Skip to content

STRATEGY_BACKTEST_RESULTS.md

Latest commit

 

History

History
111 lines (86 loc) · 4.83 KB

STRATEGY_BACKTEST_RESULTS.md

File metadata and controls

111 lines (86 loc) · 4.83 KB

Strategy Backtest Results

This page records published backtest evidence for built-in Marketpal strategy configs. Backtests are research artifacts, not investment advice or trading instructions. Results depend on the supplied universe, date range, price history, fee/slippage assumptions, and strategy version.

May 2026 Watchlist Momentum Search

This run searched weekly pure-momentum configs on a 95-ticker high-conviction watchlist universe.

  • Primary window: 2025-12-03 to 2026-05-12
  • Initial capital: 100,000
  • Rebalance cadence: weekly
  • Price mode: close
  • Backtest costs: 5 bps fee and 10 bps slippage
  • Strategy family: pure momentum, with quality/value/reversion weights set to zero

The full-window CAGR figures are annualized from a short period. They should not be read as a sustainable annual return estimate.

Strategy Role Full Return CAGR Sharpe Max DD Trades Final Equity Late Validation Recent 1-Week
aggressive_momentum_weekly_v1 Top raw full-window result 116.67% 352.88% 3.81 -11.25% 149 216,671.04 46.97% -0.27%
active_momentum_weekly_v1 Selected robust config 112.31% 335.25% 3.80 -12.02% 157 212,310.96 49.24% -0.27%
momentum_only_v1 Built-in benchmark 44.74% 105.91% 2.82 -10.75% 168 144,735.95 22.27% -0.04%

Late validation covers the later slice of the same research period. The recent one-week check is intentionally narrow; the benchmark was slightly better there because it stayed mostly in cash while the active configs deployed more capital.

Configs

aggressive_momentum_weekly_v1 maps to the top raw-search config:

mom2_mb0p55_mh0p1_mp10_mw0p18_mn4_st0p02_tr0p05_to0p45

Key settings:

  • max_positions: 10
  • max_position_pct: 0.18
  • min_buy_score: 0.55
  • min_hold_score: 0.10
  • max_new_positions_per_run: 4
  • starter_position_pct: 0.02
  • max_single_trade_pct: 0.05
  • turnover_budget_pct: 0.45
  • cash_buffer_pct: 0.02

active_momentum_weekly_v1 maps to the selected robust config:

mom2_mb0p55_mh0p1_mp14_mw0p18_mn4_st0p02_tr0p04_to0p45

Key settings:

  • max_positions: 14
  • max_position_pct: 0.18
  • min_buy_score: 0.55
  • min_hold_score: 0.10
  • max_new_positions_per_run: 4
  • starter_position_pct: 0.02
  • max_single_trade_pct: 0.04
  • turnover_budget_pct: 0.45
  • cash_buffer_pct: 0.02

aggressive_momentum_weekly_v1 produced the highest full-window return. The slightly more diversified active_momentum_weekly_v1 was kept because it had a stronger late-validation result and a lower single-trade cap.

Final Holdings

Final holdings at 2026-05-12, sorted by ending portfolio weight:

Strategy Final Holdings
aggressive_momentum_weekly_v1 DOCN 18.4%, INTC 17.7%, LITE 16.2%, BE 16.1%, AMD 16.0%, APLD 6.8%, CRWV 6.4%, HZN.AX 0.4%, DRO.AX 0.0%, ISRG 0.0%, GOOGL 0.0%, cash about 2.0%
active_momentum_weekly_v1 DOCN 18.8%, BE 15.0%, INTC 14.7%, LITE 14.0%, MU 13.8%, AMD 8.3%, AMZN 5.3%, GNP.AX 3.0%, KAR.AX 2.2%, ARU.AX 1.5%, HZN.AX 1.0%, PLS.AX 0.2%, GOOGL 0.0%, ISRG 0.0%, cash about 2.0%
momentum_only_v1 DOCN 12.2%, INTC 11.7%, LITE 11.4%, BE 11.3%, KAR.AX 10.4%, MU 8.9%, AMD 7.2%, SXE.AX 5.7%, GOOGL 4.5%, LYC.AX 4.4%, SNDK 3.3%, NHC.AX 3.1%, AMZN 2.0%, WDS.AX 1.2%, MIN.AX 0.6%, cash about 2.0%

Small near-zero residual positions can appear from price movement, minimum trade thresholds, and cash-buffer constraints.

Built-In Strategy Coverage

The table below is intentionally conservative. A strategy is marked "documented" only when this repo records a comparable result set for that strategy.

Strategy Documented Result Status
aggressive_momentum_weekly_v1 Documented in the May 2026 watchlist momentum search
active_momentum_weekly_v1 Documented in the May 2026 watchlist momentum search
momentum_only_v1 Documented as the benchmark in the May 2026 watchlist momentum search
momentum_profile_v1 No comparable published result in this page
simple_score_v1 No comparable published result in this page
best_weekly_swing_v1 No comparable published result in this page
best_monthly_swing_v1 No comparable published result in this page
engine_weekly_swing_v1 No comparable published result in this page
engine_quality_swing_rebuild_v1 No comparable published result in this page
portfolio_low_churn_swing_v1 No comparable published result in this page
engine_quality_value_reversion_v1 No comparable published result in this page
portfolio_quality_value_reversion_v1 No comparable published result in this page

Before presenting a strategy as having historical evidence, run a fresh backtest on the target universe and record the same assumptions, date range, metrics, and final holdings.