Skip to content
All library documents

Calculating RSI Across Market Closures and Trading Gaps

Article Quant Q&A · Author: Separius
Read the original

Summary

The answer recommends calculating RSI on one continuous series of observations for the chosen data frequency, excluding weekends, holidays, and other times when the market is closed. For hourly equity data, this means joining observations from successive open-market periods rather than inserting artificial bars for closed hours. For daily data, non-trading days are omitted from the series.

The explanation warns that overnight price gaps can materially affect RSI, especially at finer sampling intervals. It does not provide a numerical example or compare alternative RSI conventions, and it leaves the handling of large gaps to a trader’s own decision rules. The guidance is a general convention, with room for a different method when a particular indicator design calls for one.

Key ideas

  • Calculate RSI on a continuous sequence of market observations at the selected frequency.
  • Omit weekends, holidays, and other periods when the market is closed.
  • Overnight price gaps can significantly affect the indicator, particularly for hourly data.
  • Consider rules for significant overnight moves when using RSI to make trading decisions.

Tags

EquitiesTechnical indicatorsUS markets
Full text
# How to calculate RSI while considering market close and holidays?


# How to calculate RSI while considering market close and holidays?












I was trying to calculate RSI over hourly OHLC bars for a symbol (AAPL as an example) and got stock, first how should I handle closing hours? (does it make sense to ignore them all together and assume that next day's data is following today's data [2018-01-01/16:00 -> 2018-01-02/09:00 -> 2018-01-02/10:00]) and second how should I handle holidays and weekends? (assuming I'm going to use daily data instead of the hourly data)

## Answer by Jared M (score 0, accepted)

https://quant.stackexchange.com/a/41271

Traditionally, you should have one continuous time-series for any frequency of data and calculate your indicator based on that. That is, unless you have some other desired or proprietary method for calculating your indicators.

You should generally exclude times when the market is closed, such as weekends and holidays, as this would produce undesired bias to the indicator's value.

So in your case, calculate the RSI based on a single continuous time-series that does not include weekends, holidays or times when the market is not open.

Note - as you probably suspected, gaps in the overnight price can have a significant impact on the reading of your indicator, especially depending on the frequency of the data. So, if you are making trading decisions based on this information, you should keep that in mind or have some sort of rules based criteria in the event of significant overnight changes.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.