In order to analyze some controversies on the equivalence between ballistic
deposition (BD) and the Kardar-Parisi-Zhang (KPZ) theory, we simulated the
BD model in one and two dimensions. Effective exponents beta (L) were obta
ined in he growth regions, which were rigorously determined for various len
gths L. Effective exponents alpha (L) were obtained from saturation widths
in the steady-state regimes. In d = 1 we found beta (L) = beta + AL(-lambda
) and alpha (L) = alpha + BL-Delta, with asymptotic exponents consistent wi
th the KPZ values beta = 1/3 and alpha = 1/2, and correction-to-scaling exp
onents 0.2 less than or similar to lambda less than or similar to0.4 and 0.
6 less than or similar to Delta less than or similar to0.8. These strong fi
nite-size corrections explain the previous discrepancies between numerical
estimates for ED and the exact KPZ results. In d = 2 we could only obtain r
eliable estimates of alpha (L), which are consistent with KPZ values if fin
ite-size corrections with Delta approximate to 0.4 are considered.