Crypto cross-sectional factors
CryptoAdvanced
Academic work on cryptocurrencies finds that the cross-section of coin returns has factor structure, including momentum and attention-related effects, in the samples those papers study. A rules-based version ranks a survivorship-safe set of coins on a precommitted factor, such as recent return, and holds the leaders against the laggards. It is not a product, a staking yield, or a bot.
Not investment advice. Nothing here is a recommendation to buy or sell anything. Backtests and historical studies are not guarantees, and a rule that worked in one sample can fail in the next. Trading can lose money. Links are for education, not endorsements. Do your own research.
- Rules
- Define the universe by coins that were listed and tradable on the signal date, with a volume floor. Rank on a stated factor such as a multi-week return. Long the top group and short or underweight the bottom group if borrow or inverse exposure actually exists. Rebalance on a calendar. Ignore coins that appear in the database only after they succeed.
- Where AI fits
- AI can build a point-in-time universe, which is the hardest part, and can apply the same rank rule every period. It should not add a coin because a current index includes it. Backtests that start from today's top coins by market cap are not evidence. The model can report when the factor book is just bitcoin in disguise.
- Markets
- Large, liquid cryptocurrencies on exchanges you can actually access. Tiny coins are not the academic universe once a volume floor is honest.
- Timeframes
- Weekly or monthly rebalance.
- Risks
- Crypto factor results are sample-specific and can be dominated by one coin or one era. Shorting altcoins may be impossible or unsafe. Exchange failure, forks, and stablecoin or custody risk are not in an equity-factor backtest. Published patterns are not expected returns, and nothing here is a guaranteed yield.