Methodology 1.0.0

Worked example. Congress Buys:

• Autopilot Sharpe 0.71 / CAGR 20% / start 2024-04-11: official live composite of Autopilot follower accounts since launch on Autopilot.
• Quiver Sharpe ~1.12 / CAGR ~37% / start 2020-04-01: hypothetical backtest, including years before the strategy existed as an Autopilot portfolio.
• Difference is window + live vs backtest. Do not treat 0.71 and 1.12 as two estimates of the same number.

Return type

Time-weighted return (TWR) of actual Autopilot client portfolios. When at least 10 clients follow the strategy, the composite is the 10 oldest follower accounts, equal-weighted. If fewer than 10 clients are invested, all clients following the strategy are used. Money-weighted returns are rejected because cash-flow timing is not the strategy.

Valuation and calendar

Daily valuation from the published return series. 252 trading days per year. Risk-free rate is an explicit zero assumption. Timezone and corporate actions are those of the upstream adjusted series; we do not invent adjustments. Missing prices are not filled with a fabricated path.

Formulas

• Total return = last cumulative − 1
• CAGR = (1 + total)^(365.25 / days_live) − 1, only when the CAGR is labeled reliable
• Ann. vol = std(daily) × sqrt(252)
• Sharpe = (mean(daily) × 252 − 0) / ann. vol
• Sortino = excess / downside-deviation, annualized, from the same daily series as Sharpe
• Beta = cov(r, r_SPY) / var(r_SPY) on overlapping dates, n ≥ 30; otherwise n/a
• Max drawdown = min((equity − peak) / peak)
• Calmar = CAGR / |max drawdown| from the published pair
• INV-1: below 90 days live or 30 return points, performance fields are null

Gross and net

Returns are shown net of Autopilot's Advisory Fee, calculated using a model fee. The exact model amount used for the net column is pending CCO confirmation of actual versus model fee. This content_version applies a working model of a $99.99 annual Base Advisory & Licensing Fee at a $10,000 allocation. That is a model used for calculation, not a confirmed actual fee and not an advertised price. Gross returns, shown alongside, do not deduct that fee. Broker-dealer transaction costs, fund expenses, and taxes are not deducted from either figure. Because the Base Advisory & Licensing Fee is a fixed amount, smaller accounts experience a higher effective fee percentage than larger accounts.

Autopilot's Advisory Fee includes a fixed Base Advisory & Licensing Fee, charged in advance on a quarterly or yearly cycle. At the top subscription tier the Advisory Fee also includes an AUM Fee, currently 0.5%, applied only to assets above the threshold the base fee covers; below the top tier the AUM Fee is 0.00%. Basic Tier clients are not charged an Advisory Fee. Your tier is based on your invested assets, not on how many Pilots you follow, and every paid tier includes access to all Pilots. Broker-dealer transaction fees, wire fees, margin costs, and fund expense ratios are charged by your broker-dealer and apply on top of any Autopilot fee. Specific fees are set out in your Investment Advisory Agreement. See the complete fee schedule in our Form ADV Part 2A.

Benchmark

Compliance has not approved SPY as a universal overlay. Risk-band mapping (High = QQQ or SPY, Medium = AOR, Low = SCHD or CPI) is a V1 proposal pending CCO. Beta versus SPY is a reference statistic when overlapping daily observations exist (n at least 30). Total-return versus price-return is not confirmed in the published source series and is not invented. Excess versus SPY is n/a until that overlapping period return is in the published spine.

Technology and system risks

Autopilot's services depend on technology systems, third-party APIs, and internet connectivity. System outages, API disruptions, or connectivity issues may temporarily affect service availability or trade execution timing.

Common questions, including the Congress Buys 0.71 versus 1.12 case, are answered on the FAQ. Metric definitions live on /definitions.

What is not here

No hypothetical or backtest pages. Vendor backtests appear only as measurement.external_series. No per-follower dispersion. No lifetime trade counts from a week of tickets.