The Lighthouse Preview · November 2026

The 53-Sharpe Mirage

Could a simple stock-market formula really produce a Sharpe ratio above 53? We rebuilt the famous contrarian idea. The headline did not survive—but something more useful did.

The reveal

The MIT paper’s spectacular number was a diagnostic result under conditions its authors called unrealistic after costs and constraints. Our daily recreation lost money. Moving to weekly candles and removing the short side produced a research lead worth studying.

A claim too good to ignore

What worked, what failed—and how that helps you.

The 2007 Khandani–Lo paper used a deliberately simple contrarian portfolio to investigate the August quant meltdown. It was not offering investors a ready-made system. Yet one number leaps off the page: a 53.87 annualized Sharpe ratio for the all-stock portfolio in 1995. Read on and the experiment becomes more useful: the daily idea failed, the weekly version survived, and the size of the stock universe changed the conclusion. Knowing why can help you question spectacular strategy claims, recognize the difference between a temporary pullback and persistent weakness, and make better-informed decisions.

01 · THE CLAIM

Buy yesterday’s laggards

Short yesterday’s leaders and expect the relative gap to close. Rebalance every day across thousands of stocks.

02 · THE CATCH

The authors warned us

The paper called its 1.38% average daily return unrealistic after trading costs, market impact, short constraints, and other limitations.

03 · OUR TEST

Make the signal wait

We used only a completed period, traded at the next adjusted open, limited position size, and charged for turnover.

Six tests, one surprise

Daily reversal failed. Weekly reversal changed the conversation.

The daily long/short variants lost money in both the concentrated sample and the broad current universe. All three full-universe weekly variants were positive. The strongest bought weekly laggards relative to their own sectors and omitted the short side.

The number is not the conclusion.

Current-universe survivorship, historical classifications, simplified execution, and high turnover remain unresolved. Those limitations prevent treating 22.83% as achievable performance.

Daily all-stock L/S
−9.86%
Daily sector L/S
−14.55%
Daily sector long
13.34%
Weekly all-stock L/S
13.83%
Weekly sector L/S
10.84%
Weekly sector long
22.83%

Net annualized results under the common test assumptions. Bar length represents absolute magnitude.

The twist

A 100-stock sample nearly hid the effect.

The weekly sector-relative long-only test returned only 4.46% in a carefully balanced liquid sample, but 22.83% across the broader eligible universe. That raised an obvious suspicion: was the result merely an illiquid-stock illusion?

18.52%Lowest-liquidity quartile
16.62%Liquidity quartile 2
16.76%Liquidity quartile 3
17.20%Highest-liquidity quartile

Every liquidity and capitalization quartile remained positive. The more likely lesson is breadth: a cross-sectional effect can disappear inside a small hand-shaped sample.

The WealthVelocity distinction

Reversal can be interesting without replacing strength.

Reversal buys a stock because it fell behind and expects a partial recovery. WealthVelocity buys demonstrated strength and stays while the evidence remains constructive. Those are different promises to the investor.

Primary philosophy

Strength selection

  • Buys demonstrated strength
  • Expects trend persistence
  • Holds while the trend remains constructive
  • Requires exchange, sector, trajectory, liquidity, and timing confirmation
VERSUS

Research alternative

Weekly reversal

  • Buys relative underperformance
  • Expects partial mean reversion
  • Needs a limited recovery window
  • Requires a measurable peer dislocation

The more compatible experiment

Strength first. Pullback second.

Start with global stocks that already pass WealthVelocity’s strength and quality gates. Then use a temporary weekly lag as an entry-timing candidate. A brief pullback may be opportunity; persistent failure to rejoin the group is rejection.

The Lighthouse view

The best finding was not the biggest number.

It was the sequence of reversals in our own conclusion. The daily rule failed. Weekly data changed the result. The sample nearly hid the effect. Quartile tests weakened the easy liquidity explanation. Each test made the story less magical and more useful.

A statistical effect can deserve investigation without deserving capital.

Weekly reversal now belongs in the research laboratory and investor education. WealthVelocity selections remain grounded in visible, sustained strength.

Primary sources

Read past the headline.

Research cutoff: October 29, 2026. Historical evidence in this report is U.S.-only. Backtested results are hypothetical and do not represent actual trading. Informational research only; not personalized investment advice.