Direct Answer
An RSI screen filters a universe of securities to find those whose Relative Strength Index reading falls within a specified range, most commonly below 30 (potentially oversold) or above 70 (potentially overbought). It is a filtering tool, not a trading system: it narrows a large watchlist down to a shorter list worth researching further. Because an extreme RSI reading can persist through a strong trend rather than reversing right away, screen results are a starting point for research, not a standalone buy or sell signal.
Key Takeaways
- An RSI screen ranks or filters securities by their current Relative Strength Index value.
- The most common filter thresholds are RSI below 30 (potentially oversold) and RSI above 70 (potentially overbought).
- RSI is a bounded momentum oscillator that runs from 0 to 100, developed by J. Welles Wilder.
- An RSI screen's output is a candidate list, not a signal, it tells you where to look, not what to do.
- RSI can stay extreme for an extended period during a strong trend, so a low or high reading alone does not predict an imminent reversal.
- Screens are typically combined with trend, volume, or support/resistance context before a decision is made.
- Different screening tools let you set custom thresholds instead of the default 30/70 split.
How Does an RSI Screen Work?
The Relative Strength Index (RSI) is a momentum oscillator, developed by J. Welles Wilder, that compares the size of recent gains to recent losses over a lookback period (14 periods is the traditional default) and expresses the result on a 0-to-100 scale. An RSI screen applies that calculation across many securities at once and returns only the ones whose current RSI value falls inside a range the user specifies.
In its most common form, the screen has two boundaries: a lower bound, typically 30, below which a security is described as potentially oversold, and an upper bound, typically 70, above which a security is described as potentially overbought. Some screens let a user tighten these bounds (for example 20/80) to surface fewer, more extreme readings, or widen them (40/60) to capture a broader set of candidates.
A Worked Example
Suppose a screen is run across a watchlist of 500 stocks using the standard 14-period RSI with a filter set to "RSI below 30." The screen calculates RSI for every stock in the list and returns only those currently reading under 30, say, a subset of 18 names. Those 18 stocks have all seen recent losses outweigh recent gains enough to push RSI into the traditional oversold zone; the screen does not say anything about why, or whether the decline is likely to pause.
From there, a trader would typically look deeper into each of the 18: is the stock in a longer-term downtrend, or was this a short-term pullback within an uptrend? Is volume rising or falling into the decline? Is there a nearby support level? The screen's job ends at producing the shortlist; the research on each candidate begins after.
Limitations and Common Mistakes
- Treating a screen hit as a signal. An RSI screen surfaces candidates; it does not confirm that a reversal is imminent.
- Ignoring trend strength. In a strong trend, RSI can remain below 30 or above 70 for an extended stretch rather than snapping back, so acting on the reading alone can mean fighting the trend.
- Using one fixed threshold for every security. A 30/70 split that works for a range-bound stock may behave differently for a high-momentum or highly volatile name.
- Skipping further research. Volume, broader trend, and price structure around support or resistance all add context a bare RSI number cannot provide on its own.
One Threshold Across Very Different Securities
A screen applies 30 and 70 to every name in the universe, and those levels do not mean the same thing across securities. A steady, range-bound stock reaching 28 has done something unusual. A high-volatility name that visits 28 most months has not, and a momentum name in a strong trend may spend long stretches above 70 as its normal state. The screen treats all three identically, so the output mixes genuinely unusual readings with routine ones.
Calibrating against each security own RSI history is the fix, and it is work the screen does not do. Short of that, knowing which category a candidate falls into before reading anything into its number removes most of the false interest.
The persistence problem applies here as elsewhere. RSI can remain below 30 or above 70 through a sustained trend rather than snapping back, so a screen hit describes a current condition and carries no timing. Acting on the reading alone frequently means trading against the trend that produced it.
Treat the output as a place to look. Volume, the broader trend and whether price is sitting at a level that has mattered before all add context a bare oscillator value cannot contain, and that context is what turns a candidate into an idea or discards it.
Frequently Asked Questions
What RSI range counts as oversold or overbought?
The most common convention flags a reading below 30 as potentially oversold and above 70 as potentially overbought. These are widely used starting thresholds, not fixed rules, and some traders adjust them for a given security or market regime.
Is an RSI screen a buy or sell signal?
No. An RSI screen is typically used as a starting point for further research rather than a standalone buy or sell trigger. An extreme RSI reading can persist through a strong trend rather than immediately reversing, so screen results need additional confirmation.
Can a stock stay oversold or overbought for a long time?
Yes. During a strong trend, RSI can stay below 30 or above 70 for an extended stretch rather than reversing quickly, which is why screen results are treated as candidates for review, not automatic signals.
What comes after an RSI screen finds a candidate?
Traders typically layer on further research such as trend context, volume, support and resistance, or other indicators before deciding whether the RSI reading reflects an actionable setup or simply reflects a persistent trend.
What RSI period do screeners use, and is it disclosed?
Fourteen periods is the near-universal default, inherited from the original publication. Many screening interfaces do not show the setting at all, and some allow it to be changed in a place users do not look. Since the period determines how often the thresholds are crossed, two screens described identically can return quite different lists, and the difference is not visible in the output.
Do different platforms compute RSI the same way?
Not exactly. Wilder original smoothing, which is a specific recursive average, gives slightly different values from a plain simple average of gains and losses over the same window. The gap is small and it lands precisely where it matters for a screen: on securities sitting near the threshold, where a small difference decides membership. Two screeners can therefore disagree about the same security on the same day.
Does the screen use adjusted or unadjusted prices?
It changes the result, because RSI is computed from the sequence of gains and losses and price adjustment rewrites that sequence. An unadjusted series containing a split shows an enormous single-day loss that dominates the calculation for the length of the lookback. Dividend adjustment has a smaller but persistent effect. Screens rarely state the convention, and it is worth establishing before trusting a borderline reading.
How much history does RSI need before its values settle?
More than fourteen bars, because Wilder smoothing carries the influence of the seed value forward indefinitely rather than dropping it after the nominal period. Values computed near the start of a data window reflect the initialisation as much as the prices. For a screen run across a universe including recent listings, some of the returned readings are artefacts of where each series began.
Can an RSI screen be run on a ratio or spread series?
Mechanically yes, and the interpretation changes. RSI computed on a ratio between two securities measures the balance of gains and losses in relative performance, so an extreme reading describes one leg having outpaced the other consistently rather than an absolute price move. That is a legitimate and different question, and the conventional thresholds carry no established meaning on such a series.
References
Disclaimer
This page is for educational purposes only and is not personalized investment, financial, tax, or legal advice. RSI screen results are a research starting point, not a recommendation to buy or sell any security. Past price behavior, including past reversals from oversold or overbought readings, does not guarantee future results.