Democrats tell the Washington Post's Chris Cillizza that Scott Murphy will defeat Jim Tedisco by 210 votes, once all of the absentee votes are counted in the 20th District. This number hasn't been pulled out of thin air, but it's a dubious prediction.
Cillizza reports that the Democrats' projection is "based off of the county performances by Murphy and state Assemblyman Jim Tedisco (R) on election night". In other words, if Tedisco edged out Murphy 52 percent to 48 percent in a county with 1,000 absentee ballots, the Democrats project that Tedisco will get 520 absentee votes and Murphy will get 480.
If you run the numbers for each county this way--I confirmed the Democratic math using figures provided by the state board of elections--Murphy comes out 145 votes ahead among absentees (plus a 65*-vote lead equals a 210 vote victory). Here are some problems with this projection:
1. The counties are currently recanvassing--checking to make sure there weren't any typos when they wrote down vote totals. *So far, Murphy has "lost" 40 votes when the totals in one county were corrected. That means his current lead is down to just 25 votes. Further corrections could swing the vote in either direction. During the Minnesota statewide Senate contest recanvassing, Norm Coleman "lost" about 500 votes.
2. The Democratic projection is based solely on the approximately 6,000 votes that have already been returned--it does not account for the 4,000 ballots mailed out that may or may not be returned. As of this morning, only 15 percent of the 1,000 military ballots have been returned, and their inclusion could significantly alter any projections.
3. The Democrats' projection assumes that absentee voters behave exactly the same as voters in their respective counties. We simply don't know if that's true.
Republicans used to have a "big advantage in absentee ballots," says Michael Barone, author of The Almanac of American Politics, "but those days are gone." (Read Barone's take on last night's results here). In 2008, the Democrats had a superior "get out the vote" effort. But it's unclear if Democrats were able to continue their success in this special election or if their absentee efforts were hurt by the fact that their candidate was unknown at the start of an eight-week-long race that only became competitive during the last three weeks.
4. Another hair-splitting projection based on the turnout by party would give Tedisco the lead. As of November, the share of registered Republicans in the 20th District was 12 percentage points larger than the share of registered Democrats (42 percent to 30 percent). Among the returned absentee ballots, the Republicans' share is 13 points larger than the Democrats' share (48.5 percent to 35.6 percent). Since Republicans have done 1 point better than Democrats among those 6,000 votes, you could project Tedisco will make a net gain of 60 votes--and win by ... 35 votes.
But you wouldn't want to trust this projection. It doesn't account for a whole host of variables--just like the Democrats' projection.
"I think it's pretty much up in the air," says Barone. "I don't know how to price this toxic asset. If the government will provide 92 percent of the capital, I'll bet on it."