Turtle trading strategy
On GC futures from January 2016 to July 2026, the Turtles' System 2 made $89,101 after costs and made money in 5 of 11 years.
TrendSwingDaily barsBacktested Nov 6, 2026
Turtle trading backtest, the result
Contract
Gold
-
+$89.1K
Net after costs
After $1,559 in fees and slippage
-
5 of 11
Years that made money
January 2016 to July 2026
-
−$166K
Largest drawdown
Peak to trough, 1 contract
-
$908
Average trade
75 trades
Net after costs by month, January 2016 to July 2026.
Turtle trading rules
Buy a break of the last 55 days' high, sell short a break of their low, and exit on a 20-day break the other way or a stop 2 ATRs away.
A drawing of the rules on a made-up day, not a result.
- A one-second close above the last 55 days' high buys one contract, and a close below their low sells one short.
- A long is closed at the first close below the last 20 days' low, a short at the first close above their high.
- The stop is 2 N from the entry, where N is the 20-day average true range.
- A roll closes the position, and the new contract's first new day takes the same trade back.
- From
- Curtis Faith, on the rules Richard Dennis and William Eckhardt taught, Way of the Turtle (2007)
- Decides on
- Daily bars
- Holds
- Overnight and across days
- Parameters
- session minute 1,080, entry days 55, exit days 20, stop atr 2
Results by year
It made money in 5 of 11 years on GC. Its best year was 2025, +$105,327, and its worst 2017, −$14,042.
| Year | Net | Trades | Win rate | Drawdown | Buy and hold |
|---|---|---|---|---|---|
| 2016 | −$833 | 8 | 38% | −$14,096 | +$6,888 |
| 2017 | −$14,042 | 6 | 17% | −$14,382 | +$13,120 |
| 2018 | +$8,330 | 7 | 57% | −$9,108 | −$5,270 |
| 2019 | +$9,108 | 8 | 50% | −$10,182 | +$19,820 |
| 2020 | −$8,699 | 10 | 30% | −$35,010 | +$30,139 |
| 2021 | −$5,105 | 6 | 33% | −$16,506 | −$8,111 |
| 2022 | +$5,989 | 7 | 57% | −$24,705 | −$4,541 |
| 2023 | −$12,565 | 9 | 33% | −$25,359 | +$13,369 |
| 2024 | +$6,935 | 6 | 50% | −$33,863 | +$43,279 |
| 2025 | +$105,327 | 5 | 80% | −$49,630 | +$151,669 |
| 2026, Jan to Jul | −$5,343 | 3 | 33% | −$166,083 | −$38,803 |
Turtle trading backtest results
On GC it made $90,660 before costs. Fees and slippage took $1,559 of that.
- Net = before costs − fees − slippage
- GC: $90,660 − $309 − $1,250 = $89,101
The interval holds zero: this test cannot tell the average trade from no edge.
Turtle trading code and Pine Script
The strategy as the backtest ran it, and the Pine Script our compiler wrote from it for TradingView.
1from atlantive import BUY, NEW_YORK, SELL, Bars, RollingMinMax, Session, Strategy2 3 4class TurtleTrading(Strategy):5 """The Turtles' System 2: a 55-day breakout, a 20-day exit and a stop 2 ATR from the entry.6 7 A day is the CME trading day, 18:00 to 17:00 New York, built from one-second bars. The first8 one-second close above the last 55 days' high buys one contract, and the first close below their9 low sells one short. A long is closed at the first close below the last 20 days' low, or 2 N below10 its entry; a short mirrors it. N is the 20-day ATR at the entry. One contract: no units are added,11 which the Turtles' rules also did as a position won. Every level is read off finished days, so12 today's own range never moves it. A roll closes the position; the trade the rules hold is kept13 apart from it, and the new contract's first new day takes it back.14 """15 16 data = "bars"17 18 class Params:19 session_minute: int = 108020 entry_days: int = 5521 exit_days: int = 2022 stop_atr: float = 2.023 clip: int = 124 25 class State:26 clock: Session(zone=NEW_YORK)27 bars: Bars(interval_ns=1_000_000_000)28 high55: RollingMinMax(window=55)29 low55: RollingMinMax(window=55)30 high20: RollingMinMax(window=20)31 low20: RollingMinMax(window=20)32 atr: float33 seed: float34 days: int35 day_key: int36 day_high: float37 day_low: float38 day_close: float39 last_close: float40 held: int41 stop: float42 43 def on_event(self, view, ev, orders):44 self.clock.update(ev)45 if not self.bars.update(ev):46 return47 day = self.clock.prev_day48 if self.clock.prev_minute >= self.p.session_minute:49 day = day + 150 high = float(self.bars.prev_high)51 low = float(self.bars.prev_low)52 close = float(self.bars.prev_close)53 new_day = day != self.day_key54 if new_day:55 if self.day_key != 0:56 self.high55.update(self.day_high)57 self.low55.update(self.day_low)58 self.high20.update(self.day_high)59 self.low20.update(self.day_low)60 # N: Wilder's 20-day average of the true range, seeded by the mean of the first 20.61 true_range = self.day_high - self.day_low62 if self.days > 0:63 true_range = max(true_range, abs(self.day_high - self.last_close), abs(self.day_low - self.last_close))64 if self.days < 20:65 self.seed += true_range66 self.atr = self.seed / 20.067 else:68 self.atr = (self.atr * 19.0 + true_range) / 20.069 self.last_close = self.day_close70 self.days += 171 self.day_key = day72 self.day_high = high73 self.day_low = low74 else:75 self.day_high = max(self.day_high, high)76 self.day_low = min(self.day_low, low)77 self.day_close = close78 if self.days < self.p.entry_days:79 return80 opened = 081 if self.held > 0:82 if close < self.low20.min() or close < self.stop:83 self.held = 084 elif self.held < 0:85 if close > self.high20.max() or close > self.stop:86 self.held = 087 elif close > self.high55.max():88 self.held = 189 self.stop = close - self.p.stop_atr * self.atr90 opened = 191 elif close < self.low55.min():92 self.held = -193 self.stop = close + self.p.stop_atr * self.atr94 opened = 195 if self.held == 0 and view.position > 0:96 orders.ioc(side=SELL, qty=abs(view.position))97 elif self.held == 0 and view.position < 0:98 orders.ioc(side=BUY, qty=abs(view.position))99 elif self.held != 0 and view.position == 0 and (opened == 1 or new_day):100 if self.held > 0:101 orders.ioc(side=BUY, qty=self.p.clip)102 else:103 orders.ioc(side=SELL, qty=self.p.clip) 1//@version=62// TurtleTrading, exported from Atlantive.3// Chart simulation: decisions use completed bars; market orders fill at the next bar open.4// Native fills, fees, data gaps and session boundaries can differ. Re-test on the target symbol.5// Use standard time-based candles and the strategy's timeframe, not synthetic chart prices.6// Pine v6 trims old trade records at its order limit; the strategy keeps calculating.7// Margin is explicitly disabled to retain the native signal simulation's sizing.8strategy("TurtleTrading", overlay=true, initial_capital=100000.0, commission_type=strategy.commission.cash_per_contract, commission_value=1.96, default_qty_type=strategy.fixed, default_qty_value=1, calc_on_every_tick=false, process_orders_on_close=false, slippage=0, margin_long=0, margin_short=0)9 10if timeframe.in_seconds() != 111 runtime.error("TurtleTrading was exported for the 1 second timeframe; set the chart to 1S.")12 13session_minute = input.int(1080, "session_minute", minval=-9007199254740991, maxval=9007199254740991)14entry_days = input.int(55, "entry_days", minval=-9007199254740991, maxval=9007199254740991)15exit_days = input.int(20, "exit_days", minval=-9007199254740991, maxval=9007199254740991)16stop_atr = input.float(2.0, "stop_atr")17clip = input.int(1, "clip", minval=-9007199254740991, maxval=9007199254740991)18 19_q_safe_int(x) =>20 if math.abs(x) > 900719925474099121 runtime.error("an integer left the exact chart range; this run cannot be compared faithfully")22 int(x)23_q_int(x) => _q_safe_int(x < 0 ? math.ceil(x) : math.floor(x))24_q_iadd(a, b) => _q_safe_int(a + b)25 26_q_high = _q_int(math.round(high * 1000000000.0))27_q_low = _q_int(math.round(low * 1000000000.0))28_q_close = _q_int(math.round(close * 1000000000.0))29 30var float atr = 0.031var float seed = 0.032var int days = 033var int day_key = 034var float day_high = 0.035var float day_low = 0.036var float day_close = 0.037var float last_close = 0.038var int held = 039var float stop = 0.040// clock: completed chart bar's opening timestamp (America/New_York).41_q_clock_prev_minute = hour(time, "America/New_York") * 60 + minute(time, "America/New_York")42_q_clock_prev_second = _q_clock_prev_minute * 60 + second(time, "America/New_York")43_q_clock_prev_weekday = (dayofweek(time, "America/New_York") + 5) % 744_q_clock_prev_day = int(timestamp("UTC", year(time, "America/New_York"), month(time, "America/New_York"), dayofmonth(time, "America/New_York"), 0, 0) / 86400000)45// Current clock: this chart bar's scheduled close, known without a future bar.46// A native next event after a data gap can carry a later time.47_q_clock_minute = hour(time_close, "America/New_York") * 60 + minute(time_close, "America/New_York")48_q_clock_second = _q_clock_minute * 60 + second(time_close, "America/New_York")49_q_clock_weekday = (dayofweek(time_close, "America/New_York") + 5) % 750_q_clock_day = int(timestamp("UTC", year(time_close, "America/New_York"), month(time_close, "America/New_York"), dayofmonth(time_close, "America/New_York"), 0, 0) / 86400000)51_q_clock_seen = 152var _q_high55_buf = array.new_float(55, 0.0)53var int _q_high55_idx = 054var int _q_high55_count = 055var int _q_high55_skipped = 056var _q_high55_hiq = array.new_int(55, 0)57var int _q_high55_hih = 058var int _q_high55_hin = 059var _q_high55_loq = array.new_int(55, 0)60var int _q_high55_loh = 061var int _q_high55_lon = 062var _q_low55_buf = array.new_float(55, 0.0)63var int _q_low55_idx = 064var int _q_low55_count = 065var int _q_low55_skipped = 066var _q_low55_hiq = array.new_int(55, 0)67var int _q_low55_hih = 068var int _q_low55_hin = 069var _q_low55_loq = array.new_int(55, 0)70var int _q_low55_loh = 071var int _q_low55_lon = 072var _q_high20_buf = array.new_float(20, 0.0)73var int _q_high20_idx = 074var int _q_high20_count = 075var int _q_high20_skipped = 076var _q_high20_hiq = array.new_int(20, 0)77var int _q_high20_hih = 078var int _q_high20_hin = 079var _q_high20_loq = array.new_int(20, 0)80var int _q_high20_loh = 081var int _q_high20_lon = 082var _q_low20_buf = array.new_float(20, 0.0)83var int _q_low20_idx = 084var int _q_low20_count = 085var int _q_low20_skipped = 086var _q_low20_hiq = array.new_int(20, 0)87var int _q_low20_hih = 088var int _q_low20_hin = 089var _q_low20_loq = array.new_int(20, 0)90var int _q_low20_loh = 091var int _q_low20_lon = 092int day = na93float high_l = na94float low_l = na95float close_l = na96bool new_day = false97float true_range = na98int opened = na99 100bool _q_running = true101day := _q_int(_q_clock_prev_day)102if _q_clock_prev_minute >= session_minute103 day := _q_int(_q_iadd(day, 1))104high_l := _q_high * 1.0105low_l := _q_low * 1.0106close_l := _q_close * 1.0107new_day := day != day_key108if new_day109 if day_key != 0110 _q_sample1 = day_high111 if na(_q_sample1) or math.abs(_q_sample1) > 1.7976931348623157e308112 _q_high55_skipped := _q_high55_skipped + 1113 else114 if _q_high55_count >= 55 and _q_high55_hin > 0 and array.get(_q_high55_hiq, _q_high55_hih) == _q_high55_idx115 _q_high55_hih := (_q_high55_hih + 1) % 55116 _q_high55_hin := _q_high55_hin - 1117 if _q_high55_count >= 55 and _q_high55_lon > 0 and array.get(_q_high55_loq, _q_high55_loh) == _q_high55_idx118 _q_high55_loh := (_q_high55_loh + 1) % 55119 _q_high55_lon := _q_high55_lon - 1120 array.set(_q_high55_buf, _q_high55_idx, _q_sample1)121 while _q_high55_hin > 0122 _q_back1_hi = array.get(_q_high55_hiq, (_q_high55_hih + _q_high55_hin - 1) % 55)123 if array.get(_q_high55_buf, _q_back1_hi) <= _q_sample1124 _q_high55_hin := _q_high55_hin - 1125 else126 break127 array.set(_q_high55_hiq, (_q_high55_hih + _q_high55_hin) % 55, _q_high55_idx)128 _q_high55_hin := _q_high55_hin + 1129 while _q_high55_lon > 0130 _q_back1_lo = array.get(_q_high55_loq, (_q_high55_loh + _q_high55_lon - 1) % 55)131 if array.get(_q_high55_buf, _q_back1_lo) >= _q_sample1132 _q_high55_lon := _q_high55_lon - 1133 else134 break135 array.set(_q_high55_loq, (_q_high55_loh + _q_high55_lon) % 55, _q_high55_idx)136 _q_high55_lon := _q_high55_lon + 1137 _q_high55_idx := (_q_high55_idx + 1) % 55138 _q_high55_count := math.min(_q_high55_count + 1, 55)139 _q_sample2 = day_low140 if na(_q_sample2) or math.abs(_q_sample2) > 1.7976931348623157e308141 _q_low55_skipped := _q_low55_skipped + 1142 else143 if _q_low55_count >= 55 and _q_low55_hin > 0 and array.get(_q_low55_hiq, _q_low55_hih) == _q_low55_idx144 _q_low55_hih := (_q_low55_hih + 1) % 55145 _q_low55_hin := _q_low55_hin - 1146 if _q_low55_count >= 55 and _q_low55_lon > 0 and array.get(_q_low55_loq, _q_low55_loh) == _q_low55_idx147 _q_low55_loh := (_q_low55_loh + 1) % 55148 _q_low55_lon := _q_low55_lon - 1149 array.set(_q_low55_buf, _q_low55_idx, _q_sample2)150 while _q_low55_hin > 0151 _q_back2_hi = array.get(_q_low55_hiq, (_q_low55_hih + _q_low55_hin - 1) % 55)152 if array.get(_q_low55_buf, _q_back2_hi) <= _q_sample2153 _q_low55_hin := _q_low55_hin - 1154 else155 break156 array.set(_q_low55_hiq, (_q_low55_hih + _q_low55_hin) % 55, _q_low55_idx)157 _q_low55_hin := _q_low55_hin + 1158 while _q_low55_lon > 0159 _q_back2_lo = array.get(_q_low55_loq, (_q_low55_loh + _q_low55_lon - 1) % 55)160 if array.get(_q_low55_buf, _q_back2_lo) >= _q_sample2161 _q_low55_lon := _q_low55_lon - 1162 else163 break164 array.set(_q_low55_loq, (_q_low55_loh + _q_low55_lon) % 55, _q_low55_idx)165 _q_low55_lon := _q_low55_lon + 1166 _q_low55_idx := (_q_low55_idx + 1) % 55167 _q_low55_count := math.min(_q_low55_count + 1, 55)168 _q_sample3 = day_high169 if na(_q_sample3) or math.abs(_q_sample3) > 1.7976931348623157e308170 _q_high20_skipped := _q_high20_skipped + 1171 else172 if _q_high20_count >= 20 and _q_high20_hin > 0 and array.get(_q_high20_hiq, _q_high20_hih) == _q_high20_idx173 _q_high20_hih := (_q_high20_hih + 1) % 20174 _q_high20_hin := _q_high20_hin - 1175 if _q_high20_count >= 20 and _q_high20_lon > 0 and array.get(_q_high20_loq, _q_high20_loh) == _q_high20_idx176 _q_high20_loh := (_q_high20_loh + 1) % 20177 _q_high20_lon := _q_high20_lon - 1178 array.set(_q_high20_buf, _q_high20_idx, _q_sample3)179 while _q_high20_hin > 0180 _q_back3_hi = array.get(_q_high20_hiq, (_q_high20_hih + _q_high20_hin - 1) % 20)181 if array.get(_q_high20_buf, _q_back3_hi) <= _q_sample3182 _q_high20_hin := _q_high20_hin - 1183 else184 break185 array.set(_q_high20_hiq, (_q_high20_hih + _q_high20_hin) % 20, _q_high20_idx)186 _q_high20_hin := _q_high20_hin + 1187 while _q_high20_lon > 0188 _q_back3_lo = array.get(_q_high20_loq, (_q_high20_loh + _q_high20_lon - 1) % 20)189 if array.get(_q_high20_buf, _q_back3_lo) >= _q_sample3190 _q_high20_lon := _q_high20_lon - 1191 else192 break193 array.set(_q_high20_loq, (_q_high20_loh + _q_high20_lon) % 20, _q_high20_idx)194 _q_high20_lon := _q_high20_lon + 1195 _q_high20_idx := (_q_high20_idx + 1) % 20196 _q_high20_count := math.min(_q_high20_count + 1, 20)197 _q_sample4 = day_low198 if na(_q_sample4) or math.abs(_q_sample4) > 1.7976931348623157e308199 _q_low20_skipped := _q_low20_skipped + 1200 else201 if _q_low20_count >= 20 and _q_low20_hin > 0 and array.get(_q_low20_hiq, _q_low20_hih) == _q_low20_idx202 _q_low20_hih := (_q_low20_hih + 1) % 20203 _q_low20_hin := _q_low20_hin - 1204 if _q_low20_count >= 20 and _q_low20_lon > 0 and array.get(_q_low20_loq, _q_low20_loh) == _q_low20_idx205 _q_low20_loh := (_q_low20_loh + 1) % 20206 _q_low20_lon := _q_low20_lon - 1207 array.set(_q_low20_buf, _q_low20_idx, _q_sample4)208 while _q_low20_hin > 0209 _q_back4_hi = array.get(_q_low20_hiq, (_q_low20_hih + _q_low20_hin - 1) % 20)210 if array.get(_q_low20_buf, _q_back4_hi) <= _q_sample4211 _q_low20_hin := _q_low20_hin - 1212 else213 break214 array.set(_q_low20_hiq, (_q_low20_hih + _q_low20_hin) % 20, _q_low20_idx)215 _q_low20_hin := _q_low20_hin + 1216 while _q_low20_lon > 0217 _q_back4_lo = array.get(_q_low20_loq, (_q_low20_loh + _q_low20_lon - 1) % 20)218 if array.get(_q_low20_buf, _q_back4_lo) >= _q_sample4219 _q_low20_lon := _q_low20_lon - 1220 else221 break222 array.set(_q_low20_loq, (_q_low20_loh + _q_low20_lon) % 20, _q_low20_idx)223 _q_low20_lon := _q_low20_lon + 1224 _q_low20_idx := (_q_low20_idx + 1) % 20225 _q_low20_count := math.min(_q_low20_count + 1, 20)226 true_range := day_high - day_low227 if days > 0228 true_range := math.max(true_range, math.max(math.abs(day_high - last_close), math.abs(day_low - last_close)))229 if days < 20230 seed := seed + true_range231 atr := (seed * 1.0) / 20.0232 else233 atr := (atr * 19.0 + true_range * 1.0) / 20.0234 last_close := day_close235 days := _q_int(_q_iadd(days, 1))236 day_key := _q_int(day)237 day_high := high_l238 day_low := low_l239else240 day_high := math.max(day_high, high_l)241 day_low := math.min(day_low, low_l)242day_close := close_l243if days < entry_days244 _q_running := false245if _q_running246 opened := _q_int(0)247if _q_running248 if held > 0249 if close_l < (_q_low20_count == 0 ? 0.0 : array.get(_q_low20_buf, array.get(_q_low20_loq, _q_low20_loh))) or close_l < stop250 held := _q_int(0)251 else if held < 0252 if close_l > (_q_high20_count == 0 ? 0.0 : array.get(_q_high20_buf, array.get(_q_high20_hiq, _q_high20_hih))) or close_l > stop253 held := _q_int(0)254 else if close_l > (_q_high55_count == 0 ? 0.0 : array.get(_q_high55_buf, array.get(_q_high55_hiq, _q_high55_hih)))255 held := _q_int(1)256 stop := close_l - stop_atr * atr257 opened := _q_int(1)258 else if close_l < (_q_low55_count == 0 ? 0.0 : array.get(_q_low55_buf, array.get(_q_low55_loq, _q_low55_loh)))259 held := _q_int(-1)260 stop := close_l + stop_atr * atr261 opened := _q_int(1)262if _q_running263 if held == 0 and _q_int(strategy.position_size) > 0264 _q_qty5 = _q_int(math.abs(_q_int(strategy.position_size)))265 if _q_qty5 <= 0266 runtime.error("market order quantity must be a positive whole unit")267 _q_pos5 = _q_int(strategy.position_size)268 if _q_pos5 < 0 or _q_pos5 > 0 and _q_qty5 < _q_pos5269 runtime.error("this order would add to or partially reduce a position, so its trade ledger would not be faithful")270 strategy.order("q1", strategy.short, _q_qty5)271 else if held == 0 and _q_int(strategy.position_size) < 0272 _q_qty6 = _q_int(math.abs(_q_int(strategy.position_size)))273 if _q_qty6 <= 0274 runtime.error("market order quantity must be a positive whole unit")275 _q_pos6 = _q_int(strategy.position_size)276 if _q_pos6 > 0 or _q_pos6 < 0 and _q_qty6 < -_q_pos6277 runtime.error("this order would add to or partially reduce a position, so its trade ledger would not be faithful")278 strategy.order("q2", strategy.long, _q_qty6)279 else if held != 0 and _q_int(strategy.position_size) == 0 and (opened == 1 or new_day)280 if held > 0281 _q_qty7 = _q_int(clip)282 if _q_qty7 <= 0283 runtime.error("market order quantity must be a positive whole unit")284 _q_pos7 = _q_int(strategy.position_size)285 if _q_pos7 > 0 or _q_pos7 < 0 and _q_qty7 < -_q_pos7286 runtime.error("this order would add to or partially reduce a position, so its trade ledger would not be faithful")287 strategy.order("q3", strategy.long, _q_qty7)288 else289 _q_qty8 = _q_int(clip)290 if _q_qty8 <= 0291 runtime.error("market order quantity must be a positive whole unit")292 _q_pos8 = _q_int(strategy.position_size)293 if _q_pos8 < 0 or _q_pos8 > 0 and _q_qty8 < _q_pos8294 runtime.error("this order would add to or partially reduce a position, so its trade ledger would not be faithful")295 strategy.order("q4", strategy.short, _q_qty8) How it was tested
The published rules, never tuned on the data, with real costs on every fill. Every strategy here is tested the same way.
- Market data
- CME Globex daily bars, Jan 2, 2016 to Jul 31, 2026 ZN from March 2020
- Rules
- As published, with the published parameters, never tuned on this data
- Size
- 1 contract on every trade
- Contract
- The most traded one, rolled as the market moved
- Fees
- The broker's $0.50 a side, plus CME's and the NFA's fees when the window opens
- Slippage
- 1 tick a side on ES, 2 in NQ's regular session and 3 outside it, 1 on other markets, where it was not measured
- Stops and entries
- Tested on 1-second closes and filled at the next bar
- Contracts traded
- GC: 54 contracts, GCG6 to GCQ6ZN: 26 contracts, ZNM0 to ZNU66E: 43 contracts, 6EH6 to 6EU6
- Costs paid
- GC: $309 in fees, $1,250 in slippageZN: $118 in fees, $1,250 in slippage6E: $330 in fees, $831 in slippage
- Warm-up
- Each new contract first reads the last 130 trading days, moved onto its prices
- Flagged days
- GC: NoneZN: None6E: None Days the data holds too few bars for. Flagged, not removed.
Could you pass
a prop firm challenge
with the Turtles' System 2?
A backtest says whether a strategy makes money. Our AI runs its trades through the rules of every prop firm account, then tells you which one it would pass and at what size.
No card needed. Free credits every month.
- Most money: Blue Guardian Futures Direct 50K, $5,084
- Best payout odds: Funded Futures Family Straight to Funded 50K, 98%
- Fastest pass: Phidias Express to Live 50K, 11 days
$5,084
Median net across 2,000 simulated runs
- Evaluation
- None
- First payout
- 97%
- Fees paid
- $1,521
Median net across 2,000 simulated runs
Questions about the Turtles' System 2
What is the Turtles' System 2?
Buy a break of the last 55 days' high, sell short a break of their low, and exit on a 20-day break the other way or a stop 2 ATRs away. The rules follow Curtis Faith, on the rules Richard Dennis and William Eckhardt taught, Way of the Turtle (2007). Atlantive backtested it on CME futures from January 2016 to July 2026 after costs.
Is the Turtles' System 2 profitable on GC futures?
In this backtest it made money: on GC futures from January 2016 to July 2026 it made $89,101 after costs, over 75 trades, positive in 5 of 11 years; on ZN it lost $3,087 and on 6E it lost $1,855. Buying and holding one GC contract over the same window made $221,560. The average trade was $908, its 95% interval −$3,597 to $5,676. These are hypothetical results of a backtest, not of trading.
What are the Turtles' System 2's rules?
A one-second close above the last 55 days' high buys one contract, and a close below their low sells one short. A long is closed at the first close below the last 20 days' low, a short at the first close above their high. The stop is 2 N from the entry, where N is the 20-day average true range. A roll closes the position, and the new contract's first new day takes the same trade back. The rules follow Curtis Faith, on the rules Richard Dennis and William Eckhardt taught, Way of the Turtle (2007). Its parameters are the published ones, never tuned on the data it was tested on.
How was the Turtles' System 2 backtested?
On CME Globex data from January 2016 to July 2026, one contract, on daily bars, rolled to the most traded contract as the market moved, with the broker's and CME's fees and the slippage measured on each market. Stops and entries were tested on 1-second closes and filled at the next bar. Nothing was fitted to the data.
Can I run the Turtles' System 2 in TradingView?
Yes. The page carries its Pine Script, written by Atlantive's compiler from the same code the backtest ran, for TradingView's strategy tester. A chart's fills and fees can differ from the backtest's, so test it again on the symbol you trade.
Can I change the Turtles' System 2 and test it again?
Yes. Rerun it in Atlantive: a new chat opens with the strategy's rules in the box, for you to read, change and send. The agent writes the code, backtests it on the market and window you name, and reports the result with its costs.