The Godfather Of Technical Analysis: Your Stop Losses Are Costing You More Than You Think

ELI5/TLDR

Larry Connors, a legendary quant trader and author of 25+ books, argues that mechanical stop losses systematically degrade trading performance. His 45-year research found that any type of stops hurt returns. Instead, he builds positions asymmetrically—accepting short-term noise while positioning for power-law outsized returns. His edge comes from second and third-order thinking: seeing what happens after the obvious moves play out. This is a directly counter to retail trader orthodoxy.

The Full Story

The Stop Loss Thesis

The episode opens with Connors’ core finding from his Charles Dow Award-winning research: statistical evidence that “any type of stops that were put in place degraded the performance of the methodology.” This isn’t intuitive. Stop losses feel like risk management. But Connors backs the claim with an unlikely ally: Stanley Druckenmiller, arguably the greatest trader of our generation, who said plainly, “I’ve been trading for 40 years, I’ve never used a stop in my life.”

Druckenmiller’s rationale cuts deep. Stops create false certainty. When markets crack—“if something crazy in the world happens right now”—you don’t get a tidy exit. You get filled as a market order, at whatever price the panic gives you. Stops are psychological anchors that feel safe but execute at the worst moments.

First, Second, Third Order Thinking

Connors moves from the stop loss paradox into a broader framework: order of thinking.

There are usually no edges in first-order thinking. Things that are basically obvious—that’s where everyone operates.

The potential alpha appears when you jump to second-order thinking: after the first-order move, what else is required? When everyone agreed AI was coming, Connors asked: what comes after? Power. Data centers.

This cascaded into third-order thinking and beyond. He looked at Duolingo—yes, an AI can teach languages for free. But who can compete when AI commoditizes services? That’s the edge. Duolingo stock reflected assumptions; understanding what happens after those assumptions play out is how you get returns.

The RSI, Fear, and Mean Reversion

Connors pioneered the 2-period RSI and later the Connors RSI. The mechanics aren’t complex: shorter-term RSIs (2, 3, 4-period) measure fear. Markets either sell off because investors worry, or they pause the buying. The RSI captures that inflection—the moment fear peaks before mean reversion kicks in.

What they’re really doing is they’re measuring fear. They’re either selling the market off because they’re concerned about some sort of event, or they’re just pausing the buying.

His systematic trading background leaned into this mean reversion dynamic. A portion of capital (10-20%) was always reserved for special situations—thematic trades where AI disruption or infrastructure buildouts create outsized returns—but the core was systematic.

Power Law and Position Building

The internet bubble taught Connors a hard lesson. He’d buy a stock at $6, see it jump to $9 overnight (50% gain, taken), then watch it run to $50, $60, $70. Regret.

By the time AI became investable, he had reverse-engineered that mistake. The power law didn’t change—a small number of positions would deliver the bulk of returns—but his sizing did. He built into Corning at $60, watched it triple to $180. Micron at $100, then again at $400, hitting $1,000. These outsized returns only materialize if you stay in the position through volatility. Stops would have guillotined those gains early.

Buy Low vs Momentum: A False Dichotomy

When pushed on buy-low-sell-high versus momentum trading, Connors refused the binary. A thematic trade on Corning Glassware looks like momentum on a chart (stock already moved significantly) but it’s really a thematic thesis: optics are essential for AI, Corning is a prime beneficiary. The chart momentum is incidental to the order-of-thinking thesis.

The Temperament for Endurance

Connors admits he burned out at 34. Trading into market close every day, year after year, hollowed him out. The conversation touches on personality—what separates winners from the many who fail? His answer is honest: temperament matters, but more than that, winners are risk managers first.

The hedge funds he’s worked with span different personalities, upbringings, and styles. They couldn’t be more different. But they all positioned themselves for asymmetry—large potential gains, carefully bounded losses. Not through stops, but through position sizing and thesis clarity.

Learning, AI, and Accumulation

Connors makes a point worth sitting with: this isn’t a quick skill. Nobody wakes up ready to trade. Knowledge builds incrementally. The industry marketed “quit your job, retire from trading” because it sells subscriptions. The reality is harder. It’s acquired, layered, accumulated over decades.

AI changed the speed. Connors compressed years of options learning into months by using AI properly. That itself is a power law. But the underlying skill—risk management, understanding orders of thinking, patience—remains foundational.

Parting Advice

To newer traders (late 20s, early 30s, early in their career), Connors’ final word: risk management first. Full stop. Every legendary trader he knows doubled down on this. Dig in. Persist. Don’t expect to make money every month. Use AI, but understand it’s a tool for acceleration, not a substitute for rigor.

Key Takeaways

Stop losses systematically degrade performance — research finding backed by both Connors and Druckenmiller; market gaps and panic fills eliminate the safety they promise.
Second-order thinking is where edges live — what happens after the obvious move; requires seeing cascading consequences (power → data centers → competitive positioning).
RSI measures fear, not direction — 2-3-4 period RSIs capture inflection points when fear peaks and reversions are imminent.
Position sizing replaces stops — asymmetric positioning (bounded losses, outsized upside) comes from capital allocation, not mechanical exits.
Power law dominance — a small number of positions deliver bulk returns; staying in through volatility matters more than perfect entries.
Temperament + risk management = longevity — separates winners from the many who burn out or blow up; it’s not personality type, it’s discipline.
It’s an acquired skill — nobody was born to this; knowledge compounds over decades; AI accelerates learning but doesn’t replace rigor.
Thematic investing over style labels — momentum vs. mean reversion is a false binary; the real edge is order-of-thinking, not chart pattern recognition.

Claude’s Take

This is a genuinely contrarian take, and Connors has the credibility to hold it—45 years of quantified results, a Charles Dow Award, a substantial track record. The stop loss argument is provocative but evidence-backed, not emotion-driven bravado.

That said, there’s a signal-noise question worth interrogating. Connors is describing a style—concentrated, thesis-driven, systematic with discretionary exceptions—that requires conviction and capital to weather drawdowns. A retail trader with $10k and a day job cannot stay in a position for years waiting for the power law to play out. His risk management advice (capital allocation, asymmetric sizing) is universal. His specific rejection of stops is more contextual: it works for a systematic quant with enough capital that volatility is noise, not catastrophe.

The second and third-order thinking framework is genuinely useful—it’s a mental tool for spotting where the crowd’s vision stops and where returns actually hide. The Duolingo example works: most people saw “AI can replace tutoring”; he saw “Duolingo provides services that become commoditized by AI” → different conclusion.

The hour-and-45-minute format is meandering—the conversation touches on AI, power laws, temperament, and learning without drilling deeply into any single edge. If you’re looking for specific technical rules or a scalable system, you won’t get it here. What you get is philosophy from someone who has survived and thrived in markets for half a century. That’s valuable, but it requires you to translate it to your own context.

The AI accelerator framing is honest and interesting. Condensing years of learning into months is real; using AI to generate new thematic hypotheses faster is real. But the disclaimer—that this is a tool, not a replacement for rigor—is crucial and undersaid.

The temperament and risk management section lands hardest. “They were risk managers first” echoes across every legendary trader profile. That’s the through-line, not the stops or the RSI or the themes.

Score justification: Solid expertise, useful frameworks (order thinking, power law, risk-first), backed by credible track record. The thesis is contrarian and worth scrutiny. But it’s conversational and somewhat uneven—some passages are gold, others are filler. It’s a 7 because it’s worth your time if you’re thinking about how positioning and thesis clarity matter more than entry/exit mechanics, but it’s not a complete system or tightly argued manifesto.