Currency · United States · CFDs, forwards, futures, swaps · Intraday · United States Free

Price Overreactions in the Forex

Academic paper

Price Overreactions in the Forex and Trading Strategies

AuthorsGuglielmo Maria Caporale; Oleksiy Plastun

Institute

DEGerman Institute for Economic Research
DEIfo Institute for Economic Research
London South Bank University
Brunel University of London
?Brunel University London - Department of Economics and Finance
?CESifo (Center for Economic Studies and Ifo Institute)
?German Institute for Economic Research (DIW Berlin)
UASumy State University

Paper linkhttps://ssrn.com/abstract=3362142

Strategy in a nutshell

Trade AUD/USD by detecting overreaction days—when daily returns exceed the 50-day average plus two standard deviations (Wong, 1997). Open a position at 17:00 in the direction of the overreaction and close by day’s end, capturing intraday spikes in volatility.

Economic rationale

Overreaction days exhibit strong intraday momentum driven by behavioral biases. Prices tend to continue in the overreaction direction, creating exploitable market inefficiencies. This approach leverages psychological and behavioral patterns in financial markets for consistent short-term gains.

Backtest performance

Annualised return4.78%

Volatility6.3%

Beta0.143

Sharpe ratio0.76

Sortino ratio-0.197

Win rate47%

Full Python code

1from AlgorithmImports import *
2class PriceOverreactionsintheForex(QCAlgorithm):
3def initialize(self):
4self.set_start_date(2009, 1, 1)
5self.set_cash(100000)
6self.period: int = 50
7
8self.symbol: Symbol = self.add_forex('AUDUSD', Resolution.MINUTE, Market.OANDA).symbol
9self.securities[self.symbol].set_fee_model(CustomFeeModel())
10
11self.daily_returns: RollingWindow = RollingWindow[float](self.period)
12self.open_price: float | None = None
13
14self.settings.minimum_order_margin_portfolio_percentage = 0.
15self.schedule.on(self.date_rules.every_day(self.symbol), self.time_rules.at(17, 0), self.at17)
16    
17def at17(self):
18history : DataFrame = self.history(self.symbol, self.period, Resolution.DAILY)
19if len(history) == self.period:
20    if 'close' in history:
21        closes: np.ndarray = np.array(history['close'].values)
22        daily_returns: np.ndarray = closes[1:] / closes[:-1] - 1
23        avg_ret: float = np.average(daily_returns)
24        std: float = np.std(daily_returns)
25        
26        price: float = self.securities[self.symbol].close
27        ret: float = price / closes[-1] - 1
28        
29        if ret > avg_ret + 2*std: 
30            self.set_holdings(self.symbol, 1)
31            holdings_q: float = self.portfolio[self.symbol].quantity
32            self.market_on_close_order(self.symbol, -holdings_q)
33# Custom fee model
34class CustomFeeModel(FeeModel):
35def get_order_fee(self, parameters):
36fee = parameters.security.price * parameters.order.absolute_quantity * 0.00005
37return OrderFee(CashAmount(fee, "USD"))