Dauphas and Pourmand (2011) [Nature 473, 489--492] estimated the accretion timescale of Mars to be 1.8 $^{+0.9}_{-1.0}$ Myr from the W isotopes of martian meteorites. This timescale was derived assuming perfect metal-silicate equilibration between the impactor and the target's mantle. However, in the case of a small impactor most likely only a fraction of the target's mantle is involved in the equilibration, while only a small part of the impactor's core equilibrates in the case of a giant impact. We examined the effects of imperfect equilibration using results of high-resolution $N$-body simulations for the oligarchic growth stage. These effects were found to be small as long as a planetary embryo has a deep liquid magma ocean during its accretion. The effect due to partial involvement of the target's mantle in equilibration is small due to the low metal-silicate partition coefficient for W suggested from the low Hf/W ratio of the martian mantle. The effect due to partial involvement of the impactor's core is also small because a large fraction of the embryo mass is delivered from small planetesimals, which are likely to fully equilibrate in the deep magma ocean on the embryo. The accretion timescale of Mars estimated by the Hf-W chronology is shorter than that expected for the minimum mass solar nebula model as long as more than 10% of each impactor's core re-equilibrates with the martian mantle and the final stages of accretion are prolonged. This probably indicates that accretion of Mars proceeded rapidly due to solid and gas surface densities significantly larger than those for the minimum mass solar nebula or due to accretion of small fragments or pebbles.