Skip to content
QuantReplay

Connors' RSI(2) on S&P 500 futures

Micro E-mini S&P 500 (MES) · 5-minute bars · 2016 to 2026 · Mean reversion · Popularized by Larry Connors

Too few trades to judge: 35 out of sample.

Profit factor, out of sample
3.56
Average trade, out of sample
+0.18R
Win rate, out of sample
71%
Trades, 10 years
93
Max drawdown
$1,554
Pass odds, median plan
0.0%

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.

SharePost on XLinkedInReddit

Abstract

We backtested the RSI(2) mean reversion rules on Micro E-mini S&P 500 (MES) 5-minute bars from 2016-01-11 to 2026-10-30: 93 trades, with commission and a tick of slippage on every market and stop fill. In sample (to 2023-08-03) it made +0.10R per trade with a profit factor of 1.40; out of sample (from 2023-08-04) +0.18R per trade and 3.56 over 35 trades. 16 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 0% of the time across 36 tracked plans.

1Hypothesis

In a long-term uptrend (above the 200-day average), buy when the 2-day RSI drops below 10 and sell when price closes above its 5-day average.

If short, sharp pullbacks inside a long-term uptrend tend to reverse within days, buying them should earn more than holding costs and the occasional crash.

2Data

Instrument
Micro E-mini S&P 500 (MES), traded on the ES price series (ES.c.0: front month, rolled on the calendar). CME Globex via Databento.
Period
2016-01-11 to 2026-10-30: 2,746 sessions, 3,770,901 one-minute bars, built into 210,816 5-minute bars of the regular session (09:30 to 16:00 ET).
Costs
$0.75 commission a contract each way and 1 tick ($1.25) 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. Daily series: each regular session's price at 15:55 ET, high and low.
  2. Entry: at 15:55 ET, if that price is above its 200-day simple average and the 2-day RSI (Wilder's) is below rsi_entry (10), buy at the next bar's open.
  3. Exit: at 15:55 ET on the first day the price closes above its exit_sma (5) day average.
  4. Stop: 3 times the 10-day average true range below the entry. Connors' rules have no stop; this one is wide, and it gives each trade a defined risk for sizing and for R.
  5. 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
"""RSI(2) mean reversion on daily closes (Connors), traded on Micro E-mini S&P 500 (MES), long only. Each session's 15:55 ET price stands in for its daily close; positions are held overnight. Rules:  1. Daily series: each regular session's price at 15:55 ET, high and low.  2. Entry: at 15:55 ET, if that price is above its 200-day simple average and the 2-day RSI (Wilder's)     is below `rsi_entry` (10), buy at the next bar's open.  3. Exit: at 15:55 ET on the first day the price closes above its `exit_sma` (5) day average.  4. Stop: 3 times the 10-day average true range below the entry. Connors' rules have no stop; this one     is wide, and it gives each trade a defined risk for sizing and for R.  5. 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):    """Buy a 2-day RSI washout in a long-term uptrend, sell the bounce."""     symbol = "MES"    timeframe = "5m"    session = "rth"    params = {"rsi_entry": 10, "exit_sma": 5, "trend_sma": 200, "risk_usd": 250, "max_contracts": 50}    max_position = 50    overnight = True     def setup(self):        self.day = None        self.closes, self.trs = [], []        self.gain = self.loss = None        self.dh = self.dl = 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:            self.day, self.dh, self.dl = bar.date, bar.high, bar.low        else:            self.dh, self.dl = max(self.dh, bar.high), min(self.dl, bar.low)        if bar.dt.hour * 60 + bar.dt.minute != 950:   # the bar that closes at 15:55            return        c = bar.close        prev = self.closes[-1] if self.closes else None        if prev is not None:            self.trs.append(max(self.dh, prev) - min(self.dl, prev))            ch = c - prev            g, l = max(ch, 0.0), max(-ch, 0.0)            if self.gain is None:                self.gain, self.loss = g, l            else:                self.gain, self.loss = (self.gain + g) / 2, (self.loss + l) / 2        self.closes.append(c)        self.closes, self.trs = self.closes[-400:], self.trs[-400:]        if self.position.is_long:            n = self.p.exit_sma            if len(self.closes) >= n and c > sum(self.closes[-n:]) / n:                self.close_position("close above the short average")            return        n = self.p.trend_sma        if len(self.closes) < n or len(self.trs) < 10 or self.gain is None or self.orders:            return        rsi = 100.0 if self.loss == 0 else 100 - 100 / (1 + self.gain / self.loss)        if c > sum(self.closes[-n:]) / n and rsi < self.p.rsi_entry:            atr = sum(self.trs[-10:]) / 10            self.buy(self._size(3 * atr), stop_price=c - 3 * atr, tag=f"RSI2 long | f={rsi:.2f}")

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
Trades583593
Net P&L+$2,469+$6,694+$9,163
Profit factor1.403.562.05
Average trade+0.10R+0.18R+0.13R
95% interval of the average−0.05 to 0.250.03 to 0.340.02 to 0.24
t-statistic of trade R1.322.472.42
Win rate76%71%74%
Average win / average loss0.451.420.71
Max drawdown$1,554$824$1,554
Sharpe ratio (daily, annualized)2.038.044.40
Longest losing streak222
Long / short profit factor1.40 / n/a3.56 / n/a2.05 / n/a
Average holding time5593 min7367 min6260 min
Commission paid$87$53$140
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 (58 trades)Out of sample (35)Averages

Averages: +0.10R in sample, +0.18R 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 2-day RSI at entry; the dashed line is the OLS fit (r = −0.02, p = 0.83, n = 93; Spearman ρ = −0.17). Out of sample alone: r = −0.31, p = 0.07. 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. 8 were profitable in sample and 16 out of sample.

15
4.00
2.97
3.24
2.93
10
4.00
3.00
2.49
2.47
5
4.00
3.56
2.70
2.20
3
4.00
3.00
2.41
2.11
5
10
15
20

RSI entry level across, exit average (days) up. Solid outline: the defaults we tested; dashed: the in-sample best.

Hover a cell for both sides of the split.

RSI entry level 5.00 to 20.0 across · Exit average (days) 3.00 to 15.0 in depth · Profit factor, out of sample 2.11 to 4.00 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 RSI entry level and exit average (days), as a map and as a surface of the same grid. Values above 4 are drawn at 4.

What would change it

The defaults were already the in-sample best of the grid, so there is no tweak to test: profit factor 1.40 in sample and 3.56 out of sample.

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 2.5% [1.9%, 3.3%]; 79% of years ended up; median ending P&L +$830, 5th to 95th percentile −$978 to +$2,562.

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 50K0.1%1%99%49
Alpha Futures Advanced 50K Evaluation0.0%0.3%100%n/a
Alpha Futures Standard 50K Evaluation0.0%0.1%100%n/a
Alpha Futures Zero 50K Evaluation0.0%0.1%100%n/a
Apex Trader Funding 50K EOD Evaluation0.0%0.1%100%n/a
Apex Trader Funding 50K Intraday Trailing Drawdown Evaluation0.0%0.1%100%n/a
AquaFutures Beginner 50K Evaluation0.0%0%100%n/a
Bulenox 50K Qualification Account (Legacy, Option 1, trailing drawdown)0.0%0%100%n/a
Bulenox 50K Qualification Account (Legacy, Option 2, EOD drawdown)0.0%0%100%n/a
Earn2Trade Gauntlet Mini $50K0.0%0.1%100%n/a
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

  • The daily close is approximated by the 15:55 ET price.
  • Long only; the added 3-ATR stop changes the original rules slightly.
  • Few trades a year.
  • A backtest on the past, with fills simulated from one-minute bars; live fills, queues and outages differ.

8Conclusion

Too few trades to judge: 35 out of sample. The defaults were already the in-sample best of the grid, so there is no tweak to test: profit factor 1.40 in sample and 3.56 out of sample. The result is too close to break-even to tell apart from noise after costs.

9References

  1. Connors, L., and Alvarez, C. (2009). Short Term Trading Strategies That Work. TradingMarkets Publishing Group.
  2. Wilder, J. W. (1978). New Concepts in Technical Trading Systems. Trend Research.

Every number on this page comes from the MES 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.