"any votes one candidate would and one candidate would not get in a contest with the republican would seem to be important" - that's a tautology, but it's also a different issue. You cannot take the fact that someone voted for A over B in the Democratic primary, to conclude that that voter would vote for the Republican over B in a general election. Statistically, the mere fact that someone voted in a Democratic primary means they're much more likely to vote for the Democratic nominee over the Republican, than someone who did not vote in either party's primary. And we don't know the preferences of those people.