> > When I use linear probing (probe decrement always equals 1) I get 672 collisions >in my table > (which > > has table size 2003)
> Some questions to ponder on a rainy Sunday afternoon: > > - Are you really sure you are counting the collisions correctly? > It's only a colission if two words A and B hash to the same table address > (I know you know that). But you also shouldn't count the word A more than once. > (So, are you counting the insertion of "the" only once, disregarding the >search/insertion for "the" > all the other times?) I just tested this. The number of collisions to expect with a table of about 2003, linear probing, and a probe decrement of 1, is indeed that high.