QuantRocket VMOTTrend: SPY Weekly Two-Rule Short Hedge

Hypothesis

Faithful port of QuantRocket's Apache-2.0 VMOTTrend, the Trend component of Alpha Architect's Value/Momentum/Trend approach. Trade SPY.USEQ using daily session-close bars. QuantRocket identifies source SID FIBBG000BDTBL9 as SPY: https://www.quantrocket.com/docs/. Preserve the source's 252-session moving average, 252-session return, weekly Sunday-labeled resampling, equal half-weight hedge signals, and one-session position delay. This is the standalone short hedge component, not the complete VMOT long portfolio. Replace the source's $0.005/share commission with platform USEQ zero commissions and modeled spread/impact. Replace SMART MOC routing with the platform's market execution at the prescribed daily session close; this preserves close-based simulated timing but does not reproduce auction execution. Source: https://github.com/quantrocket-codeload/vmot/blob/4c579d703c4dec99c8e3b0f42031ea2dbf210207/vmot/vmot_trend.py.

A submitted exact-rule strategy takes precedence over autonomous diversity preferences. Preserve SPY and the original hedge mechanism rather than changing symbols or adding filters because of prior failures. The proposal contributes short-only equity exposure outside the dominant Binance long-only bucket. Corpus concentration was supplied as an empty object, so no instrument-specific concentration claim is made.

negative_expectancy: decline optimization of this fixed standalone SPY short hedge. Over 48 closed positions, realized net losses total approximately $37,856, with PF 0.517 and negative per-trade economics. This verdict covers the isolated hedge allocation, not QuantRocket's complete VMOT portfolio or the hedge's incremental portfolio contribution. Evaluate that contribution separately against an otherwise identical unhedged portfolio while preserving the source rules.

Implementation

Standalone SPY short hedge translating QuantRocket VMOTTrend: each negative 252-session trend condition contributes 50% short exposure. Sunday-labeled weekly signals are forward-filled and delayed one session. At every daily session close, whole-share holdings are adjusted to the delayed target fraction of current mark-to-market equity.

Addresses the critical Layer 4 allocation mismatch by removing the unchanged-fraction early return and recalculating the share target every daily close. Initial entries also use mark-to-market equity, which equals settled equity while flat. Indicator definitions, weekly updates, half weights, one-session delay, instrument and fixed research parameters are preserved. USEQ market orders approximate source close execution without reproducing SMART MOC auction execution; platform spread/impact and zero commissions replace source commissions. Sandbox losses alone do not establish a coding defect or measure this hedge's contribution to the complete VMOT long portfolio. No verification or reserved final-exam data reads were performed. The required outbox file could not be written because this session's filesystem permissions are read-only; the complete bound JSON is returned here.

Verification Results

The sandbox reports -25.77% return and -2.44% average closed-position return. These are advisory results for a standalone short hedge; evaluate full-history economics and portfolio hedging contribution separately.

Backtest Review

Sharpe

-0.05

Total return

-37.86%

Max drawdown

63.42%

Trades

48

Win rate

12.5%

Profit factor

0.52

Complete audited history spanning 1994–2024; 48 closed short positions with multiweek holds are reasonable for this slow hedge.

Trade direction matches the hypothesis. Negative beta (-0.5003) supports potential hedging utility, which requires a combined-portfolio test.

Standalone net return is -37.86%, profit factor 0.517 and average trade return -1.891% of notional versus the USEQ review floor of +0.05%. QA's standalone economic concern is borne out.

Drawdown is 63.42% versus the declared 30% risk limit; this reinforces the economic finding rather than serving as the sole rejection reason.

The faithful specification has no tunable parameters, so optimization has nothing to search.

Iteration History

At each prescribed daily close, calculate the whole-share target from current mark-to-market equity and the delayed weekly fraction, then trade the difference from actual holdings even when the fraction has not changed. Retain the weekly signal updates and one-session execution delay.

should_exit() skips all quantity adjustments whenever the target fraction is unchanged. Consequently, the hedge holds a fixed share count through potentially months of unchanged signals, rather than maintaining the source's forward-filled 50% or 100% portfolio weight. Price changes and short-position PnL cause exposure to drift. The source computes daily returns using the delayed target weight each day; this implementation therefore changes its allocation and return mechanics. Source: https://raw.githubusercontent.com/quantrocket-codeload/vmot/4c579d703c4dec99c8e3b0f42031ea2dbf210207/vmot/vmot_trend.py

Outcome Summary

Evaluate the hedge's incremental contribution against an otherwise identical unhedged portfolio while preserving the source rules.

Backtest review voted abandon because the isolated hedge had negative expectancy and exceeded its declared 30% drawdown limit; its fixed specification offered no parameters to optimize. Optimization, post-optimization analysis, and risk review were not reached.

Port QuantRocket's standalone SPY short hedge using equally weighted 252-session moving-average and return conditions, weekly Sunday-labeled signals, and a one-session position delay.

The 1994–2024 backtest returned -37.86% with Sharpe -0.055 across 48 closed short positions. Maximum drawdown was 63.42%, profit factor was 0.517, and average trade return was -1.891% of notional.

Analysis

VMOTTrend