We present an on-line investment algorithm that achieves almost the sa
me wealth as the best constant-rebalanced portfolio determined in hind
sight from the actual market outcomes. The algorithm employs a multipl
icative update rule derived using a framework introduced by Kivinen an
d Warmuth. Our algorithm is very simple to implement and requires only
constant storage and computing time per stock in each trading period.
We tested the performance of our algorithm on real stock data from th
e New York Stock Exchange accumulated during a 22-year period. On thes
e data, our algorithm clearly outperforms the best single stock as wel
l as Cover's universal portfolio selection algorithm. We also present
results for the situation in which the investor has access to addition
al ''side information.''