Skip to content

Fading sweeps of yesterday's high and low

Micro E-mini Nasdaq-100 (MNQ) · 5-minute bars · 2016 to 2026 · ICT and liquidity · A common liquidity-sweep setup

No edge after costs: profit factor 0.84 out of sample.

Profit factor, out of sample
0.84
Average trade, out of sample
−0.11R
Win rate, out of sample
31%
Trades, 10 years
602
Max drawdown
$16,267
Pass odds, median plan
1.3%

Hypothetical results. These results are based on simulated or hypothetical performance results that have certain inherent limitations. Unlike the results shown in an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under- or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown.

Abstract

We backtested the previous-day high/low sweep fade rules on Micro E-mini Nasdaq-100 (MNQ) 5-minute bars from 2016-01-11 to 2026-10-30: 602 trades, with commission and a tick of slippage on every market and stop fill. In sample (to 2023-08-03) it made −0.06R per trade with a profit factor of 0.92; out of sample (from 2023-08-04) −0.11R per trade and 0.84 over 173 trades. 5 of the 16 parameter settings we tried were profitable out of sample. Sized at $250 of risk a trade, simulated 50K evaluations passed 0% to 3% of the time across 36 tracked plans.

1Hypothesis

When price pokes through yesterday's high or low and closes back inside within three bars, fade the poke, stop beyond its extreme, target twice the risk.

If quick pokes through yesterday's extremes are stop runs that fail, fading the reclaim should reach 2R often enough to pay for costs.

2Data

Instrument
Micro E-mini Nasdaq-100 (MNQ), traded on the NQ price series (NQ.c.0: front month, rolled on the calendar). CME Globex via Databento.
Period
2016-01-11 to 2026-10-30: 2,745 sessions, 1,056,797 one-minute bars, built into 210,814 5-minute bars of the regular session (09:30 to 16:00 ET).
Costs
$0.75 commission a contract each way and 1 tick ($0.50) of slippage on every market and stop fill. Limit orders fill when price touches them; market orders fill at the next bar's open and stops at their price, each with one tick of slippage; when a bar touches both the stop and the target, the stop counts.
Position size
About $250 of risk a trade: micro contracts, rounded down, at least one. R is a trade's net P&L divided by its risk at entry.
Out of sample
The last 30% of weekdays: 2023-08-04 on (846 days), after 1,974 in-sample days. The rules were fixed before either side was run.

3Method

Our reading of the publicly described rules; the creator may trade it differently. The rules, exactly as the code runs them:

  1. Yesterday's high and low are its regular-session (09:30-16:00 ET) extremes. The day must open inside them.
  2. After a bar trades above yesterday's high, sell the first close back below it if that close comes within reclaim_bars (3) bars, counting the bar that went above. Yesterday's low is the mirror image (buy). Each level gets one try a day, from 09:30 to 12:00 ET.
  3. Stop: one tick beyond the extreme of the sweep (the highest high since price went above).
  4. Target: target_r (2) times the risk. Trades risking under 0.05% or over 0.5% of price are skipped.
  5. Exit: the stop, the target, or flat at 15:55 ET. One trade a day.
  6. Size: about risk_usd dollars of risk a trade (micro contracts, rounded down, at least 1 and at most max_contracts).
bot.py, the file the engine ran
"""Previous-day high and low sweep fade on Micro E-mini Nasdaq-100 (MNQ) 5-minute bars, regular session. Rules:  1. Yesterday's high and low are its regular-session (09:30-16:00 ET) extremes. The day must open     inside them.  2. After a bar trades above yesterday's high, sell the first close back below it if that close comes     within `reclaim_bars` (3) bars, counting the bar that went above. Yesterday's low is the mirror     image (buy). Each level gets one try a day, from 09:30 to 12:00 ET.  3. Stop: one tick beyond the extreme of the sweep (the highest high since price went above).  4. Target: `target_r` (2) times the risk. Trades risking under 0.05% or over 0.5% of price are skipped.  5. Exit: the stop, the target, or flat at 15:55 ET. One trade a day.  6. Size: about `risk_usd` dollars of risk a trade (micro contracts, rounded down, at least 1 and at     most `max_contracts`)."""import qy  class Bot(qy.Strategy):    """Fade a failed poke through yesterday's high or low."""     symbol = "MNQ"    timeframe = "5m"    session = "rth"    params = {"reclaim_bars": 3, "target_r": 2.0, "risk_usd": 250, "max_contracts": 50}    max_position = 50    entry_window = ("09:35", "12:00")    flatten_time = "15:55"    overnight = False     def setup(self):        self.day = None        self.h = self.l = None        self.yh = self.yl = None     def _size(self, risk_points):        per_contract = risk_points * self.point_value        n = int(self.p.risk_usd // per_contract) if per_contract > 0 else 1        return max(1, min(self.p.max_contracts, n))     def on_bar(self, bar):        if bar.date != self.day:            if self.h is not None:                self.yh, self.yl = self.h, self.l            self.day, self.h, self.l = bar.date, bar.high, bar.low            self.ok = self.yh is not None and self.yl < bar.open < self.yh            self.up = self.dn = None          # [bars since the poke, extreme]            self.tried_up = self.tried_dn = False        else:            self.h, self.l = max(self.h, bar.high), min(self.l, bar.low)        if not self.ok or self.trades_today >= 1 or not self.position.is_flat or self.orders:            return        t = bar.dt.hour * 60 + bar.dt.minute + 5        if t >= 720:            return        tick = self.tick_size        if not self.tried_up:            if self.up is None and bar.high > self.yh:                self.up = [0, bar.high]            if self.up is not None:                self.up = [self.up[0] + 1, max(self.up[1], bar.high)]                if bar.close < self.yh:                    self.tried_up = True                    self._fade(-1, bar.close, self.up[1] + tick, self.up[0] - 1)                    return                if self.up[0] >= self.p.reclaim_bars:                    self.tried_up = True        if not self.tried_dn:            if self.dn is None and bar.low < self.yl:                self.dn = [0, bar.low]            if self.dn is not None:                self.dn = [self.dn[0] + 1, min(self.dn[1], bar.low)]                if bar.close > self.yl:                    self.tried_dn = True                    self._fade(1, bar.close, self.dn[1] - tick, self.dn[0] - 1)                    return                if self.dn[0] >= self.p.reclaim_bars:                    self.tried_dn = True     def _fade(self, d, close, stop, bars):        risk = abs(close - stop)        if not 0.0005 * close <= risk <= 0.005 * close:            return        f = risk / close * 100        if d > 0:            self.buy(self._size(risk), stop_price=stop, target_r=self.p.target_r, tag=f"PDL sweep long | f={f:.4f}")        else:            self.sell(self._size(risk), stop_price=stop, target_r=self.p.target_r, tag=f"PDH sweep short | f={f:.4f}")

4Results

Table 1. In sample (2016-01-11 to 2023-08-03) against out of sample (2023-08-04 to 2026-10-30), after costs.
StatisticIn sampleOut of sampleAll
Trades429173602
Net P&L−$5,573−$4,229−$9,802
Profit factor0.920.840.90
Average trade−0.06R−0.11R−0.07R
95% interval of the average−0.19 to 0.08−0.32 to 0.09−0.19 to 0.04
t-statistic of trade R−0.82−1.10−1.28
Win rate35%31%34%
Average win / average loss1.721.851.76
Max drawdown$10,544$6,876$16,267
Sharpe ratio (daily, annualized)−0.60−1.29−0.78
Longest losing streak101111
Long / short profit factor0.98 / 0.880.83 / 0.840.94 / 0.87
Average holding time34 min44 min37 min
Commission paid$8,048$1,098$9,146
Cumulative P&LDrawdown from the highOut of sample from 2023-08-04

Drag across the chart to zoom into a period, double-click to reset.

Figure 1. Cumulative P&L after costs and its drawdown from the running high, 2016-01-11 to 2026-10-30, at about $250 of risk a trade. The shaded part is out of sample.
In sample (429 trades)Out of sample (173)Averages

Averages: −0.06R in sample, −0.11R out of sample.

Figure 2. The distribution of trade results in R, as a share of each side's trades. Bars are in sample; the line is out of sample.
In sampleOut of sampleQuintile mean, 95% intervalOLS fitDensity contours

Drag a box to zoom, double-click to reset.

Figure 3. Trade R against the stop's distance, log x-axis; the dashed line is the OLS fit (r = −0.03, p = 0.44, n = 602; Spearman ρ = 0.36). Out of sample alone: r = −0.05, p = 0.54. Squares are the means of five equal-count bins, with 95% intervals.

5Robustness

We reran the strategy on a grid of 16 settings of its two main parameters, each on the full history with the same split. 0 were profitable in sample and 5 out of sample.

3
1.02
0.94
0.88
0.83
2
0.90
0.85
0.84
0.80
1.5
0.86
0.80
0.81
0.76
1
1.26
1.05
1.11
1.06
1
2
3
5

Bars to reclaim across, target (R) up. Solid outline: the defaults we tested; dashed: the in-sample best.

Hover a cell for both sides of the split.

Bars to reclaim 1.00 to 5.00 across · Target (R) 1.00 to 3.00 in depth · Profit factor, out of sample 0.76 to 1.26 as height and shade. Drag or use the arrow keys to turn it.

Drag to rotate; ctrl or ⌘ and scroll to zoom.

Figure 4. Profit factor across bars to reclaim and target (R), as a map and as a surface of the same grid. Values above 4 are drawn at 4.

What would change it

Nothing on the grid helps. Its in-sample best, bars to reclaim 2 and target (R) 3, had profit factor 0.97 in sample but 0.94 out of sample, against 0.84 for the defaults.

In-sample yearOut-of-sample year95% interval
Figure 5. Average trade by calendar year with its 95% interval: a year-by-year walk through the history. Lighter bars are in-sample years.
Figure 6. Average trade in three volatility regimes: terciles of the 20-session average of the daily range (% of the open), up to the day before the trade (regular session). The cut-offs use the whole period, so this describes, it doesn't forecast.
Figure 7. Average trade by weekday and by the hour of entry (New York time), with 95% intervals.
Median25th to 75th percentile5th and 95thRuin: down $2,000
Figure 8. 2,000 simulated years of 252 sessions, drawn session by session from the whole backtest (days without trades included), with 48 paths drawn; red paths lost $2,000 at some point. Risk of ruin 53.5% [51.4%, 55.7%]; 36% of years ended up; median ending P&L −$945, 5th to 95th percentile −$4,793 to +$2,919.

6Prop-firm implications

The backtest's sessions run through the rules of 36 tracked 50K evaluations: 2,000 runs of up to 60 sessions each, with the drawdown and daily limits checked against each trade's intraday low. Intervals are Wilson 95% intervals of the simulation. A simulation of the past, not the chance that anyone passes.

PlanPass odds, 95% intervalPassFailOut of timeMedian sessions
FundedNext Futures Flex Challenge 50K2.6%34%63%40
Alpha Futures Standard 50K Evaluation1.3%15%84%46
Alpha Futures Zero 50K Evaluation1.3%15%84%46
Apex Trader Funding 50K EOD Evaluation1.3%15%84%46
Apex Trader Funding 50K Intraday Trailing Drawdown Evaluation1.3%21%78%46
AquaFutures Beginner 50K Evaluation1.3%5%94%46
Bulenox 50K Qualification Account (Legacy, Option 1, trailing drawdown)1.3%8%91%46
Bulenox 50K Qualification Account (Legacy, Option 2, EOD drawdown)1.3%5%94%46
Earn2Trade Gauntlet Mini $50K1.3%15%84%46
Earn2Trade Trader Career Path TCP501.3%15%84%46
Scale 0 to 10%.
Figure 9. Simulated pass odds by plan, best first, with each plan's fail and out-of-time shares and the median sessions to a pass.

7Limitations

  • One try per level a day, mornings only.
  • One market (MNQ).
  • A backtest on the past, with fills simulated from one-minute bars; live fills, queues and outages differ.

8Conclusion

No edge after costs: profit factor 0.84 out of sample. Nothing on the grid helps. Its in-sample best, bars to reclaim 2 and target (R) 3, had profit factor 0.97 in sample but 0.94 out of sample, against 0.84 for the defaults. On this reading of the rules, with these costs, the historical record gives no reason to expect a profit.

Every number on this page comes from the MNQ bot above, run by our backtest engine on our own data. We publish what doesn't work as well as what does, and we don't tune rules to look good.

Hypothetical results. These results are based on simulated or hypothetical performance results that have certain inherent limitations. Unlike the results shown in an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under- or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown.