Big‑picture summary (Turtle Trading Style)
Time Span for Test: 1/1/2023 to 30/06/2025
Goal: Trade a Turtle‑style breakout on a top‑25, high‑liquidity universe using daily data.
Entry: Go long when the current price breaks above the rolling 30‑day highest close.
Exit (priority order):
Stop‑loss at −25% from average entry price.
Take‑profit at +30% from average entry price.
Exit if price falls below the rolling 15‑day lowest close.
Position sizing: Target 20% of portfolio value per position.
——— QuantConnect Code Generated by ChatGPT is below ———-
BackTest results from QuantConnect:
https://www.quantconnect.cloud/backtest/eebcd7650ae287029cf1d929f66de19e/?theme=chrome
Embed Code:
https://www.quantconnect.cloud/backtest/eebcd7650ae287029cf1d929f66de19e/?theme=chrome
In case the result links can be lost I copied the result front end.
As you can see 1.000.000 USD Becomes 2.718.001 USD in 30 months in test cycle.
from AlgorithmImports import *
from datetime import timedelta
from QuantConnect.Data.Consolidators import TradeBarConsolidator
class TurtleSP500(QCAlgorithm):
def Initialize(self):
# --- Backtest setup ---
self.SetStartDate(2023, 1, 1)
self.SetEndDate(2025, 6, 30)
self.SetCash(1_000_000)
# Warm up enough bars for Donchian windows
self.SetWarmUp(60, Resolution.Daily)
self.SetBrokerageModel(BrokerageName.INTERACTIVE_BROKERS_BROKERAGE, AccountType.MARGIN)
# --- Universe ---
self.UniverseSettings.Resolution = Resolution.Daily
self.AddUniverse(self.CoarseSelectionFunction)
# --- Strategy parameters ---
self.entryLen = 30 # 30-day high for entry (close-based)
self.exitLen = 15 # 15-day low for exit (close-based)
self.positionSize = 0.20 # 10% per trade
# --- 1% TP/SL parameters ---
self.takeProfitPct = 0.30 # +20% from avg entry
self.stopLossPct = 0.25 # -20% from avg entry
# --- Storage ---
self.symbolData = {}
# --- Schedule daily trading ---
self.Schedule.On(
self.DateRules.EveryDay(),
self.TimeRules.AfterMarketOpen("SPY", 30),
self.ScanAndTrade
)
# === Universe selection: Top 25 by dollar volume ===
def CoarseSelectionFunction(self, coarse):
sorted_by_dollar_volume = sorted(coarse, key=lambda x: x.DollarVolume, reverse=True)
top = sorted_by_dollar_volume[:25]
return [c.Symbol for c in top]
# === Handle securities added/removed from universe ===
def OnSecuritiesChanged(self, changes):
for sec in changes.AddedSecurities:
symbol = sec.Symbol
if symbol not in self.symbolData:
self.symbolData[symbol] = SymbolData(self, symbol, self.entryLen, self.exitLen)
for sec in changes.RemovedSecurities:
symbol = sec.Symbol
if symbol in self.symbolData:
# Clean up consolidator and liquidate if needed
self.symbolData[symbol].Dispose()
self.Liquidate(symbol)
del self.symbolData[symbol]
# === Daily scan & trade logic ===
def ScanAndTrade(self):
if self.IsWarmingUp:
return
for symbol, data in list(self.symbolData.items()):
if symbol not in self.Securities:
continue
security = self.Securities[symbol]
if not data.IsReady or not security.IsTradable:
continue
price = security.Price
upper = data.entryDonchianHigh() # 20-day highest close (rolling)
lower = data.exitDonchianLow() # 5-day lowest close (rolling)
holding = self.Portfolio[symbol]
invested = holding.Invested
# --- Entry: price breaks above 20-day high ---
if not invested and price > upper:
self.SetHoldings(symbol, self.positionSize)
self.Debug(f"BUY {symbol.Value} @ {price:.2f}")
# --- Exits: 1% stop-loss, 1% take-profit, or Donchian 5-day low ---
elif invested:
avg_price = holding.AveragePrice
if avg_price and avg_price > 0:
sl_level = avg_price * (1 - self.stopLossPct)
tp_level = avg_price * (1 + self.takeProfitPct)
# Precedence: stop-loss -> take-profit -> Donchian
if price <= sl_level:
self.Liquidate(symbol)
self.Debug(f"SELL {symbol.Value} @ {price:.2f} [StopLoss {self.stopLossPct*100:.0f}%]")
continue
if price >= tp_level:
self.Liquidate(symbol)
self.Debug(f"SELL {symbol.Value} @ {price:.2f} [TakeProfit {self.takeProfitPct*100:.0f}%]")
continue
# Donchian exit (secondary)
if price < lower:
self.Liquidate(symbol)
self.Debug(f"SELL {symbol.Value} @ {price:.2f} [Donchian-{self.exitLen}d]")
# === Helper class for Donchian channels (with daily rolling updates) ===
class SymbolData:
def __init__(self, algo: QCAlgorithm, symbol: Symbol, entryLen: int, exitLen: int):
self.algo = algo
self.symbol = symbol
self.entryLen = entryLen
self.exitLen = exitLen
# Close-based windows (match original definition)
self.entryWindow = RollingWindow[float](entryLen) # forighest close
self.exitWindow = RollingWindow[float](exitLen) # for 5-day lowest # 1) Seed with historical daily bars (DataFrame -> iterate itertuples)
lookback = max(entryLen, exitLen) + 1
history = algo.History(symbol, lookback, Resolution.Daily)
if not history.empty:
for bar in history.itertuples():
# bar.close is a pandas namedtuple field
self._add_close(float(bar.close))
# 2) Add a daily consolidator so the windows roll forward each day
self.consolidator = TradeBarConsolidator(timedelta(days=1))
self.consolidator.DataConsolidated += self._on_daily_bar
algo.SubscriptionManager.AddConsolidator(symbol, self.consolidator)
def _on_daily_bar(self, sender, bar: TradeBar):
self._add_close(float(bar.Close))
def _add_close(self, close: float):
self.entryWindow.Add(close)
self.exitWindow.Add(close)
@property
def IsReady(self) -> bool:
return (
self.entryWindow.Count == self.entryWindow.Size and
self.exitWindow.Count == self.exitWindow.Size
)
# --- Donchian methods (close-based) ---
def entryDonchianHigh(self) -> float:
return max(x for x in self.entryWindow)
def exitDonchianLow(self) -> float:
return min(x for x in self.exitWindow)
def Dispose(self):
# Clean up the consolidator when symbol leaves the universe
if getattr(self, "consolidator", None) is not None:
self.algo.SubscriptionManager.RemoveConsolidator(self.symbol, self.consolidator)
self.consolidator = None