Larry Connors and Cesar Alvarez
RSI(2): a system that wins three times in four, and why that is the problem
Most systems in this library lose on the majority of their trades. This one wins on around three quarters of them, which is precisely what makes it dangerous: a high hit rate feels like safety, and here it is the thing hiding where the risk actually is. Connors's published rules have no stop loss at all, and understanding why — and what that costs — is the whole of learning this strategy.
- Style
- Mean reversion
- Approach
- Mechanical
- Difficulty
- Intermediate
- Horizon
- Swing (days to weeks)
- Holding period
- 2-5 trading days
- Time needed
- 15 minutes a day
- Markets
- Index ETFs · Large caps
- Source
- Short Term Trading Strategies That Work — Larry Connors and Cesar Alvarez
The rule set
- Trend filter: only consider the instrument when price is above its 200-day moving average
- Entry: buy when the 2-period RSI falls below a low threshold — Connors used 10, and 5 for a more selective version
- Exit: sell on a close above the 5-day moving average, or when RSI(2) returns to neutral
- There is no profit target beyond that and no attempt to hold for a larger move
- Fix the risk per trade in advance, and never average down into a single name
What makes it distinctive
- A high hit rate with short holding periods and small average winners — the opposite payoff shape to every breakout system here
- The 200-day trend filter is not optional: without it the same rules buy every stage of a decline
- Entirely mechanical, which makes it unusually easy to backtest honestly and to review after the fact
When it works
Indices and large caps in an established long-term uptrend that pull back regularly — an environment where a sharp two-day drop really is noise rather than the start of something.
When it fails
At the beginning of a genuine trend reversal it buys repeatedly and loses each time, and one tail event can erase dozens of small wins. The original rules carry no hard stop, which is what makes that tail unbounded rather than merely painful.
How a decision moves through it
Input
Daily closing prices
Closes only. Every component of this system — RSI, both moving averages, the exit — is computed from the close, which is what makes it a fifteen-minute-a-day routine.
Measure
200-day moving average
The regime filter, and the single most important component. It decides whether the instrument is eligible at all.
Measure
RSI over two periods
A two-period lookback makes RSI extremely fast — it reaches single digits after two or three down days, which is exactly what it is meant to detect.
Decide
Oversold, and above the 200-day
Both together. The oversold reading on its own is the same signal that fires all the way down a bear market.
Act
Buy the close, exit on the bounce
Holding periods of two to five days. The exit is deliberately quick and takes a small gain rather than waiting for a larger one.
The payoff shape, and why it inverts everything
A trend-following system wins on perhaps 40% of trades, loses small, and occasionally wins enormously. RSI(2) is the mirror image: it wins on roughly 70–80% of trades, gains a little each time, and occasionally loses a great deal.
Both shapes can have the same expectancy. What differs is where the danger sits — and a high win rate puts it in the tail, where it is invisible until it arrives.
| Breakout system | RSI(2) | |
|---|---|---|
| Win rate | ~40% | ~75% |
| Average win | Large | Small |
| Average loss | Small, bounded by a stop | Small — until it is not |
| Worst case | One stop, defined in advance | Undefined without a stop |
| How it feels | Constantly wrong | Reliably right, then suddenly not |
The missing stop loss, which is deliberate and dangerous
Connors's published rules contain no stop loss, and his stated reasoning is testable: adding a stop to a mean-reversion system reduces its returns. That is true and it is easy to see why. The system buys weakness expecting a bounce; a stop below the entry sells the position at the precise moment the setup has become more extreme.
This is not an argument that Connors was wrong about the arithmetic. It is an argument about which risk you would rather carry, and it is a decision the strategy makes on your behalf unless you make it yourself. The sizing page covers the practical compromises.
The 200-day filter is the strategy, not a refinement
Buying oversold readings without a regime filter is the textbook way to lose money in a bear market: RSI(2) goes below 10 repeatedly all the way down, and each signal is technically correct and financially fatal.
The 200-day filter converts the question from 'is this oversold' to 'is this oversold within something that has been going up for a year'. In the second case, a sharp two-day drop is plausibly noise. In the first, it may be the trend.
Every mean-reversion system in this library needs an equivalent of this filter, and the ones that fail catastrophically are the ones that dropped it because it was rejecting signals.
Five ways into this system
- Two indicators, one filter, and an exit that takes the small gainA short rule set with one component that behaves very differently from the version most people know.7 min read
- Sizing a system with no stop loss, which is the only real decision hereThe strategy declines to define its own worst case. That leaves you three options, and choosing none of them is itself a choice.7 min read
- Where a two-day drop is noise rather than newsThe strategy needs an instrument where a sharp short-term fall is usually noise, and gains small enough that transaction costs are a first-order concern.5 min read
- The turn, the tail, and a published edge that has thinnedThe failure is not gradual. Seventy small wins accumulate quietly, and the loss that matters arrives all at once at a trend reversal.7 min read
- RSI(2) for beginners: why winning most of the time is not the same as making moneyThe most useful lesson available to a beginner is embedded in this strategy: a high win rate tells you nothing on its own.6 min read
The ideas behind it
This system assumes you already know these. Each one is explained from scratch in Investing 101.
Compare with
- Bollinger Band Squeeze BreakoutWait until the bands compress to the narrowest they have been in months, then trade whichever direction price breaks out of the compression.
- 200-day MA Macro TrendHold the index while it closes above its 200-day average, and sit in cash or short-term bonds while it closes below.
These are documented methods described for study. Nothing here is investment advice, a recommendation, or a claim about future returns — every system on this page has losing periods, and the pages say where.
Reading about a system is not having one.
Plutux is where you write your own rules down, test them against real data, and keep the record your memory would otherwise rewrite. Join the waitlist for early access.