1.)Getting a server close to the exchange 2.)Finding a good predictor. Predicting up or down is easy on paper. Finding a predictor that a.)beats the spread,

1.)Getting a server close to the exchange 2.)Finding a good predictor. Predicting up or down is easy on paper. Finding a predictor that a.)beats the spread, b.)factors in lag time to execution, and c.)factors in commission, is hard. 3.)Market regime changes. What works a few months ago isn't guaranteed to work now.