The two tools answer different questions
The Pro section currently has two prototype features. Many people assume they do the same thing — both "find strong stocks" — but they do not.
- The New-high Screener asks: where does this stock stand against its own past?
- RS Ranking asks: how fast is it moving compared with everything else over the same window?
One is absolute strength, the other relative strength. Each covers the other's blind spot.
A stock can have a high RS and still sit far below its prior high — meaning it merely fell less than others, or bounced out of a deep hole, with a wall of trapped supply still overhead. Conversely, a stock can have just made a 6-month high with an RS of only 60 — meaning everything is rising and it is simply going along for the ride.
Stocks that satisfy both are far fewer, but that is what "the market is actually buying this" looks like.
What RS is — and what it is not
This has to come first, because RS is one of the most misread names in the business.
RS is not RSI. RSI is an oscillator built from a single stock's own gains and losses, telling you about overbought and oversold. RS ignores all of that.
RS is also not IBD's proprietary RS Rating. That is a commercial metric with its own methodology and database. The RS here is our own price percentile — similar in spirit, different in number.
The actual definition is simple: take the top 300 companies by market cap, compare their price returns over the selected window, and convert that into a percentile from 1 to 99. Weakest 1, strongest 99.
So RS 90 means this stock's return over that window beat roughly 90% of the stocks with usable data.
Three things to keep in mind:
- Price only — no dividends, no fundamentals. It measures where money is flowing, not how good the company is.
- The universe is fixed at the top 300. This is relative strength *among large caps*, not the whole market. Small- and mid-cap movers are not in this pool.
- RS is a relative score, not an absolute return. In a bear market an RS of 99 can still be negative — it simply fell less than the rest. This is RS's biggest trap, and it comes up again below.
Four windows are really four different questions
RS offers 20, 60, 120 and 250 trading days. The default is 60.
- 20 days (short term): money flow over the past month. Fast-moving and noisy; useful for spotting something that has just started.
- 60 days (swing): about a quarter. The default, and the most practical single setting — it matches the holding period of most momentum positions.
- 120 days (intermediate): about half a year. Long enough to separate a real trend from a rebound.
- 250 days (long term): about a year. Shows who has led the market across the whole year.
No single window tells you enough on its own. The useful move is to read two windows against each other:
- Long high + short high: the leaders are still leading; the trend continues.
- Long high + short low: a former leader is fading. These are the dangerous ones — the reputation remains, the money has left.
- Long low + short high: could be laggard catch-up, could be a new group starting. Judge it alongside new highs and sector — if several names in the same industry turn strong together, that is worth attention.
The method is simple: on the same day, run 250 days and then 20 days, and compare which names appear on both lists and which appear only once.
How to use the new-high screener
The New-high Screener also scans the top 300 and offers last day, last 3 days, or last 5 days. Any single qualifying session inside that window counts; results show the actual date and the highest tier reached (all-time > 5-year > 3-year > 1-year > 6-month).
- Last day: the names that appeared today. A short list — good for a daily check.
- Last 3 / 5 days: the fuller picture of the week. Good for a weekly review.
Above the results there are two dropdowns — industry and new-high tier — and those are where the value is.
Use the industry filter to spot groups. If three or five names from one industry show up on the new-high list, that is usually not a single-stock move but a whole group turning. Group moves persist far better than isolated ones. The reverse also holds: a stock making a new high while its peers do nothing deserves an extra dose of skepticism.
Use the tier filter as a quality bar. A 6-month high is a watch item; one year and above carries more weight. The meaning of each tier and the reasoning behind the 2% tolerance are covered separately in What the "New High" Column Means.
A complete working sequence
Here is how the site's tools chain together in practice. The point is that each step narrows the list — no step goes looking for new names.
Step 1: check the market stage. Open Today's Market on the home page and see where the Market Lifecycle currently sits. This step decides not what to buy but whether to buy, and how large. In a headwind stage, an RS ≥90 list still returns 30 names — that does not make them tradable.
Step 2: narrow with RS. Start at 60 days, RS ≥90. If the list is too long, raise to ≥95; too short, drop to ≥80. Adjust the threshold, not the window — changing the window changes the question; the threshold only changes how tight it is.
Step 3: confirm position with new highs. Cross-reference the RS list against the new-high list and prioritise names on both. This step confirms that a stock is not just fast but also has nothing overhead.
Step 4: converge by industry. Within the intersection, see which industry appears most often. Names clustered in the same group move up the priority list.
Step 5: check trend structure in the Screener. Verify the moving-average alignment. RS and new highs are outcomes; the MA structure tells you whether those outcomes rest on a stable trend.
Step 6: wait for an entry with Pullback Buy Points. Chasing on the breakout day usually means the worst price of the move. Waiting for a pullback toward a moving average gives a far better risk-reward.
Step 7: decide the exit before you enter. Set the stop on the risk-management page under My Watchlist (it provides 14-day ATR with 2×ATR initial and trailing stop suggestions), then set a price alert. The stop is written down before the purchase, not after the loss.
Seven steps sounds like a lot, but running RS and the new-high screen takes under two minutes; everything after that is confirmation work on one list.
Adapt to the market stage instead of reusing one setting daily
The same parameters mean different things in different environments:
- Uptrend / tailwind: RS ≥90 at 60 days, paired with new highs. Normal operation — this is where the workflow works best.
- Bull pullback: lead with new highs plus pullbacks, use RS as support. Short-window RS jumps around during pullbacks and is less informative.
- High-risk consolidation: raise the threshold to ≥95, keep only the strongest, and cut position size. A shorter list is information, not a malfunction.
- Headwind market: the judgment that matters most here is "don't". RS is relative — an RS of 99 in a bear market may well be −10% while everything else is −25%. Treating that list as a buy list means shopping for the slowest decliners in a downtrend. This is the classic misuse.
- Early / confirmed recovery: this is where 250-day RS earns its keep. Coming off a bottom, whatever held a high long-term RS through the entire decline is the candidate leadership for the next advance.
Four common misuses
1. Treating RS as a trade signal. A high RS is not a reason to buy — only a reason to look closer. It carries no entry, no stop, and no time frame. All three are yours to supply.
2. Cherry-picking only the highest RS. RS 99 is the most crowded position in the market. Crowded does not mean doomed, but it does mean most people are already inside, leaving less marginal buying. A stock at RS 85–95 that has just made a 1-year high often sits in a more comfortable spot.
3. Using 20-day RS to chase what has already exploded. The highest short-window RS names are usually whatever rose hardest in the past few days. Entering there puts your stop as far away as it can possibly be.
4. Forgetting the universe is only 300 names. This RS answers "who is strongest among large caps". Genuine small-cap movers are not in the pool — not seeing them does not mean they are not there.
Known limitations, stated plainly
- Both features are prototypes. Subscription access is not enabled yet, and the conditions may change.
- The RS universe is *today's* top 300. Membership shifts with market cap, so an RS calculated today and one three months from now are not measured against an identical peer set.
- 250-day RS needs 251 sessions of closes. The first run — or one right after a deployment, before the cache is warm — will be noticeably slower. It is fast afterwards.
- New highs use roughly six years of long-term closing data, a separate data path from the moving-average screen, so the two data dates can occasionally differ by a day.
- This workflow has not been fully backtested. It documents usage logic, not verified performance. Any process that reads this smoothly still has to be recorded and reviewed against your own results.
One last thing
New highs and RS both do exactly one job: shrink thousands of stocks down to a list you can actually read.
They will not tell you when to buy, how much, or when to admit you were wrong. Those three remain your work — and they matter far more to your results than whether you set the threshold at RS 90 or RS 95.
Want the same workflow for US stocks?
The sequence transfers, but the parameters do not. On US Stock Coffee the moving averages are 10/20/50/150, new highs top out at five years (no "all-time high" tier, because the history is five years), and the market-stage call uses market breadth rather than margin balance — the US has no equivalent of Taiwan's daily margin data.