Most traders still run the R3 playbook like it’s 2009. But with modern futures, volatility filters, and smarter entry tweaks—you can triple your trades and cut your drawdown.
TL;DR Summary
📉 Classic Larry Connors R3 still works as a mean reversion strategy, but it's outdated for modern markets.
📈 Index futures like ES, NQ, and RTY outperform ETFs due to leverage and speed.
🔄 Relaxing RSI rules increased trade count 4x without hurting performance.
🔥 Volatility filters outperformed the 200-day MA , and added diversification.
🧠 Combining filtered variants produced over 4300 trades with a 70.7% win rate.
What Is the R3 Strategy by Larry Connors?
The R3 strategy was published in 2009 by Larry Connors as a simple mean reversion system tailored for ETFs. It uses the RSI 2-period to catch pullbacks in strong uptrends.
Original Rules:
- The instrument is above its 200-day moving average.
- RSI(2) is below 10.
- RSI(2) has declined 3 days in a row.
- RSI(2) was below 60 three bars ago.
- Entry: Buy on the close when all conditions are true.
- Exit: Sell when RSI(2) closes above 70 or if Close is lower than SMA50
It was easy to track and yielded excellent results back then.
SPY Equity curve and annual performance stats for the original Larry Connors R3 strategy on the SP500 ETF
🎥 Prefer to watch instead of read?
I broke down everything in this article , from the classic R3 rules to the modern futures strategy , in this detailed YouTube video: 👉 Watch the R3 Strategy Rebuilt for Futures
Why Most Traders Still Misuse R3 (And Lose Edge)
The strategy works. But the markets have evolved. Many traders still apply the original R3 on ETFs, expecting the same results as 2009.
Here’s the issue:
- ETFs are slow. Low leverage, slippage, and tight ranges limit the mean reversion payoff.
- RSI(2) below 10 happens rarely, especially with strong directional filters.
- Too few trades make your portfolio brittle.
So I applied the R3 on US index futures, S&P 500, Nasdaq, Dow, Russell, and MidCap, and rebuilt it with smarter filters.
Modernizing R3 for Futures
Futures offer leverage, low fees, and high liquidity. Perfect for mean reversion. Here’s how R3 looks on the S&P 500 futures (ES):
Initial Results:
- 110 trades since 2006
- 58% win rate
- $287 average trade
- Return/Drawdown: 2.9
Problem?
Only 110 trades in 32 years. That’s ~6 per year. Great stats, but not enough opportunities.
The Trade-Off: More Trades vs. Tighter Filters
To get more trades, I relaxed the RSI conditions and ran an optimization with 1920 combinations using the following variables:
- RSI(2) < {Level}
- At least {X} of the last {Y} RSI bars must be declining
- Exit after {Bars} bars OR if RSI(2) > {Level}
We find the following:
· All of the combinations are profitable which says a lot.
· More than 1000 combinations produced better Return/DD ratio.
· More than 1800 of the combinations produced better average trade
· Top 50 strategies produced +5 Return/DD ratio.
✅ One of the best versions (picked for higher number of trades):
- 432 trades
- 73% win rate
- $563 average trade
- 6.4 Return/DD
ES Equity curve and annual performance stats for relaxed R3 rules on ES futures (SP500), showing a major increase in trades and profit
Choosing the Right Filter: Directional vs. Volatility
Time to bring back the filter, I tested:
- Directional Filter: using my proprietary Market Regime Directional filter
- Volatility Filter: using my proprietary Market Regime Volatility filter
- Volume Filter: using a simple volume oscillator
Many variations worked, but not equally.
Table comparing 4 R3 strategy variants with metrics like win rate, average trade, Return/DD. Highlights benefits of relaxed rules and filters
Building a Robust Strategy Portfolio With Filter Variants
Here’s the magic: same strategy logic with three variant strategy filters, on 5 instruments, US Indexes.
When combined, the three variants produced:
- Over 4300 trades
- 70.7% win rate
- $470+ average trade
- Smoothed equity curve with low correlation between most strategies.
Correlation matrix for R3 filters across multiple instruments, showing low overlap between R3 variants. Below is a combined portfolio equity curve
Portfolio equity curve of Larry Connors R3 strategy with different filters applied to 5 US Indexes
📌 One reason these R3 variants worked well is that they were tested across instruments, filters, and timeframes, not just overfitted to past trades.
If you want to understand why most traders skip this step (and what it costs them), check out this breakdown:
👉 Why Most Traders Fail: The Missing Piece Called Robustness Testing
“Even with the same instrument and same strategy, just changing the filter gave us uncorrelated results.” – Ali Casey
Which Volatility Filter Works Best?
I tested several filters, but one stood out:
- Based on 20-day true range percentile
- Captured contractions before price reversals
📖 Related post: Read our full guide on volatility filters that boost strategy performance.
Combining Multiple Futures Markets for Broader Edge
This R3 variant works across:
- S&P 500 (ES)
- Nasdaq (NQ)
- Dow (YM)
- Russell 2000 (RTY)
- MidCap 400
Each instrument showed different characteristics. For instance:
- ES had higher average trade with directional filter
- NQ performed better under volatility filter
- Low correlation between same-strategy filters
Perfect for micro futures traders too!
Key Takeaway: Same Strategy, Different Filter = New Edge
A simple shift in your filter logic can create:
- More trades
- Uncorrelated results
- Better portfolios
“You don’t need a new indicator , you need a new lens.” – Ali Casey, StatOasis
Get the fully-tested R3 variants, with relaxed entry and custom filters
→ Click here to download the strategy pack
FAQs
Q: What’s the original Larry Connors R3 setup?
A mechanical RSI2-based pullback strategy for ETFs using 200-day MA and RSI triggers.
Q: Is Larry Connors R3 strategy still profitable in 2025?
A: Yes, with updated entry logic and better filters, it can outperform the original ETF version.
Q: Why trade index futures instead of ETFs?
More leverage, tighter spreads, and consistent volatility make futures ideal for fast reversion.
Q: Which RSI levels work best now?
RSI(2) < 40 with bar decline logic gave significantly more trades without compromising win rate.
Q: What’s a voltility filter?
A rule based on range, true range, or price contraction. It adapts better to modern markets.
Q: How do I build a portfolio from this?
Use the same entry logic across different instruments and filters. Combine results, track correlation, and allocate capital accordingly.
Q: Can I use Larry Connors R3 strategy on micro futures like MES or MNQ?
Absolutely. The logic is identical, and the smaller contract size makes it ideal for small accounts.
Q: What platform can I use to automate this?
Any backtesting platform that supports RSI and bar logic will work. I use MultiCharts, but you can adapt it in TradeStation, StrategyQuant X, NinjaTrader, or TradingView.
Q: Is this strategy suitable for live trading?
Yes, but only after paper trading. Try at least 3 months of demo testing to understand behavior and slippage.
Q: How do I know when a strategy is no longer working?
Track rolling metrics like win rate, profit factor, and max drawdown. If they deviate significantly from historical norms for 30+ trades, it may be time to review.
Conclusion: Stop Using R3 Like It’s 2009
Larry Connors gave us a timeless framework. But frameworks evolve.
With just a few tweaks, relaxed entries, index futures, volatility filters, you can create a great portfolio.
