Connors RSI(2) Mean Reversion with a Long-Term Trend Filter
Summary
This implementation presents a basic mean-reversion approach using a two-period RSI and a 200-period simple moving average as a directional filter. It seeks long entries when price is above the long-term average and RSI is at or below 10; short entries require price below that average and RSI at or above 90. Positions are closed when price crosses back through a five-period simple moving average, creating a short-term exit rule for both directions.
The code sizes each position using the available balance and current price, with the fee rate included in the quantity calculation. It does not specify protective stops, profit targets, a maximum holding period, or additional position-risk limits. The document identifies the method as originating with Larry Connors and provides an implementation, but reports no backtest or live-trading evidence. Results would depend on the traded market, timeframe, fees, shorting conditions, and execution assumptions.
Key ideas
- The strategy uses RSI with a two-period lookback to identify extreme short-term moves.
- A 200-period moving average determines whether long or short setups are permitted.
- Long entries require oversold RSI above the trend filter, while short entries require overbought RSI below it.
- Price crossing a five-period average triggers position liquidation.
- The code gives no test results or explicit stop-loss and position-risk limits.
Tags
Full text
# RSI2
# RSI2
Connors' RSI2 Strategy
2020-06-27 original implementation - FengkieJ (fengkiejunis@gmail.com)
This strategy is a very basic mean reversion strategy, mainly using RSI(2) overbought and oversold levels to determine entry signals.
Originally developed by Larry Connors.
Reference: - RSI(2) [ChartSchool]. Retrieved 27 June 2020, from https://school.stockcharts.com/doku.php?id=trading_strategies:rsi2
## Source (MIT)
```python
"""
Connors' RSI2 Strategy
2020-06-27 original implementation - FengkieJ (fengkiejunis@gmail.com)
This strategy is a very basic mean reversion strategy, mainly using RSI(2) overbought and oversold levels to determine entry signals.
Originally developed by Larry Connors.
Reference: - RSI(2) [ChartSchool]. Retrieved 27 June 2020, from https://school.stockcharts.com/doku.php?id=trading_strategies:rsi2
"""
from jesse.strategies import Strategy
import jesse.indicators as ta
from jesse import utils
class RSI2(Strategy):
def __init__(self):
super().__init__()
self.vars["fast_sma_period"] = 5
self.vars["slow_sma_period"] = 200
self.vars["rsi_period"] = 2
self.vars["rsi_ob_threshold"] = 90
self.vars["rsi_os_threshold"] = 10
@property
def fast_sma(self):
return ta.sma(self.candles, self.vars["fast_sma_period"])
@property
def slow_sma(self):
return ta.sma(self.candles, self.vars["slow_sma_period"])
@property
def rsi(self):
return ta.rsi(self.candles, self.vars["rsi_period"])
def should_long(self) -> bool:
# Enter long if current price is above sma(200) and RSI(2) is below oversold threshold
return self.price > self.slow_sma and self.rsi <= self.vars["rsi_os_threshold"]
def should_short(self) -> bool:
# Enter long if current price is below sma(200) and RSI(2) is above oversold threshold
return self.price < self.slow_sma and self.rsi >= self.vars["rsi_ob_threshold"]
def should_cancel_entry(self) -> bool:
return False
def go_long(self):
# Open long position and use entire balance to buy
qty = utils.size_to_qty(self.balance, self.price, fee_rate=self.fee_rate)
self.buy = qty, self.price
def go_short(self):
# Open short position and use entire balance to sell
qty = utils.size_to_qty(self.balance, self.price, fee_rate=self.fee_rate)
self.sell = qty, self.price
def update_position(self):
# Exit long position if price is above sma(5)
if self.is_long and self.price > self.fast_sma:
self.liquidate()
# Exit short position if price is below sma(5)
if self.is_short and self.price < self.fast_sma:
self.liquidate()
```Shown in full with attribution under the source's licence. Licence: MIT
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.