- Researchers warn collective crypto trading strategies are vulnerable to exploitation.
- These schemes promise to place institutional-grade investment strategies within the reach of retail investors.
A new academic paper warns that collective crypto trading schemes — which pool user funds and trade automatically with them — are inherently vulnerable to losing profits and being exploited by insiders, no matter how they are designed.
These findings come from research by academics at Cornell Tech. In their research, they examined the fundamental trade-offs between profitability and economic fairness in such systems.
The team of eight researchers who co-authored the January 2 paper found that these Collective Investment Algorithms, or CoinAlgs, posed unavoidable risks, regardless of whether they choose to keep their trading strategies private or opt for full transparency.
“CoinAlgs can either be transparent, and risk losing profits; or be private, and open the door for unfair value extraction by insiders,” the researchers said. “Even in scenarios where a CoinAlg may seem benign, subtle backdoors can be inserted that lead to profits from adversaries.”
Institutional-grade software
Collective Investment Algorithms are not a new phenomenon. They have existed in the traditional financial world for years, and are used at asset management firms like BlackRock and Renaissance Technologies to execute trades across multiple client portfolios.
But in crypto trading circles they’ve surged in popularity in recent years, supercharged by the proliferation of artificial intelligence.
They promise to place institutional-grade investment strategies or specialised trading intelligence within the reach of retail investors, although these tools often lack the investor protections of regulated intermediaries.
Many CoinAlgs operating across DeFi skew heavily towards speculative assets, with a preference for emerging tokens and highly volatile memecoins.
Value extraction on Uniswap
The researchers used historical transaction data from Uniswap, the biggest decentralised exchange, to simulate a profitable CoinAlg that correctly predicts future asset prices.
In this situation, they found that even minimal information leakage from a private CoinAlg
Enabled significant value extraction by arbitrageurs, investors who exploit temporary price differences to make small but consistent profits.
At the same time, those who use private CoinAlgs experience the additional risk of insiders using privileged information to conduct similar arbitrage or frontrunning trades to extract profits.
Transparent CoinAlgs didn’t fare much better.
They avoid issues of insider trading by ensuring all algorithmic trading models and data is open source and public. Yet this makes it even easier for arbitrageurs to extract profits from the CoinAlgs’s trades, with little ability for recourse.
“Even moderate defensive strategies against arbitrage incur a significant cost of transparency and thus [eroding] profits,” the researchers said.
Yet despite the issues, CoinAlgs are likely here to stay. Many investors are drawn to them because they advertise the ability to generate lucrative returns by outsourcing difficult investment decisions to AI algorithms.
The interest in such products is only set to increase as companies like OpenAI, Anthropic and Google continue to pour money and resources into AI development.
“CoinAlgs are an inevitable part of the financial landscape,” the researchers said. “The search for principled guardrails remains an important avenue for future work.”
Tim Craig is DL News’ Edinburgh-based DeFi Correspondent. Reach out with tips at tim@dlnews.com.
