The strategy,
with the evidence attached.

What the algorithm does.

Stock selection

PIPPO uses a confirmed bullish internal structure break and a close above SMA111 for entry eligibility. Confirmed bearish structure breaks supply exits. The study allows up to five equal-weight positions per sleeve and score-based replacements.

Tactical momentum

The daily variant gates exposure using the prior-session IHSG SMA25 > SMA50 relationship. Its research universe is unrestricted; it includes illiquid stocks, so this is not an executable stock basket.

Defensive core

Point-in-time KOMPAS100 membership plus stocks outside it with a 20-session median traded value of at least Rp 5 billion. It combines a prior completed monthly MA55 gate, market breadth, turnover controls, a 10-session minimum hold and a 12% volatility target. A target is not a guaranteed volatility or loss limit.

Separate sleeves—not a combined curve

The displayed curve places the normalized initial capital of 100 entirely in the tactical sleeve. The separate combined study starts one-third tactical and two-thirds defensive, with no transfers between sleeves; its weights drift as returns differ. Its results are not the tactical-only figures shown above.

Read this before using the numbers.

• Costs are modeled, not verified fills. The displayed tactical test includes 0.09% commission and 0.10% ordinary slippage per transaction. Separately, the defensive simulator uses the same ordinary costs and configures 0.25% additional stop slippage and opening-price fills for gaps through stops. The combined account uses already-costed sleeve returns; no extra top-level charge is added. Taxes, capacity and all real-world execution costs are not independently verified.
• Liquidity is the main constraint. The all-stock tactical universe contains securities that may not fill at modeled prices or in practical size. High simulated returns do not remove that limitation.
• Losses are substantial. This daily-gated tactical sleeve lost 55.15% from a prior peak at its worst. The defensive component itself had a 33.68% maximum drawdown; “defensive” does not mean safe.
• Historical research is not independent validation. These reports were produced by the strategy’s own research project. Parameter selection, incomplete security lifecycle/corporate-action treatment and historical data quality can bias results. No survivorship-free full-IDX claim or verified out-of-sample/live return is established.
• Backtest, paper and live are different. This curve is not the workspace’s current cash balance, trade ledger or user account. Actual trading can differ because of fills, capacity, spreads, costs and changing market conditions.
• Past performance does not guarantee future results. This is shared research, not personalized investment advice. Karsa does not execute broker orders.

Where the evidence comes from.

The original pemantau-garis artifact names this series “Tactical unfiltered - daily SMA25/50.” This is the daily-gated tactical component, not the report’s combined account. We chose the entire available cohort, 2016-02-01 to 2026-08-28, rather than its higher-return later subsets.

The combined daily-gate portfolio reported 69.05% CAGR and −51.54% maximum drawdown over the same period, starting one-third tactical and two-thirds defensive. The previous monthly-gate combined variant reported 70.21% CAGR and −69.24% maximum drawdown over the same period. The earlier single-sleeve PIPPO result (23.29% CAGR, −72.73% drawdown, 2016–2026) is a different experiment with different execution assumptions. None of those alternative results is substituted into this tactical curve.

Source repository

thoriqfy/pemantau-garis

Research script revision

89e12e5 · 1 September 2026; artifact hashes below establish file identity, not proof that the output was committed at that revision.

Source files

outputs/deployed_two_sleeve_daily_tactical/summary.csv
equity_curves.csv and strategy_manifest.json

CAGR = (final equity / initial capital)^{1 / elapsed years} − 1, with elapsed years measured using 365.25 days. Drawdown uses the full daily series and running peaks. The first plotted session already includes modeled costs and is slightly below the initial capital of 100. The chart uses a logarithmic scale to show proportional changes, not equal currency increments.