Demonstration · not investment advice

Research desk

The same ranking, run separately on four quarters of the universe, each holding its top 10 stocks. If the edge is real, all four groups should earn it.

Simulated monthly prices, Oct 2014 to Sep 2026 · Backtest through Dec 2024, live since Jan 2025 · Last run Oct 7, 2026, 6:10 a.m. CT

Ranking system

Backtest

Period

Growth of $1 in each group

Top 10 stocks by rank in each quarter of the universe, after trading costs. Log scale.

    How these are calculated
    Returns
    Monthly, after trading costs of 0.20% a side on the positions that change at each rebalance.
    Sharpe
    Mean monthly return above a 2% annual risk-free rate, divided by its standard deviation, times √12.
    Adjusted Sharpe
    Pezier and White's adjustment: SR × [1 + (S/6)·SR − ((K − 3)/24)·SR²], where S and K are the skewness and kurtosis of monthly excess returns. Negative skew and fat tails lower it.
    Sortino
    Annualized excess return divided by the annualized deviation of the months below the risk-free rate.
    Max drawdown
    The largest fall from a peak in the growth of $1 during the selected period.

    Rolling 12-month Sharpe

    Whether each group's edge holds up over time, or fades.

      Live since Jan 2025

      Fixed range: compared with the full backtest.

      Live excess return against the backtest

      All 40 stocks against the benchmark, inside the range the backtest led you to expect.

        Live turnover against the backtest

        The share of the 40 holdings replaced each month, averaged since Jan 2025. Trading costs rise with it.

          Alerts

          Checked after each monthly run.

            Assumptions

            What every number on this screen depends on.

            Universe
            400 synthetic stocks, each with a random StockID
            Groups
            Four, by StockID mod 4; each runs the same ranking on its own quarter
            Holdings
            Top 10 by rank in each group, equal weight
            Rebalance
            Monthly
            Trading costs
            0.20% a side on positions that change
            Turnover range
            The 5th to 95th percentile of average turnover over every stretch of the backtest as long as the live period
            Benchmark
            Equal-weight average of all 400 stocks, without costs
            Risk-free rate
            2% a year, held constant
            Point in time
            Each ranking sees only what was known at its rebalance. With real data, fundamentals are dated by when they were filed. The backtest check shows what dating them by the quarter does.
            Survivorship
            No stock in this simulation delists. A real universe has to include stocks that were later delisted; the backtest check shows what leaving them out does.