Bitcoin trading bot reveals top-performing strategies

A team built a Bitcoin trading bot and tested various strategies, finding that a simple rand-cost-average approach remains tough to beat. The bot used AI to test strategies on Bitcoin (BTC) and Ethereum (ETH), focusing on their volatility, liquidity, and popularity among traders. The best strategy over the period from 1 August 2018 was holding a 70/30 mix of BTC and ETH, rebalanced quarterly, yielding a total return of 1,011% (34.3% annually). However, this came with significant drawdowns, which might be uncomfortable for some investors.

Another effective strategy involved switching between BTC and ETH based on which was trading above its 200-day moving average. This delivered a 1,004% return (34.2% annually), but also included substantial drawdowns. The bot's rules for switching were straightforward, with cash held in USDT during bear trends, which helped boost performance.

Dollar-cost averaging (DCA), where a fixed amount of BTC or ETH is bought weekly, provided a solid 381% return. The team also tested the Gaussian channel strategy, which outperformed others with a 2,955% return from 1 January 2018. However, this strategy's success depends on BTC continuing to have major bull runs, and its performance drops significantly without its best trades.

The team cautioned that while these strategies showed promising results, they are not investment advice. The success of these bots in real-world conditions remains uncertain, especially given varying costs and the assumption of perfect order fills.