Vision Weekly Optimization: Stop-Loss Testing Results¶
Date: October 1st, 2025 Author: Daniel Shanklin, Director of AI and Technology, AIC Holdings Tags: Short Selling, Risk Management, Vision Research, Statistical Analysis, Trading Optimization
Sources: - Vision Research LLC weekly short recommendations (2024-2025) - Internal analysis of 72 realized Vision Weekly trades - Statistical validation using 25 train/test splits
Vision Research LLC provides weekly short recommendations to institutional investors, but our analysis of 72 realized trades revealed a performance paradox: 56.9% win rate generating -9.6% average returns. The culprit was extreme outliers, with the worst trade losing -152% while the best gained +63%.
Testing Methodology¶
We analyzed 252 different exit strategies using rigorous train/test methodology across 25 data splits. Each strategy combined stop-loss levels (1% to 50%) with stop-gain levels (5% to 50%, plus no limit) to create a comprehensive 7x9 matrix of possibilities.
Key Finding: 3% stop-loss with no stop-gain limits emerged as optimal in 100% of test splits (25/25), transforming -9.6% baseline returns into +12.6% optimized returns—a 22.1 percentage point improvement.
Matrix Results: Stop-Gains Destroy Performance¶
| Stop-Loss | 5% Gain | 10% Gain | 20% Gain | 50% Gain | No Limit |
|---|---|---|---|---|---|
| 1% | +2.3% | +4.9% | +9.3% | +13.4% | +13.4% |
| 3% | +1.5% | +4.1% | +8.4% | +12.6% | +12.6% |
| 10% | -1.3% | +1.3% | +5.7% | +9.8% | +9.8% |
| No Stop | -20.6% | -18.1% | -13.7% | -9.6% | -9.6% |
Critical Insight: Reading across any row shows stop-gains consistently reduce returns. Even with optimal 1% stop-loss, implementing a 5% stop-gain reduces returns from +13.4% to +2.3%—an 11.1 percentage point loss.
Figure 1: Heatmap visualization clearly shows the optimal strategy (green top-right) and demonstrates how stop-gains consistently destroy performance as you move left across any row.
Three Trade Examples¶
ANET (Disaster Prevention): Actual -152.1% return → -3.0% with stop-loss = $14,909 saved per $10K
LULU (Winner Preserved): Actual +24.5% return → +24.5% unchanged = $0 impact (stop never triggered)
DKS (Moderate Loss): Actual -38.9% return → -3.0% with stop-loss = $3,592 saved per $10K
Why Stop-Gains Hurt Short Selling¶
Short selling has asymmetric risk: unlimited downside, limited upside. Stop-gains cap Vision's biggest winners (PTON declined 48.7%, TREX declined 40.4%) while stop-losses prevent unlimited disasters. Traditional symmetric risk management destroys value in this context.
Implementation Challenges¶
While 1% stop-loss provides optimal theoretical returns (+23.0% improvement), practical concerns favor 3% implementation:
- Market Noise: 1% moves often represent normal volatility, not fundamental deterioration
- Transaction Costs: Frequent stop triggers increase trading costs
- Liquidity Risk: Institutional positions need execution reliability during stress periods
The 3% threshold captures 96% of theoretical benefits (+22.1% vs +23.0%) while addressing real-world implementation constraints.
Takeaways for Institutional Investors¶
-
Research Provider Optimization: Skilled stock selection can coexist with poor risk management. Systematic exit strategies can unlock latent analytical value.
-
Asymmetric Risk Management: Short selling demands different approaches than long strategies. Eliminate downside disasters while preserving unlimited upside potential.
-
Stop-Gain Avoidance: Profit caps consistently reduce performance by preventing capture of research providers' biggest analytical successes.
-
Implementation Balance: Theoretical optimization (1% stops) may prove less valuable than practical optimization (3% stops) after considering real-world constraints.
-
Validation Requirements: Always test optimization strategies across multiple data splits before implementation to prevent overfitting to specific datasets.
The Vision Weekly analysis demonstrates how institutional investors can enhance external research relationships through systematic risk management while preserving analytical independence. The 22.1 percentage point improvement proves substantial enough to justify implementation complexity for institutional portfolios utilizing external short-selling research.