Good EPS (Rising Earnings + Cheap P/E)

Strategy summary

Core filter: (1) TTM EPS is rising quarter-over-quarter for N consecutive quarters. (2) Current P/E is below the historical median P/E over a chosen window (e.g. 5Y).

This is a screening strategy: you get a list of candidates, then investigate why mispricing may exist—e.g. market neglect, temporary negative news, one-off items, or genuine improvement that the market has not yet repriced.

Strategy logic

Inputs:

• EPS lookback: N consecutive quarters of rising TTM EPS (e.g. 1Q to 40Q; typical default 4).
• Median P/E window: 6M, 1Y, 2Y, 5Y, or 10Y—over which the median P/E is computed.
• Entry threshold (relative to median): e.g. P/E < median, or P/E < median×(1−10%), or P/E < median×(1−20%).
• Exit rules: (a) P/E mean reversion (exit when P/E reaches median or above, or a chosen premium); (b) EPS trend break (first quarter TTM EPS does not rise); (c) time stop (e.g. 1Y, 2Y, 3Y).
• Portfolio construction: Equal-weight vs score-weighted (e.g. by discount to median); number of holdings; rebalance frequency (e.g. monthly/quarterly).

Backtest design checklist

Before trusting backtest results, address:

• EPS quality: Exceptional items, negative EPS handling (exclude/reset?), minimum EPS magnitude, TTM vs quarterly, seasonality; stability (e.g. no down quarters vs allow one miss).
• P/E robustness: Outlier handling (winsorize/trimmed median); TTM vs forward EPS; regime shifts (rates, growth); sector-relative median.
• Universe & bias: Liquidity/market-cap filters; survivorship bias (include delisted names in history).
• Timing & costs: Signal date (result date vs month-end); execution (slippage, spread, impact); signal refresh frequency.
• Risk controls: Sector caps, max single-name weight, vol targeting, drawdown stops.
• Robustness: Parameter sensitivity grid, walk-forward / out-of-sample, sub-period and regime performance; crowding (how many names qualify in euphoric vs panic regimes).

Common failure modes

• EPS quality: One-offs, accounting changes, cyclicality—rising TTM EPS can reverse.
• Median P/E instability: Valuation bands shift with rates and growth expectations; historical median may be a poor anchor.
• Cheap for a reason: Governance, demand shock, or margin compression ahead—always ask why the discount exists.
• Crowding: In stress, many names may qualify; in euphoria, few—affects capacity and execution.