Systematic Strategies & Alpha

The efficient-market hypothesis in its weak, semi-strong, and strong forms, why it is untestable in isolation (the joint-hypothesis problem), why it cannot be literally true (Grossman-Stiglitz), and what "efficiency" actually means for a systematic trader.

The idea that expected returns are earned by exposure to a small set of systematic factors, the distinction between risk premia and anomalies, how factor portfolios are constructed, and the "factor zoo" multiple-testing critique that haunts the field.

The three- and five-factor models that replaced the CAPM as the empirical benchmark, the construction of SMB, HML, RMW, and CMA, the time-series regressions used to test them, and the GRS test for whether alphas are jointly zero.

The tendency of assets that have performed well (poorly) to continue performing well (poorly) over the next period.

Buying cheap and selling expensive stocks, the book-to-market signal, the risk-based versus behavioral explanations for why value earns a premium, and the anatomy of value's decade-long drawdown from 2007 to 2020.

The return you earn if prices do not move, defined consistently across currencies, rates, and commodities, why it is best understood as a risk premium for bearing crash and liquidity risk, and how the carry trade blows up.

The empirical fact that low-risk stocks earn higher risk-adjusted returns than high-risk stocks, a direct contradiction of the CAPM, and the leverage-constraint (betting-against-beta) explanation that resolves it.

Paying up for good businesses, profitable, growing, safe, well-managed firms, and why they earn higher returns than junk despite being "better." The QMJ framework, its link to the dividend-discount identity, and its role as the complement to value.

Turning a raw predictor into a tradeable signal, winsorizing, z-scoring and ranking, neutralizing unwanted exposures, and measuring its quality with the information coefficient and the fundamental law of active management, IR = IC·√breadth.

Why signals lose their edge over time, the half-life of predictive power, post-publication crowding, and the turnover-versus-decay tradeoff that determines how fast you must trade a signal before transaction costs eat it.

The two fundamental ways to build a systematic strategy, ranking assets against each other (cross-sectional) versus each asset against its own history (time-series), how demeaning determines market-neutrality, and the algebraic identity linking them.

The tendency of prices or returns to revert to a long-run average after deviations.

The original convergence trade, find two cointegrated assets, trade the mean-reverting spread with z-score entry and exit rules, size by the half-life of reversion, and understand why the relationship eventually breaks.

Cross-sectional mean reversion at scale, strip out common factors with PCA, model the idiosyncratic residual as a mean-reverting process, and trade its s-score across hundreds of names. The Avellaneda-Lee framework and why breadth is the whole game.

Time-series momentum traded across dozens of futures markets, the CTA strategy that buys what's going up and sells what's going down, why trends persist, and the long-volatility convexity that makes it "crisis alpha."

Identifying and adapting to different market states (trending vs mean-reverting, risk-on vs risk-off).