Dual-Regime Weekly Strategy Backtest UNH 40.81% Return With 4.15% Max Drawdown

Get the full strategy library: This article uses one strategy from the Mega Backtrader Strategy Pack, a package of 500+ Backtrader-ready Python trading strategies with batch backtesting, dashboards, metrics, charts, and documentation.

Download the package here: Mega Backtrader Strategy Pack

If you want to move from one-off trading ideas to a repeatable strategy research workflow, the package gives you strategy code, batch runners, metrics, charts, and publishable reports in one place.

Markets do not behave the same way all the time. Some periods reward trend following. Other periods punish breakouts and reward mean reversion. That is the idea behind a dual-regime strategy: detect the current market condition first, then choose the trading logic that fits that condition.

This article walks through DualRegimeWeeklyStrategy, one of the strategies included in the Mega Backtrader Strategy Pack. The strategy was tested on UNH daily candles over a 1-year period, with SPY used as the benchmark.

The result:

This is a strong example because the strategy beat both the asset buy-and-hold return and the SPY benchmark while keeping maximum drawdown low. The run ended with one open position, so the final equity includes the marked value of that open trade. The win rate is based on the 13 closed trades.

Strategy Idea

DualRegimeWeeklyStrategy uses a weekly decision schedule. It does not react to every daily candle. Instead, it checks conditions once per week, closes the previous position, determines the current market regime, and then decides whether to enter a new trade.

The strategy uses two different playbooks:

1. Trending regime: If ADX is above the threshold, the strategy uses price versus EMA to decide direction.
2. Ranging regime: If ADX is below the threshold, the strategy looks for Bollinger Band and RSI extremes.

That structure makes the strategy easy to understand. It first asks whether the market is trending or ranging. Only after that does it choose which signal type matters.

Strategy Parameters

The strategy exposes its behavior through Backtrader params:

params = (
    ('adx_period', 14),
    ('adx_threshold', 25),
    ('rebal_day', 0),
    ('ema_period', 50),
    ('bb_period', 20),
    ('bb_devfactor', 2.0),
    ('rsi_period', 14),
    ('rsi_oversold', 30),
    ('rsi_overbought', 70),
    ('atr_period', 14),
    ('atr_mult', 3.0),
)

These settings control:

• The ADX lookback and threshold used to classify trend strength.
• The weekday used for the weekly rebalance check.
• The EMA period used during trending markets.
• The Bollinger Band and RSI settings used during ranging markets.
• The ATR settings available for volatility-aware extensions.

Code Walkthrough

The strategy starts by creating the indicators it needs for both regimes:

self.adx = bt.indicators.ADX(self.data, period=self.p.adx_period)
self.ema = bt.indicators.ExponentialMovingAverage(
    self.data.close,
    period=self.p.ema_period,
)
self.bb = bt.indicators.BollingerBands(
    self.data.close,
    period=self.p.bb_period,
    devfactor=self.p.bb_devfactor,
)
self.rsi = bt.indicators.RSI(self.data.close, period=self.p.rsi_period)
self.atr = bt.indicators.ATR(self.data, period=self.p.atr_period)

ADX is the regime filter. EMA is used when the market is trending. Bollinger Bands and RSI are used when the market is ranging.

The strategy then limits decision-making to one chosen weekday:

dt = self.data.datetime.date(0)
current_week = dt.isocalendar()[1]

if dt.weekday() != self.p.rebal_day:
    return
if self.last_week == current_week:
    return

self.last_week = current_week

With rebal_day set to 0, the strategy evaluates on Monday. This avoids constant churn from daily signal noise and gives the system a cleaner weekly rhythm.

Before opening a new trade, the strategy closes any existing position:

if self.position:
    self.close()

That makes each weekly decision explicit. The strategy does not keep stacking unrelated entries. It exits the old position first, then evaluates the new setup.

The regime switch is controlled by ADX:

regime = 'trending' if self.adx[0] > self.p.adx_threshold else 'ranging'

When the market is trending, the strategy uses the EMA as the directional filter:

if regime == 'trending':
    if self.data.close[0] > self.ema[0]:
        self.order = self.buy()
    else:
        self.order = self.sell()

In the original strategy code, that bearish branch can issue a sell order. In the package batch run used for this article, the runner was in its default long-only mode, so sell orders are treated as exits rather than new short entries. That means the reported result is a long-only interpretation of the strategy logic.

When the market is ranging, the strategy waits for oversold or overbought extremes:

else:
    if self.data.close[0] <= self.bb.bot[0] and self.rsi[0] < self.p.rsi_oversold:
        self.order = self.buy()
    elif self.data.close[0] >= self.bb.top[0] and self.rsi[0] > self.p.rsi_overbought:
        self.order = self.sell()

The long entry needs price near the lower Bollinger Band and RSI below the oversold threshold. The sell branch captures the opposite condition, and in the long-only batch run it functions as an exit signal rather than a fresh short.

Why This Strategy Is Interesting

Many simple strategies use only one market assumption. A moving-average strategy assumes trend continuation. A Bollinger Band reversion strategy assumes mean reversion. Both ideas can work, but neither describes every market.

DualRegimeWeeklyStrategy is useful because it combines both ideas without making the code hard to follow:

1. Use ADX to decide whether the market is trending.
2. Use EMA direction in trending markets.
3. Use Bollinger Band and RSI extremes in ranging markets.
4. Recheck the decision once per week.

That weekly schedule is important. It reduces overtrading and makes the strategy easier to inspect in the results dashboard.

Backtest Setup

The run used the package batch backtester with these settings:

The package generated the metrics table, equity curve, benchmark comparison, drawdown chart, rolling return chart, and daily return distribution automatically.

Results

The key number is not only the 40.81% return. The drawdown stayed at 4.15%, which is low for a strategy that outperformed both UNH buy-and-hold and SPY over the test window.

The single open trade at the end matters. Closed-trade statistics describe the 13 completed trades, while the final portfolio value includes the open position's mark-to-market value at the final bar.

Equity Curve vs Benchmark

The equity curve shows the strategy growing from $10,000 to $14,080.92. UNH buy-and-hold also performed well, but the strategy finished higher with a smoother path.

Drawdown vs Benchmark

The drawdown chart is the strongest part of this case study. The strategy's maximum drawdown was only 4.15%, while both the asset and benchmark had deeper pullbacks during the same period.

Rolling Return vs Benchmark

The rolling return chart shows how performance developed through the test window. The strategy did not rely on a single lucky spike. It built the final result across multiple weekly decisions.

Daily Return Distribution

The daily return histogram gives a compact view of day-to-day movement in the strategy equity curve. It is useful for quickly checking whether the result came from frequent smaller moves or a handful of extreme days.

What This Shows

This backtest is a good example of why a large strategy library is useful. The idea is not only to find one strategy and stop. The value is in being able to test many different structures quickly:

1. Trend following.
2. Mean reversion.
3. Volatility breakout.
4. Weekly rebalancing.
5. Regime switching.
6. Long-only and short-capable variants.

DualRegimeWeeklyStrategy shows how a compact Backtrader class can combine multiple market assumptions in one readable strategy. It is still simple enough to modify, but it already has enough structure to behave differently across market regimes.

Get the Complete Strategy Library

DualRegimeWeeklyStrategy is one of 500+ Backtrader-ready strategies included in the Mega Backtrader Strategy Pack.

Get the full package here: Mega Backtrader Strategy Pack

The package includes:

• 500+ Python strategy modules in the strategies/ package.
• Fast batch backtesting with multiprocessing.
• Per-strategy metrics, equity curves, drawdown plots, and CSV outputs.
• A browsable HTML strategy catalog.
• A Markdown strategy package manual.
• Windows and shell batch runners for automated research.

If you want to test more strategy ideas faster, this package gives you the code and the research workflow.

Get the Mega Backtrader Strategy Pack

Disclaimer

This article is for research and educational use only. Backtest results are not financial advice, investment advice, or a guarantee of future performance. Always validate strategy logic, data quality, execution assumptions, costs, slippage, and risk before using any trading system.