Skip to content
All library documents

Rebasing Price Series to Compare Instruments on One Chart

Article Quant Q&A · Author: Vladimir

Summary

This exchange addresses how to compare several instruments with different nominal price levels on a shared chart. The questioner finds that raw prices and log prices retain scale differences, while single-period returns are too small to make the paths easy to compare visually.

The suggested solution is to rebase each price series to a common starting value, effectively converting cumulative returns into an index such as a base of 100. This makes relative performance paths directly comparable; plotting the rebased series or its logarithm are both suggested. The method assumes the instruments cover the same start and end dates so they can share a common base period. It is a visualization technique for comparing indexed performance, not a normalization that makes volatility or risk equivalent across instruments.

Key ideas

  • Rebase each instrument’s price series to a common starting value for visual comparison.
  • A shared index base expresses subsequent prices as cumulative relative performance.
  • Rebased series can be plotted directly or on a logarithmic scale.
  • The suggested comparison assumes the instruments share aligned start and end dates.
  • Rebasing does not equalize volatility or risk across instruments.

Tags

Full text
# Different range price data on one chart


# Different range price data on one chart












I'd like to evaluate 3-4 instruments on one price chart. For example:







I try to use:

- Ln(close) - too different values, to put them on one chart, same results with Close(0)-Average(10)

- returns Close(0)-Close(-1), (Close(0)-Close(-1))*100/Close(-1) - deviations are too small for evaluating Is there any way to normalize different price data on one Y axis to get look like this

## Answer by Viquar (score 2)

https://quant.stackexchange.com/a/10854

have you tried to rebase the timeseries? if the start and end dates are same, you could basically have 100*(1+returns). easy to then plot either directly or log of the new time series.

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.