Drawdown Enhancements to Reporting Script

Summary of Changes

The reporting script has been enhanced to calculate and report:

1. ✅ Maximum Drawdown for Entire Strategy (already existed)
2. ✅ Maximum Drawdown for Each Year (newly added)

What Was Already There

Overall Maximum Drawdown

• Function: compute_drawdown() (line 370)
• Calculated in: compute_performance_metrics() (line 417-418)
• Output file: performance_summary.csv
• Column name: max_drawdown

Example performance_summary.csv:

start_date,end_date,trading_days,start_equity,end_equity,total_return,cagr,ann_vol,sharpe,max_drawdown
2020-01-02,2023-12-29,1015,100000,145000,0.45,0.1324,0.1856,0.713,-0.1245

The max_drawdown column shows -12.45%, meaning the worst peak-to-trough decline was 12.45%.

What Was Added

New Function: compute_yearly_drawdowns()

def compute_yearly_drawdowns(portfolio: pd.DataFrame) -> pd.DataFrame:
    """
    Compute maximum drawdown for each calendar year.
    
    Parameters:
    -----------
    portfolio : DataFrame with 'date' and 'equity' columns
    
    Returns:
    --------
    DataFrame with columns: year, max_drawdown
    """

This function:

1. Groups portfolio data by calendar year
2. Calculates the maximum drawdown within each year
3. Returns a DataFrame with one row per year

New Output File: yearly_drawdowns.csv

Example output:

year,max_drawdown
2020,-0.0856
2021,-0.0432
2022,-0.1567
2023,-0.0723

This shows:

• 2020: worst drawdown was -8.56%
• 2021: worst drawdown was -4.32%
• 2022: worst drawdown was -15.67% (worst year)
• 2023: worst drawdown was -7.23%

Output Files Summary

After running the enhanced script, you will have:

Overall Performance

• performance_summary.csv - Contains CAGR, overall max drawdown, Sharpe ratio, etc.

Returns

• yearly_returns.csv - Return for each calendar year
• monthly_returns.csv - Monthly return matrix

Drawdowns (NEW!)

• yearly_drawdowns.csv - Maximum drawdown for each calendar year

Other Files

• rolling_annualized_volatility.csv - 2-month rolling volatility
• daily_account_ledger.csv - Daily equity, margin, free cash
• trade_impact_ledger.csv - Trade-by-trade impact analysis
• Charts (if enabled)

How Drawdown is Calculated

Drawdown at time t is:

DD(t) = Equity(t) / RunningMax(Equity[0:t]) - 1.0

Maximum drawdown is the minimum value (most negative) over the period.

For example:

• Start with $100,000
• Peak at $120,000
• Drop to $108,000
• Drawdown = $108,000 / $120,000 - 1.0 = -0.10 or -10%

The yearly maximum drawdown finds the worst such decline within each calendar year.

Usage

Simply run the enhanced script as before:

python 6-reporting_script_with_yearly_drawdowns.py

The script will automatically generate all the files including the new yearly_drawdowns.csv.

Integration with Daily Statement Report

If you want to also display yearly drawdowns in your daily statement PDF report, you can:

1. Load yearly_drawdowns.csv in the daily statement script
2. Add a summary page showing yearly statistics (returns, volatility, drawdowns)
3. Display the max drawdown for the specific year on each daily statement page

Summary of Changes to Daily Report

Added load_summary_statistics() function - Loads performance metrics from CSV files Added draw_summary_statistics_page() function - Renders the summary page with all statistics Modified write_daily_statement_pdf() signature - Added summary_stats parameter Modified PDF generation - Summary stats page is now first, then volatility chart, then daily statements Modified main() function - Loads summary stats and passes them to PDF writer

The summary page will now show:

CAGR Total Return Sharpe Ratio Annual Volatility Overall Maximum Drawdown Table of yearly returns and drawdowns

All formatted cleanly on the first page of the PDF report!

USD Risk Volume Filtering

USAGE EXAMPLES: Example 1: No filtering (include all instruments) pythonCONFIG = { "trade_symbol_allowlist": None, "min_usd_risk_volume": 0, # 0 means no filtering } Example 2: Only trade instruments with >$1M daily USD risk pythonCONFIG = { "trade_symbol_allowlist": None, "min_usd_risk_volume": 1_000_000, # $1M minimum } Example 3: Only highly liquid instruments (>$50M daily USD risk) pythonCONFIG = { "trade_symbol_allowlist": None, "min_usd_risk_volume": 50_000_000, # $50M minimum # Will likely include: ES, NQ, CL, GC, etc. # Will exclude: Most agricultural, minor currencies, etc. } Example 4: Combine with allowlist (both filters must pass) pythonCONFIG = { "trade_symbol_allowlist": {"MES", "MNQ", "MCL", "MGC"}, # Specific micros "min_usd_risk_volume": 20_000_000, # AND must have >$20M USD risk # Instrument must be in allowlist AND meet USD risk threshold } Example 5: Test with single instrument pythonCONFIG = { "trade_symbol_allowlist": {"MES"}, "min_usd_risk_volume": 0, # No USD risk filtering }

OUTPUT: When you run the backtest, you'll see filtering messages like: Loading instruments... ✅ ES (trades as MES): USD risk $2,450,123,456 >= min $50,000,000 - INCLUDED ✅ NQ (trades as MNQ): USD risk $1,876,543,210 >= min $50,000,000 - INCLUDED 🚫 6A (trades as M6A): USD risk $15,234,567 < min $50,000,000 - FILTERED OUT ✅ CL (trades as MCL): USD risk $1,234,567,890 >= min $50,000,000 - INCLUDED 🚫 ZC (trades as MZC): USD risk $8,456,789 < min $50,000,000 - FILTERED OUT ... This lets you see exactly which instruments are being included/excluded based on your threshold.

IMPORTANT NOTES:

Set min_usd_risk_volume = 0 to include all instruments (no filtering) USD risk is based on most recent data in the file (last non-NaN value) Both filters apply: If you set both trade_symbol_allowlist and min_usd_risk_volume, the instrument must pass BOTH filters Make sure you ran 4a-compute_risk_volume.py first or you'll get warning messages The filtering happens at instrument loading time, so instruments are included/excluded for the entire backtest period based on their most recent USD risk value

TESTING YOUR THRESHOLD: To find a good threshold, you can:

Set min_usd_risk_volume = 0 and run backtest Look at the output - it will show USD risk for each instrument Adjust threshold based on which instruments you want to include Re-run with your chosen threshold

Typical thresholds:

$0: All instruments (no filtering) $10M: Filter out very illiquid markets $50M: Moderately liquid markets only $100M: Highly liquid markets only $500M+: Institutional-grade liquidity only