Regression to the Mean

Regression to the mean is a statistical phenomenon. It states that if a variable is extreme on its first measurement, it will tend to be closer to the average on its second measurement.

It sounds technical. But it explains why we believe in magic, curses, and useless medicine.

The Symptom: The "Sports Illustrated Jinx"

There is a famous myth in American sports: The Sports Illustrated Jinx. The belief is that appearing on the cover of the magazine is bad luck. Athletes almost always perform worse the following season.

Is it a curse? No. It is Regression to the Mean.

To get on the cover, you must have an outlier season. You must perform exceptionally well—better than your average, and better than everyone else. It is statistically unlikely you will repeat that "perfect storm" of performance. The next season, you don't "get worse" because of a curse. You simply return to your normal average.

The Mechanism: The Medical Trap

This phenomenon traps doctors and patients constantly. It is the primary reason why anecdotal evidence ("It worked for me!") is dangerous.

The "Doctor Visit" Effect

1. The Peak: Patients go to the doctor when their symptoms are at their worst. When the back pain is unbearable. When the flu is peaking.
2. The Intervention: The doctor gives a pill, performs a therapy, or waves a magic wand.
3. The Regression: Because the pain was at an extreme peak, it was likely to improve naturally the next day anyway. It regresses to the mean.
4. The Fallacy: The patient thinks: "The pill worked!"

Reality: Time worked. The body worked. The pill might have been a sugar pill (placebo), but it gets the credit.

The "Sophomore Slump"

Musicians often release a brilliant first album, followed by a mediocre second one. Why? Because they had their whole lives to write the first one (selecting only the best songs), and it was an outlier of creativity. The second album represents their "average" output.

The Prescription: The Control Group

How do we know if a drug actually works, or if it's just regression to the mean?

We need a Randomized Controlled Trial (RCT) with a Control Group.

• Treatment Group: Takes the drug. Improves by 20%.
• Control Group: Takes a placebo. Improves by 20% (due to regression to the mean).

Result: The drug is useless. The "improvement" was just natural fluctuation.

If the Treatment Group improves by 50% and the Control Group by 20%, then we know the drug has a real effect (30%).

What You Can Do

1. Be Skeptical of "Miracle Cures" for Chronic Pain: Pain fluctuates. If you try a new diet when your pain is worst, and it gets better, it might just be the timing.
2. Look for the Control Group: When reading a study, ask: "Did they compare this to doing nothing?"
3. Accept the Average: In investing, sports, and health, extreme highs are rarely sustainable. Don't bet the house that the streak will continue.

Related Content