The eight house strategies

Each house strategy is a fixed rule set run by the portfolio simulator and tracked live from its launch date. They share a base: the liquid universe (price ≥ $5, market cap ≥ $300M), the variable cost model, total return with dividends, $100,000 of model capital, equal weights, a 30-day buy-back cool-off and SPY as the benchmark. Where a published method could not be expressed exactly, the strategy page lists the approximation; they are repeated here. None of the descriptions below is a performance claim: the live record on each strategy's page is the only one.

Core Combination

Idea. Own the stocks that look good on everything at once: value, quality, momentum, growth and low risk. Origin. Multi-factor investing: each style has decades of evidence and long droughts, and the droughts do not line up, so a blend is smoother than any single style (Asness, Moskowitz and Pedersen on value and momentum together). Rules. Ranked by Core: Combination; 25 positions; monthly; sold when the rank falls below 70; no sector above 30% of the portfolio. What can go wrong. Fully invested with no hedge, so drawdowns close to the market's; a blend of five styles can trail a market led by a handful of expensive growth stocks for years.

Quality at a Reasonable Price

Idea. Good businesses, not at any price. Origin. The quality factor (high and stable profitability, low leverage, Asness, Frazzini and Pedersen's "Quality Minus Junk"; Novy-Marx's gross profitability) combined with a value check, the approach often associated with Buffett's "wonderful company at a fair price". Rules. Ranked by the Quality at a Reasonable Price system (Quality 60%, Value 40%); 20 positions; quarterly; sold when the rank falls below 70. Approximation. The sell rule was added: the simulator sells only on rules. What can go wrong. Quality lags in sharp rallies led by low-quality, beaten-down stocks ("junk rallies", as in 2009 and late 2020); 20 names and no sector caps.

Magic Formula

Idea. Good businesses at bargain prices, in one rule: rank on earnings yield and return on capital equally. Origin. Joel Greenblatt, The Little Book That Beats the Market (2005). Rules. Market cap ≥ $300M, sector not Financials; ranked equally on EV/EBIT (lower better) and ROIC (higher better), NA policy negative; 30 positions; annual rebalance at year end; sold when the rank falls below 90 (out of the top 10%) or the stock fails the buy rules. Approximations. Greenblatt re-forms the whole basket each year; here a holding is kept while it stays in the top 10%. Utilities are not excluded, only Financials. Greenblatt uses his own definitions of EBIT, enterprise value and tangible capital; Stoxa uses its standard EV/EBIT and ROIC fields. What can go wrong. A deep-value tilt with long stretches behind the index. Portfolio123's 2020 study found the formula lagged the market for a decade after publication; it is kept as a reference, and the live record will show whether that continued.

Piotroski-style Value

Idea. Cheap stocks are often cheap for good reason; simple accounting health checks separate the recoveries from the value traps. Origin. Joseph Piotroski (2000), whose nine-point F-score, applied to the cheapest fifth of stocks by book to market, separated winners from losers. Rules. ROA > 0, free cash flow > 0, current ratio > 1, debt to equity < 1; ranked by Core: Value (industry-relative); 20 positions; quarterly; sold when a health test fails or the rank falls below 60. Approximations. Four of the nine F-score tests can be expressed today, and two of them as levels rather than improvements: the current ratio and leverage tests check a level, not a year-over-year change. Not available yet: cash flow above net income (it compares two fields), and the changes in ROA, leverage, current ratio, shares outstanding, gross margin and asset turnover (no year-over-year fields yet). What can go wrong. Value can trail growth for years; without the "change" tests the screen is weaker than Piotroski's.

Dividend Growth

Idea. Dividend increases come from companies with a moderate yield, spare payout and growing earnings, not from the highest yields, which are often a warning of a cut. Origin. Practitioner dividend-growth investing (the "dividend aristocrat" style); the academic case is weaker than for value or momentum, and largely runs through quality and profitability. Rules. Dividend yield between 1.5% and 6%, payout ratio below 70%, EPS growth above 0; ranked on Quality 60% and Momentum 40% (NA policy negative); 20 positions; quarterly; sold when a buy rule fails. Approximation. Portfolio123's equivalent ranks on analyst sentiment; Stoxa has no analyst data, so price momentum stands in. What can go wrong. Dividends are not contractual; a cut takes the stock out only at the next quarterly rebalance. Sector concentration in utilities, staples and financials is possible with no caps.

Low-Volatility Quality

Idea. Steady, profitable companies with small price swings. Origin. The low-volatility anomaly: Blitz and van Vliet (2007) found that the least volatile stocks earned market-like returns with much lower risk, against the textbook expectation that risk is rewarded (also Ang, Hodrick, Xing and Zhang; Frazzini and Pedersen's "Betting Against Beta"). Requiring quality avoids stocks that are merely quiet on the way down. Rules. Ranked on Low Risk 50% and Quality 50% (NA policy negative); 30 positions; quarterly; sold when the rank falls below 60 or the close is 25% below the highest close since purchase. Approximation. The rank sell rule was added. What can go wrong. Lags in strong rallies; low-volatility stocks behave like long-duration bonds and suffer when interest rates rise quickly; the trailing stop turns a fast fall-and-rebound into a realised loss.

Momentum with a Quality Floor

Idea. Ride the strongest price trends, but only in profitable, moderately levered companies, and halve exposure in bear markets. Origin. Jegadeesh and Titman (1993) on medium-term momentum; Daniel and Moskowitz on momentum crashes, which cluster at market rebounds and in speculative names; Faber's moving-average rule for the hedge. Rules. ROIC > 0, net debt to EBITDA < 3; ranked by Core: Momentum (6- and 12-month returns); 20 positions; monthly; sold when the rank falls below 60; while SPY closes below its 200-day average, 50% of the portfolio is in cash. What can go wrong. Momentum reverses sharply at market turns; the hedge misses part of V-shaped recoveries and whipsaws in sideways markets; monthly trading and high turnover make it the most cost-sensitive of the eight.

Market-Implied Growth Gap

Idea. Buy companies whose price assumes much less growth than they have actually delivered. Origin. Stoxa's own, built on the valuation engine: the reverse DCF solves for the revenue growth that makes the model's value equal the price (the expectations-investing idea of Rappaport and Mauboussin), and the gap compares it with the record:

The three-year revenue CAGR uses the last four fiscal-year revenues filed by the date. Rules. Growth gap ≥ 5 points and the valuation data check not "error"; ranked by Core: Value; 20 positions; quarterly; sold when the gap turns negative or the rank falls below 50. Restrictions. Implied growth is computed nightly for the liquid universe only (about a second per company), and for past quarter ends a few dates per night, so the backtest before launch is thin where dates are not computed yet. Financials and REITs are valued with a dividend model and have no implied revenue growth, so they are never bought. The equity risk premium and sector norms inside the valuation defaults are a bundled snapshot, not values as of each date. What can go wrong. Past growth can stop, and the market is often right that it will: a low implied growth is sometimes a correct forecast of a shrinking business. Unlike the others, this strategy has no published track record outside Stoxa.