MOPTA 05 -- Modeling & Optimization:  Theory and Applications
Windsor, ON, Canada / 25 - 27 July 2005

Model-Based and Model-Free Optimization 
of a Stock Day-Trading Strategy

Optimization of a stock day-trading strategy 
for a single asset is considered.   
Any position must be closed out at the day's end.  
Interest may be earned on days not in the market; 
note that this problem differs from that of merely 
optimizing the return in excess of the interest rate.
 
Model-free, data-based optimization of the strategy turns out 
to lead to a relatively simple mathematical programming problem.  
Model-based optimization according to a first-order 
autoregressive model is also accomplished.  
The procedures are illustrated with actual stock rates of return.