Simons emphasized that successful quantitative trading required more than just prediction. Three elements were essential:
**1. Prediction Models**
The firm developed sophisticated algorithms to forecast price movements. These weren't elaborate equations but rather machine learning systems that tested hypotheses against historical data. If a predictor showed statistical significance over long periods, it joined the model.
**2. Transaction Cost Modeling**
"The average person will buy 200 shares and won't move the market," Simons explained. "But if you want to buy 200,000 shares, you're going to push the price." Understanding market impact was crucial. How much would a trade move the price? At what point do transaction costs overwhelm profit?
**3. Portfolio Optimization**
Minimizing volatility across the entire portfolio required "fairly sophisticated applied mathematics" - statistics, probability theory, and optimization techniques. The goal was generating consistent returns with surprisingly low risk.
The firm maintained strict capacity limits. The Medallion Fund could only manage a certain amount before its trades would move markets too much. This discipline prevented the fund from becoming an uncontrollable behemoth.
Infrastructure was paramount: massive data collection, custom software for rapid hypothesis testing, and the best computational resources available.