jawsnfl / backtrader-finstrat

Trading strategies, back-tested with Backtrader

backtrader-finstrat

Active Stratefy - Residual (idiosyncratic) momentum based, with tactical asset allocation, pyramid positioning (Bollinger Bands)

Source code - strategy_idiosync_m_gtaa_py_bb_atr_stoploss.py

Risk Management

0. Quarterly (rebalance) checks

   • If the global market is in a downtrend, don't rebalance and continue holding safe assets.
   • Individual stocks - exclude from consideration stocks in downtrend (SMA < 200 or 150)

1. Monthly checks (beginning of each month)

   • if the global market is in a downtrend (< 200SMA), exist equities and buy safe assets.
   • If the global market switched to an uptrend (> 200SMA), rebalance back into risky assets, even before the official rebalance.

Reasoning - don't catch a falling knife, however, monthly intervals allow room for the market to recover if briefly touching SMA200.

2. Daily checks
   • Checks trailing high against deviation of scaled average true range ATR(90) of the stock.
     • ATR is more robust than using constant trailing % loss, since it takes into account historical volatility of specific stock.
     • The larger is number of days from trailing high, the more likely price to break the ATR
       • To count that, ATR is scaled by number of days (up-to 6) from the high
     • ATR approach is more sound theoretically and is robust under different universes, as opposed to constant 20%, which was more tuned to the extended universe
     • It is mechanism that backs up the global (monthly) risk management in a robust manner.

Positioning

• If stock is in a downtrend at the purchase decision, wait till it reaches the lower limit of its Bollinger Band before allowing the purchase.
• If the purchase is enabled, perform pyramid positioning - if the asset's price moved X% above the current target price, buy an asset's proportion and set a new target price.

Profit taking

Profit-taking is happening in 2 cases:

1. Sell at the rebalance, due to the current asset wasn't picked by the momentum algorithm.
2. Profit-taking if an asset appreciated 30% compared to its buy price (for any continuous period, not just a quarter)

Past strategy - Absolute momentum based, with tactical asset allocation, pyramid positioning (Bollinger Bands)

Source code - strategy_momentum_gtaa_py_bb_atr_stoploss.py

• Shares the same features as the main strategy, just the momentum calculation is different.

Past strategy 2 - momentum based, with tactical asset allocation, pyramid positioning

Source code - strategy_momentum_gtaa_pyramiding.py

Risk Management

1. Monthly checks (beginning of each month)
   • if the global market is in a downtrend (< 200SMA), exist equities and buy safe assets.
   • If the global market switched to an uptrend (> 200SMA), rebalance back into risky assets.

Reasoning - don't catch a falling knife, however, monthly intervals allow some room for the market to recover if briefly touching SMA200.

2. Daily checks

   • If individual asset's trailing decline > 25%, close position.

3. Quarterly - if the global market is in a downtrend, don't rebalance, continue holding safe assets.

Positioning

• Pyramid positioning - if the price of the asset moved X% above starting price, buy a portion of the asset and set a new target price. While absolute returns are a bit lower, the strategy has improved risk-adjusted returns and significantly lower maximum drawdown. Backtested on multiple universes.

Profit taking

Profit-taking is happening in 2 cases:

1. Sell at the rebalance, because the current asset wasn't picked by the momentum algorithm.
2. Profit-taking if an asset appreciated 30% compared to its buy price (for any continuous period, not just 3 months)