Calculation Highlights
RPS on this site is calculatedunder the same market, same period and same rulesby computing each issue’s price performance and converting the descending rank into a percentile. Japanese and U.S. stocks are scored in separate universes and are never combined into a single mixed ranking.
1. Calculate the price-change rate over the specified period
Stock i as of t for n The n-day RPS input is the price-change rate from the same base date n back n trading days.
This assumes split-adjustment conventions (and similar) are aligned across stocks. Issues missing a start or end price, or lacking the required listing history, are not forced into that period’s valid universe.
2. Rank in descending order within the same market
Japan-stock RPS compares only valid Japanese issues; U.S.-stock RPS compares only valid U.S. issues. The highest-return stock is ranked 1 and the rank is then converted to a percentile score with the following formula.
Under this definition, RPS90 indicates a position that outperformed approximately 90% of valid issues over the period. It does not mean the price rose 90%.
3. Worked examples for 100 stocks
| Rank | Number of valid issues | Calculation | RPS | How to read it |
|---|---|---|---|---|
| Rank 1 | 100 | (1−1÷100)×100 | 99 | Near the very top of the market |
| Rank 10 | 100 | (1−10÷100)×100 | 90 | Outperformed roughly 90% of issues |
| Rank 20 | 100 | (1−20÷100)×100 | 80 | Outperformed roughly 80% of issues |
| Rank 50 | 100 | (1−50÷100)×100 | 50 | Near the middle of the universe |
4. Ties, missing values and the universe
When multiple stocks share the same period return, this site assigns the entire tied group the highest rank. Stocks lacking the required history are excluded from that period’s ranking. Consequently the valid-issue count may differ between RPS20 and RPS250.
To preserve comparability, store the market, period, observation date and number of valid issues together with the score. The latest valid sample can be checked in thepublic data directory.
5. Differences among RPS20, 50, 120 and 250
| Period | Time horizon captured | Key caveats |
|---|---|---|
| 20 days | Approx. 1 months | Significant impact from short-term news and temporary volatility. |
| 50 days | Approximately 2 to 3 months | Confirmed across multiple periods as a bridge between short-term and medium-term. |
| 120 days | Approximately six months | Slow to react to short-term reversals. |
| 250 days | Approximately 1 years | Cannot generate valid values with short listing history. |
6. Difference from weighted formulas or benchmark index ratios.
In Japanese search results, there are also ways to weight the 63・126・189・252 day returns or to call the ratio of individual stock prices divided by the market index as RS/RPS. Both are different definitions with research purposes. This site does not claim to fully reproduce IBD's proprietary RS Rating using the publicly available cross-sectional ranking rules for each period individually.
Frequently Asked Questions on RPS Calculation
How is RPS calculated?
Calculate the price change rates for stocks in the same market over the same period, rank them in descending order, and convert to percentiles.
Why is the highest RPS value near 99 rather than 100?
In the definition dividing rank by the number of valid stocks, the 1 th rank is still less than 100. The upper bound expression may differ due to rounding in the display or service formula.
How are stocks with the same return handled?
This site assigns the highest rank to the top of the tied rank group. Please also check the tie-breaking rules when comparing.
Are RPS and IBD's RS Rating the same?
They are not necessarily the same. This site does not claim to reproduce IBD's proprietary calculations using the publicly disclosed cross-sectional percentile rules.