The Four Horses of Learning: Why Some People Change and Others Don't

TL;DR: The Buddha once described four types of horses: those that run at the shadow of the whip, those that run when the whip touches them, those that run only when struck, and those that run only when beaten to the point of death. Learners fall into the same four categories. The dangerous delusion isn't being a slow learner—it's thinking you're a fast one.

James here, CEO of Mercury Technology Solutions. Hong Kong — August 2026

A reader wrote me with a puzzle.

He understood that learning matters. But he also sensed that something about how he learned wasn't working. He'd read strategies, internalize them, try them—and fail. The gap between "I get it" and "I can do it" felt unbridgeable.

His question: What actually happens inside different people when they learn? Why do the same instructions produce radically different results?

The answer is 2,500 years old. And most people hate it.

The Buddha's Four Horses

In the Bamboo Grove at Rājagṛha, the Buddha once classified all beings into four types of horses.

The first horse sees the whip's shadow and runs.

The second feels the whip brush its tail and runs.

The third must be struck to move.

The fourth only runs when beaten to the point of death—when survival itself is at stake.

The Buddha wasn't talking about horses. He was describing the full spectrum of human responsiveness to truth.

Most learning theory today—spaced repetition, active recall, deliberate practice—assumes you're the second horse. Maybe the first. It optimizes for people who actually learn when exposed to information.

But what if you're the third? Or the fourth?

The Four Grades of Learners

Grade One: 蓋天授也 (Heaven-Gifted)

Zhang Liang once described Liu Bang with this phrase: gai tian shou ye—"heaven-bestowed."

The story: Liu Bang was losing to Xiang Yu everywhere, barely holding on while waiting for Han Xin to save him. Finally Han Xin's messenger arrived. Not with troops. With a demand: Han Xin had conquered Qi and wanted to be named "Acting King of Qi" to legitimize his rule.

Liu Bang exploded. He was about to have the messenger killed when Zhang Liang caught his eye and whispered: We still need him.

Liu Bang's response, instantly: "You know why I was angry? Because a real man doesn't settle for acting king. He becomes actual king. Go tell Han Xin: I name him King of Qi."

Most people read this story and mock Liu Bang as a hypocrite. The same people, when they encounter real pressure, freeze for hours—or days—before adjusting.

Grade One learners don't freeze. They're told something, they understand it, they execute it. No gap. No "processing time." No "I need to sit with this."

You ever sit in a classroom where the teacher explains a concept once, and someone immediately solves every variant of that problem forever? That's Grade One.

You've probably never met one.

Grade Two: 不貳過 (Never Twice)

Confucius ranked his 3,000 disciples and named Yan Hui first. The others asked why.

"Yan Hui never makes the same mistake twice," Confucius said.

Meaning: Yan Hui might not nail a concept on first exposure. But once he got a problem wrong, that category of error was permanently eliminated. You could test him any way you wanted—he'd never repeat it.

Think about what this means in practice. Someone with this trait studies the curriculum once, misses some problems, learns from each miss, and emerges with permanent mastery. No decay. No "I used to know this."

The scary thing about Grade Two: Their errors have a half-life of one. Today's mistake is tomorrow's impossible mistake.

Most "top students"—provincial champions, competition medalists—aren't Grade Two. They're Grade Three with lower costs.

Grade Three: The Cost Curve

Here's where most high achievers actually live.

A Grade Three learner needs repetition. But the quantity of repetition determines their ceiling:

• 10 problems → 90% accuracy

• 20 problems → 95% accuracy

• More practice → marginal gains, but permanent

Within this category, people compete on cost. The ones who hit 95% in 10 problems become provincial champions. The ones who need 15 get into top universities. The ones who need 25 settle for good schools.

Same type. Different efficiency.

This is where the education system optimizes. Standardized tests, practice exams, drill books—all designed for Grade Three learners with varying cost curves.

The system works for them. Which is dangerous, because it lets them believe they're Grade One or Two.

Grade Four: The Living Dead

Grade Four learners don't respond to practice. You can give them 500 problems, 800 problems—their accuracy plateaus below 70%. The knowledge doesn't stick through repetition.

Their only teacher is pain.

Financial ruin. Betrayal. Health collapse. The kind of experience that rewrites your nervous system because the alternative is death.

Sometimes they learn. Often they just break.

This isn't a moral judgment. It's a cost function. Some people require life-threatening stakes to update their mental models. The market, relationships, and reality are their only effective instructors.

The Dangerous Fantasy

Here's what breaks people: They imagine themselves as Grade One or Two.

They hear a trading strategy explained clearly. "I understand it completely," they think. Then they execute, and everything unravels. The same mistakes, the same blind spots, the same emotional cascades.

They don't understand why. They heard it. They got it.

But here's the thing: You also heard every concept in school clearly. You had years to practice. You still didn't score 700 on the gaokao. The test makers literally gave you 50 points of margin—and you still couldn't close the gap.

Why would trading be different?

A decade later, most people are making the exact same errors they made ten years ago. Same cognitive traps. Same emotional patterns. Same failure to update.

This is why Yan Hui was terrifying. Not because he was perfect—because his tomorrow was different from his today. His errors evolved. Most people's errors fossilize.

If you could simply ensure that this year's mistakes are different from last year's, you'd be employable as a professional trader tomorrow. The bar is that low. Most people can't clear it.

The Agarwood Principle

I call this the Agarwood Principle.

Grade One people aren't agarwood trees. They're already fragrant wood. Exposure to pressure doesn't create their capability—it merely reveals what was always there.

Grade Two people are agarwood with 99% resin density. A little pressure, a little time, and the fragrance floods out.

The rest of us are Grade Three and Four agarwood. We need infection, injury, years of pressure before we produce anything valuable. The process is slow, painful, and non-negotiable.

The question isn't whether learning works. The question is: Which grade are you actually in?

Not which grade you wish you were in. Not which grade your parents told you. Not which grade your test scores temporarily suggested.

Which grade does your behavior place you in?

• Do you hear advice and implement it same-day? (Grade One)

• Do you make a mistake once and never repeat it? (Grade Two)

• Do you improve with deliberate practice, but slowly? (Grade Three)

• Do you only change when reality beats you to the edge of death? (Grade Four)

The tragedy isn't being Grade Three or Four. Billions of successful people are.

The tragedy is thinking you're Grade One, designing your life around that assumption, and wondering why nothing sticks.

What to Do With This

If you're Grade Three—and statistically, you probably are—stop pretending you're Grade One.

Stop chasing "aha moments." They don't change behavior. Start building systems that force repetition until your error rate drops below threshold.

Stop trusting your "understanding." Understanding without execution is just entertainment. The only valid test is behavior change over time.

Stop comparing yourself to prodigies. They're playing a different game with different hardware. Your path is longer but still viable—if you accept the cost curve.

And if you're Grade Four? Stop hoping information will save you. Information doesn't reach you. Only consequences do. Your job is to engineer smaller consequences before the existential ones arrive.

Build artificial pressure. Create accountability systems that hurt enough to matter. Surround yourself with people who won't let you off the hook.

Because the whip's shadow won't move you. The whip itself won't move you. You need to feel the pain—strategically, early, and often—before life delivers it catastrophically.

Mercury Technology Solutions: Accelerate Digitality.