Skip to content

Knowledge · Evaluation

How to tell whether an alpha actually works.

Alphas are easy to find and easy to fool yourself with. This guide shows how professionals judge one: two families of alpha, the scorecard each needs, and the checks that separate an edge from a lucky backtest (real costs, unseen data, an honest count of what was tried).

A framework for reading any alpha claim, ours included. Not a capital recommendation.

0
Alpha families
0
Evidence tiers
0
Defined metrics
0
Interactive labs

Cross-sectional

The ranker

Ranks a broad universe on each date and bets on the order: which names beat which. Judged on rank IC, quantile spreads and the P&L of a simple long/short book, with turnover, cost and multiple-testing checks on top.

Time-series

The forecaster

Forecasts each asset's own next move and trades a concentrated book of a few names. Because the forecast becomes the position, it is judged on that book's out-of-sample P&L, net of costs and deflated for multiple testing.

01 — Foundations

Two families of alpha, two scorecards

Every alpha starts as one score per asset: per bar for an intraday strategy, per rebalance date for a multi-day one. How that score is judged depends on what it is built to capture. Part A covers the ranker's scorecard; Part B the forecaster's.

CROSS-SECTIONAL · across names, one daterank the universe → IC = corr across namesTIME-SERIES · one asset, over timeforecast vs realised over time → corr over time

Left: the ranker's skill is whether high scores out-ranked low scores at each date. Right: the timer's skill is whether the forecast called each asset's own path. The two skills answer different questions, so they get different scorecards.

Sign in to continue

See the full evaluation framework

You've seen the two families and why each needs its own scorecard. The rest of the guide (the eight questions, the metric deep-dives and the interactive labs) unlocks free when you sign in.

  • The eight questions every full evaluation answers
  • Part A: the cross-sectional scorecard, with rank IC, quantile spreads and typical healthy ranges
  • Part B: the three-tier time-series verdict and forecast-quality diagnostics
  • Parts C and D: costs, capacity and the frequency rulebook
  • The capital decision, six interactive labs, and the full glossary

Checking your access…