Minimum Effective Dose

In part to support my goal of a metrics-driven lifestyle, I've begun rigorously tracking my running performance. One motivating factor is to ensure I'm making progress so I don't waste months spinning my wheels with nothing to show for it. Another is to use evidence of previous improvement to drive myself through tough, demoralizing stretches. And the third reason is to identify my minimum effective dose (MED) for improvement.

Tim Ferriss defines minimum effective dose (MED) as "the minimum dose that will produce a desired outcome." Here's an analogy he uses to describe it:

If you need 15 minutes in the sun to trigger a melanin response, 15 minutes is your MED for tanning. More than 15 minutes is redundant and will just result in burning and a forced break from the beach. During this forced break from the beach, let's assume one week, someone else who heeded his natural 15-minute MED will be able to fit in four more tanning sessions. He is four shades darker, whereas you have returned to your pale pre-beach self. Sad little manatee. In biological systems, exceeding your MED can freeze progress for weeks, even months.

He then uses this in the context of exercise to show how excessive weightlifting does not increase strength over time but rather invites the possibility of fatigue, mental burnout, and injury risk. I've appropriated the same concept to examine my running habits.

Here's the thing: I don't want exercise and fitness to become a primary goal in my life. Since I'm trying to narrow my focus and only focus on a small number of projects at once, treating running like a "big project" will just sap energy from my other intermediate-term goals, like becoming better at my job and learning how to ski. Rather, I'd like it to become a habit that I perform without having to think too hard about it. Therefore, I'm seeking to understand my minimum effective dose for running frequency, which I've defined as "the minimum running frequency that will ensure improvements in proficiency over time." Ignoring for a moment the fact that everybody has good days and bad days, I want to identify the maximum number of days I can allow between runs while still increasing speed and distance each week.

To answer that question, I look carefully at my running logs. Even after only three weeks of tracking this way, I've gained some insight:

• Between my 2/15 and 2/20 runs, my course time (on the same course) dropped from 51:00 to 48:30. This means that five days between runs were sufficient to realize performance gains. My MED is at most a 5-day frequency (meaning that I could probably get away with going 6 or 7 days between runs and still seeing improvement.)
• Due to travel & work conflicts, I went 13 days between my next two runs (2/20 and 3/5). I actually ran a different course these two runs so I can't compare times directly, but I can subjectively state that the run did not go well. My feet felt raw and blistery, I was sluggish, and I felt weak and soft overall. Combining that information with the common-sense notion that "thirteen days is probably too long between runs to see consistent improvement," I can semi-confidently state that my MED is at least a 13-day frequency. So now I know my MED is likely somewhere between six and twelve days, and subsequent runs will be scheduled to measure it more precisely.

It's important to remember that MED will change over time. Incremental improvement eventually takes exponentially more effort (like how I lowered my mile time by 20 seconds last week but elite track & field athletes work for months to shave off fractions of a second.) But the goal is to use MED to schedule runs that ensure improvement without dominating my lifestyle. If, for example, my MED turns out to be 8 days, I'll run once a week until gains start to taper off.

Is this the right approach for marathon training, or to become a world-class sprinter as soon as possible? Definitely not. But I believe it's a reasonable methodology for avoiding effort that does not yield results.