Daily RSI Below 20 on US Stocks｜Comparing 3 Exits
A backtest of 100 US stocks on daily charts (5–56 years per stock) in ENTRIQ, entering after RSI(14) drops below 20. Compares 3 exits with RSI below 30.
Introduction: what RSI below 20 means
An RSI reading below 20 points to an even deeper oversold condition than a drop below 30. It is natural to think, "If it sold off more deeply, the bounce should be bigger too," but how did things actually play out in historical data?
These are the results of backtesting 100 stocks in ENTRIQ (5 to 56 years of data depending on the stock) (Dataset v2.0). Looking at cases where RSI(14) dropped below 20, we lay out the results for 3 patterns with different exits and compare them side by side with the RSI-below-30 figures.
The detailed backtest of RSI below 30 is covered in Buying When RSI Drops Below 30｜Backtest of 3 Exits and Risks, so to avoid repetition, this article focuses on what is specific to RSI 20 and on the comparison.
Test conditions
- Universe: 100 US stocks selected from the current S&P 500 constituents, taking the largest by market cap with up to 15 stocks per sector
- Period: The full history available for each stock (5 to 56 years depending on the stock); only stocks with at least 5 years of data
- Timeframe: Daily (split-adjusted)
- RSI setting: Period 14
- Entry: Buy at the open on the next trading day after RSI(14) crosses below 20
- New signals in the same stock while a position is open are skipped (only one position at a time)
Exit patterns
| Pattern | Exit definition |
|---|---|
| A_rsi50 | Exit at the open on the next trading day after RSI returns to 50 or above (if it doesn't, exit at the close on the 60th trading day). No stop loss |
| B_10bars | Exit at the close on the 10th trading day after entry. No stop loss |
| C_stop5 | Stop loss at -5%. If the stop isn't hit, exit at the close on the 10th trading day. If the stock gaps down at the open, the fill is at that open price, so a single trade's loss can exceed -5% |
About survivorship bias
This test uses the current S&P 500 constituents. Stocks that were delisted in the past, or removed from the index due to deteriorating business performance, are not included. This may make the backtest results look more favorable than they would if those stocks were included. Please keep this in mind when reading the results.
RSI below 20 results: the 3 patterns side by side
Overall summary
| Metric | A_rsi50 | B_10bars | C_stop5 |
|---|---|---|---|
| Sample size | 716 | 801 | 885 |
| Share of trades closed in profit | 69.1% | 59.3% | 40.3% |
| Average P/L | +3.45% | +1.67% | +0.19% |
| Average gain (trades closed in profit) | +9.57% | +7.52% | +7.29% |
| Average loss (trades closed at a loss) | -10.69% | -7.01% | -4.66% |
| Profit factor | 2.09 | 1.60 | 1.07 |
| Largest loss (single trade) | -87.9% | -55.4% | -25.3% |
| Longest losing streak (per stock) | 4 | 6 | 7 |
| Median MAE (maximum adverse excursion) | -6.88% | -4.52% | -4.76% |
| Average holding period | 24.3 trading days | 10.0 trading days | 6.8 trading days |
Note: Past sample; not indicative of future results.
Comparison across the 3 patterns
By sample size, the order is C > B > A. Partly because of the rule that skips overlapping positions in the same stock, sell-offs deep enough to push RSI down to 20 occur far less often than drops below 30.
The share of trades closed in profit, average P/L and profit factor all rank A > B > C. Pattern A, which waits for RSI to recover to 50, is the highest, and pattern C, which uses a stop loss and exits after 10 trading days, is the lowest.
Largest loss ranks A > B > C (largest first). Because pattern A keeps holding without a stop loss, it includes trades that suffered a large loss in a single stock. Pattern C uses a -5% stop loss, so its largest loss is the smallest, but when the stock gaps down, the fill is at that open price, so the loss can exceed the defined stop distance.
Longest losing streak (longest losing streak within a single stock) ranks C > B > A. Pattern C, where the stop loss is triggered often, had more consecutive losses.
Median MAE (median maximum adverse excursion while holding) ranks A > C > B (more negative first). Because pattern A holds for longer, its unrealized losses along the way also tend to be larger.
RSI below 20 vs RSI below 30: what's different
For the detailed backtest of RSI below 30, see Buying When RSI Drops Below 30; here we lay out the main points of comparison.
Comparison tables by pattern
| Metric | A_rsi50 (RSI below 20) | A_rsi50 (RSI below 30) |
|---|---|---|
| Sample size | 716 | 5,318 |
| Share of trades closed in profit | 69.1% | 72.2% |
| Average P/L | +3.45% | +2.21% |
| Profit factor | 2.09 | 1.92 |
| Largest loss | -87.9% | -93.3% |
| Median MAE | -6.88% | -4.73% |
| Metric | B_10bars (RSI below 20) | B_10bars (RSI below 30) |
|---|---|---|
| Sample size | 801 | 7,029 |
| Share of trades closed in profit | 59.3% | 56.9% |
| Average P/L | +1.67% | +1.20% |
| Profit factor | 1.60 | 1.51 |
| Largest loss | -55.4% | -72.5% |
| Median MAE | -4.52% | -3.57% |
| Metric | C_stop5 (RSI below 20) | C_stop5 (RSI below 30) |
|---|---|---|
| Sample size | 885 | 7,865 |
| Share of trades closed in profit | 40.3% | 45.9% |
| Average P/L | +0.19% | +0.54% |
| Profit factor | 1.07 | 1.23 |
| Largest loss | -25.3% | -37.9% |
| Median MAE | -4.76% | -3.76% |
Note: Past sample (N=716 to 7,865); not indicative of future results.
"The deeper the sell-off, the better" doesn't always hold
Drops below 20 occurred far less often than drops below 30, and this was true for all 3 patterns.
Comparing average P/L and profit factor, the figures were higher for RSI below 20 in patterns A and B. The share of trades closed in profit was higher for RSI below 20 in pattern B, but lower in pattern A (69.1% vs 72.2%). On the other hand, in pattern C all three figures were higher for RSI below 30, so the results do not support a blanket statement that "a deeper oversold condition is better."
For all 3 patterns, the largest loss was larger for RSI below 30 (more negative). With fewer occurrences, RSI below 20 had a smaller largest single-trade loss than RSI below 30. However, even for RSI below 20, pattern A's largest loss was a large -87.9%, and it still includes trades held without a stop loss.
For all 3 patterns, the median MAE was larger for RSI below 20 (more negative). The data suggests that after sell-offs deep enough to push RSI down to 20, unrealized losses tended to persist after entry, and large adverse moves along the way were common.
Breakdown by market environment and condition
By trend (pattern A)
| Trend | N | Share of trades closed in profit | Average P/L | Profit factor | Median MAE |
|---|---|---|---|---|---|
| Downtrend | 348 | 69.3% | +3.21% | 1.92 | -7.10% |
| Range | 325 | 68.3% | +2.95% | 2.01 | -6.99% |
Note: Past sample (N=348, N=325); not indicative of future results.
For pattern A, the share of trades closed in profit and average P/L were higher in downtrends. However, the median MAE was also larger (more negative) in downtrends, showing that unrealized losses along the way were larger during downtrend phases. Note as well that profit factor was higher in ranges.
By volatility (pattern A)
| Volatility | N | Share of trades closed in profit | Average P/L | Profit factor | Median MAE |
|---|---|---|---|---|---|
| High volatility | 389 | 71.2% | +5.61% | 2.62 | -10.05% |
| Low volatility | 39 | 69.2% | +1.48% | 1.95 | -2.57% |
| Normal | 288 | 66.3% | +0.81% | 1.27 | -5.30% |
Note: Past sample (N=389, N=39, N=288); not indicative of future results.
For pattern A, the share of trades closed in profit, average P/L and profit factor all rank high volatility > low volatility > normal. However, the high-volatility bucket also had the largest (most negative) median MAE, indicating periods in which unrealized losses can grow large.
Around earnings (pattern A)
| Distance from earnings | N | Share of trades closed in profit | Average P/L | Profit factor | Median MAE |
|---|---|---|---|---|---|
| Away from earnings | 187 | 70.1% | +3.22% | 2.31 | -6.32% |
| Within 5 trading days of earnings | 88 | 69.3% | +1.68% | 1.50 | -6.53% |
Note: Past sample (N=187, N=88); not indicative of future results. Earnings date data is available only from 2008 onward.
For pattern A, the share of trades closed in profit, average P/L and profit factor were all higher away from earnings. The sample within 5 trading days of earnings is small, so this is a limited observation, but results were more modest when earnings events overlapped.
Limitations and caveats
Small sample size: Drops below 20 are rare in themselves, and pattern A has a sample size of 716 trades. That is far smaller than RSI below 30 (5,318 trades for pattern A), and narrowing down by bucket can push the sample below a level that can be meaningfully interpreted.
Survivorship bias: As noted above, only current constituents are included, so the poor results of stocks that were delisted or removed from the index in the past are not reflected.
Differences from real execution: The backtest assumes mechanical fills at the open on the day after the entry signal and at the defined exit conditions. In real trading, slippage and differences in liquidity occur.
Fixed conditions: An RSI period of 14 and an entry level of 20 are the settings used in this backtest. Changing the period or level changes the results.
To see how this plays out for your own stocks, you can run the same conditions with ENTRIQ's Chart Replay. You can record and organize your results with Strategy Tags.
Summary
Here is a summary of the backtest of RSI below 20 (entering on the next trading day after RSI(14) drops below 20) across 3 patterns.
- Sample size is far smaller than for RSI below 30
- For patterns A and B, average P/L and profit factor were higher than for RSI below 30 (the share of trades closed in profit was higher only in pattern B), but for pattern C all three were lower
- Median MAE was larger (more negative) for RSI below 20 in every pattern, meaning large unrealized losses along the way were more likely
- The largest loss was larger for RSI below 30 in all 3 patterns, but even for RSI below 20, pattern A kept holding without a stop loss, so a single trade's loss could be large
- It cannot be said across the board that "the deeper the sell-off, the better the results"; the direction of the comparison changes depending on the exit conditions
Frequently asked questions (FAQ)
Q. Is RSI below 20 better than RSI below 30?
A. In this backtest, it cannot be said across the board that it is "better." For patterns A and B, average P/L and profit factor were higher than for RSI below 30, but for pattern C the order was reversed. Also, because occurrences are few, you need to check with your own stocks and period before drawing any conclusion.
Q. Why is the largest loss so large for pattern A, which has no stop loss?
A. Pattern A keeps holding without a stop loss until RSI returns to 50 or above (up to 60 trading days), so losses grow when a stock that fell sharply continues to be held. It is not that "the stop loss came too late"; the design has no stop loss in the first place.
Q. Isn't the RSI-below-20 sample too small to be useful?
A. Overall there is a reasonable number of trades, but by bucket (specific combinations of trend, volatility and earnings conditions), the numbers fall to a level that calls for caution in interpretation. Use this backtest as a reference for your own testing approach, and carry out additional checks on the stocks and conditions you trade.
Q. Can I backtest with an RSI period other than 14?
A. This article reports a backtest with a period of 14. Changing the period changes the number of occurrences and the results. You can try different settings with ENTRIQ's Chart Replay.
Q. Can I use these results directly in my own trading?
A. Results change if the stocks, period, timeframe or execution conditions differ. The figures in this article are a reference for "comparing conditions" and do not recommend any specific trade.
Related articles
This article is provided for informational purposes to examine and organize historical data and does not recommend any specific trade. The figures shown are past samples and do not indicate future results. All investment decisions are your own responsibility.
Reproduce this validation yourself
With ENTRIQ's chart replay, you can trade through past charts using the same rules.
Try the demo — no sign-upNo sign-up or card needed. Runs right in your browser. Start 14-day free trial →
Validate your method on past charts
The same validation in this article can be reproduced with your own method. On past charts, do the classics really work?
- ✓ Replay past charts
- ✓ AI feedback on your trades
- ✓ Validate methods with statistics
No sign-up or card needed. Runs right in your browser.