Quick definition Statistical arbitrage (stat arb) uses statistical models to identify mispriced securities and profits from the correction back to fair value. What it is Stat arb strategies build statistical models of security relationships. For example, if historical data shows that Apple and Microsoft prices move together, but they diverge today, a stat arb strategy might short Microsoft and buy Apple, betting they will converge again. Stat arb is automated and algorithmic. Computers identify opportunities and execute trades. Why it matters Stat arb is one of the most profitable trading strategies on Wall Street. Large hedge funds dedicate teams to stat arb research and development. Stat arb also improves market efficiency. Stat arb traders identify mispricings and trade them away, pushing prices back to fair value. Relationship types Stat arb models can exploit: - Sector relationships (tech stocks tend to move together) - Factor relationships (value stocks move together) - Pairs relationships (two competitors move together) - Mean-reversion relationships (prices revert to historical averages) Practical example A stat arb model studies the relationship between two airline stocks. Historically, they have moved together within a 5 percent range. One day, one airline announces good news and its price jumps 3 percent. The other airline's price hasn't moved. The model assumes they will converge. The strategy buys the lagging airline and shorts the leading one, betting they will converge. If they do, both sides of the trade profit. Risks Stat arb relies on historical relationships continuing. If market structure changes, the relationship breaks and the strategy fails. Stat arb also involves leverage. Strategies often use 10x leverage or more, so small losses become large ones. See also - Arbitrage - Algorithmic Trading - Pairs Trading - Mean Reversion
