considered the most acceptable of justifications for going off a diet. But it is one of the most dangerous given it will likely be a reoccurring excuse for future opportunities as well.

Revaluing the Commitment

  • “I know I said I wasn’t going to go over 50 carbs a day, but that was actually pretty conservative anyway. Keeping below 100 carbs a day is probably more than enough.”

Description: This is a re-evaluating of the commitment in the face of an enticement, minimizing the original goal as unnecessarily high.

However: It is likely your revaluation is specific to the moment and the thing you wish to ingest right then. So if you’re honest with yourself, this long-term choice was brought on by a short-term temptation.

Lemming Permission

  • “Everyone here is having this tonight.”
  • “I guess we’ll all having this for lunch… I suppose it would be rude if I didn’t as well.”

Description: A large number of people in a social situation becomes the key rationale for breaking the commitment.

However: This will clearly be an ongoing issue with anyone who finds themselves in social situations. As with any other diet, you should be able to simply insist your needs are specific and that you can’t participate in something that doesn’t meet them.

Expert Opinion Permission

  • “Claire said it would be fine if I eat these and she’s a nutritionist!”

Description: Having an opinion from someone known as a reason to break the commitment.

However: Of course, the so-called “expert” is often someone who neither knows the full context of your commitment or the progress thus far. The opinion shouldn’t have much impact unless you’re consulting them to reevaluate everything you’re doing in the first place. And once again, this shouldn’t be happening at the moment you wish to cheat, it should be in a session all its own.

Mutual Broken Commitments Permission

  • “Well, if Dale is eating off his diet then I’m eating off my diet.”

Description: Having one or more people engaging in commitment-breaking leads to justification to do likewise.

However: Someone else’s lack of ability to meet their commitment shouldn’t be something you want to emulate.


Older posts «

» Newer posts