Because there are "safety mechanisms" so that some of the errors in the bases are not translated into the proteins that are ultimately produced.
Your textbook probably has a chart of the "genetic code", please check it when reading this. If you don't have a copy with you, here it is:
http://library.thinkquest.org/C0123260/basic%20knowledge/images/basic%20knowledge/RNA/genetic%20code.jpg
The genetic code is the equivalence between sets of bases (codons) and aminoacids.
Note that there are only 20 aminoacids, but there are 64 possible combinations of 4 bases in 3 positions (i.e. codons). 3 of these possible combinations are "stop" codons, which leaves 61 codons for 20 aminoacids.
Thus, the genetic code is "degenerate" (= each aminoacid corresponds to more than one codon).
If you look at the different codons that code for the same aminoacid, you'll see that they usually differ by one base only (and it's usually the base occupying the third position). This means that even though one base may be substituted for another during replication, there are good chances that the resulting codon will still code for the same aminoacid.
As a result, when translation happens, there is no change in the protein primary structure.