A 100% win rate is a red flag, not a résumé
What running a fleet of autonomous trading bots taught me about picking the right metric. I run a small fleet of automated trading bots in paper mode. One of them — a mean-reversion strategy on crypto — closed its first 34 trades without a single loss. 34 out of 34. A 100% win rate. If I put that number on a landing page, it would sell. It's also close to meaningless. Here's why, and why it…
Running a fleet of autonomous trading bots taught me the importance of choosing the right metric. One bot, using a mean-reversion strategy on crypto, achieved an impressive 100% win rate in its first 34 trades. However, this metric alone is misleading. The strategy operates by adding to positions when they move against it, reducing the average entry price and turning losses into wins.
This approach exposes hidden risks like capital tied up in underwater positions and increased tail risk during market downturns. Additionally, the 34 trades in a calm market only reflect the strategy's performance in a stable environment, not how it would handle a downturn. The win rate, therefore, is a vanity metric that fails to capture the bot's true performance.
In evaluating autonomous systems, I now prioritize a minimum number of trades (>=100), surviving real corrections in paper mode, and assessing risks like drawdown and time under water through Monte Carlo reshuffling of trades. This approach ensures that the evaluation reveals potential failure modes before they lead to costly consequences.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
