What Edge Is Left?

Markets have always been inefficient. People process the same information differently, invest on different horizons, and panic at different thresholds. So if inefficiency creates opportunity, why do index funds still beat most active managers most of the time?

Either the inefficiencies aren't real, or most managers are looking in the wrong places. I've spent enough time in both camps to have an opinion.

There are six forms of informational advantage. Insider trading is illegal. Alternative data and speed require budgets we don't have. The three that remain are expert knowledge, processing advantage, and statistical edge. They're not equal.

In 2001, EPR Properties was trading at $11 with a 15% dividend yield. A yield that high is the market screaming that something is about to break. Every analyst I read said the same thing: theater REIT, bankruptcy exposure, stay away.

I had just spent years inside the movie theater industry negotiating the exact type of leases EPR owned. So I did something no analyst covering the stock had done. I read every lease in the portfolio. These weren't distressed assets. They were AMC megaplex leases with percentage rent provisions, backed by the very theaters that were killing off their competitors. EPR eventually traded above $50. The edge was real. It was also a one-time gift from a career I'd already left.

Same story, different decade. In 2022, mortgage REIT preferreds were priced like the world was ending. 35% discounts to par. High teens pricing. Rising rates, everyone running for the exits. But if you actually read the structures, the underlying mortgages were among the strongest in the country, and REIT rules meant hedging profits had to be distributed as cash flow. We bought RITM, TWO, NLY, AGNC, MFA, IVR, CHMI, and CIM preferreds. The analysis was right. But again: how many times does that window open?

Expert knowledge and processing advantage are genuine edges. They're also episodic. You can't build a business around waiting for the next EPR to show up.

Statistical edge is different because it shows up every day. Value, momentum, quality, low volatility. None of this is new. Graham identified value in the 1930s. Momentum has been documented across nearly every asset class studied. Firms like DFA, AQR, and Research Affiliates have built decades of track record on these factors, and none of them promise to beat the market every quarter.

The hard part isn't knowing the factors exist. It's not wrecking them. Most investors pile into value after it works and abandon it after a drawdown, which is exactly backwards. Value underperformed growth for most of the 2010s. A lot of smart people gave up on it right before it surged in 2021. Fees grind against you. And combining factors carelessly can cancel out the premiums you're trying to capture.

Our model evaluates companies across multiple factors simultaneously, looking for stocks where independent signals agree: a discount identified by one lens, confirmed by action from another. When low-correlation factors point at the same stock, the conviction is stronger than any single screen. We build portfolios for stock independence so that selection skill has room to surface rather than getting buried by market trend.

If someone on Reddit really knew the next Tesla, why would they share it? CNBC needs to fill airtime every hour, so every 2% dip becomes a 'selloff.' Backtested strategies work beautifully in hindsight and collapse on contact with new data. And if 1,000 investors flip coins to make buy/sell decisions, after ten flips one of them has a perfect record. That investor feels skilled. They have an amazing story at dinner. They're lucky.

The edge isn't having better data. It's understanding why people consistently misprocess the data they already have. That inefficiency is durable because it's human nature, and human nature doesn't get arbitraged away.

The next question is whether any particular edge is real or imagined. That's a measurement problem, and it's what the instruments here are built to answer.