Skip to content
All library documents

Why Crypto Backtests Need Delisted Coins in Historical Data

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

Summary

The post asks for a historical crypto data source that avoids survivorship bias and includes coins that were later delisted. It highlights a research concern: studying only assets that remain listed can distort conclusions about past market performance. The author wants coverage of at least the historically largest coins and is open to programming a solution, but the document does not identify a provider or describe a data collection method.

The cited paper is offered as motivation for treating survivorship and delisting bias as material issues in crypto research. No results, provider comparisons, or implementation details are included, so this is a research question rather than a practical guide. Its main lesson is to check whether a historical dataset retains failed or delisted assets before using it to evaluate strategies; the post itself does not establish which datasets meet that requirement.

Key ideas

  • Historical crypto research can be distorted when datasets omit delisted assets.
  • The post seeks coverage that includes the historically leading coins, including those no longer listed.
  • It cites research as motivation but supplies no data source or implementation approach.
  • Researchers should verify historical asset coverage before relying on crypto backtests.

Tags

CryptoStatisticsBacktesting
Full text
# Where to get crypto data without survivorship-bias (delisted coins accounted for)?


# Where to get crypto data without survivorship-bias (delisted coins accounted for)?












What are folks using for their crypto data needs.

I am looking for a provider of Crypto data - free of survivorship bias and should at least have the top 20 coins historically.

As this research paper states, survivorship and delisting bias are a big deal in crypto. Survivorship and Delisting Bias in Cryptocurrency Markets: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4287573 "Our results suggest that the survivorship and delisting bias are important biases that ought to be omitted.”

I am comfortable with programming too so if you know of any decent solution out there that I can use, that would be great.

Many Thanks. And Happy New Year!

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.