It will be weeks before the winner is known in the super-close race for attorney general. But thanks to figures provided by the Secretary of State's office, we can make a back-of-the-envelope guess as to how the remaining 2.3 million uncounted ballots will affect the outcome.
The results don't look good for Steve Cooley, who currently trails Kamala Harris by 12,000 votes. (Update at 11 p.m.: Cooley leads by 16,000 votes. See below.)
As we go county-by-county, let's assume that the uncounted ballots break for each candidate in the same proportion that the counted ones have. So, if there are 412,000 uncounted ballots in Los Angeles County, and Harris won L.A. by 53.7 to 39.2, then let's assume that she picks up 221,200 votes and Cooley picks up 161,500 votes. That would boost Harris' margin by 59,700 votes.
Similarly, in Orange County, Cooley beat Harris 60.9 to 30.7. There are 233,000 uncounted votes there, so that would give Cooley a boost of 70,400 votes.
You do that for every county and add it all up, and you arrive at the following result: Cooley picks up 1,938 votes.*
Now, that's not quite right, because three counties haven't reported their uncounted ballot totals to the state, and those counties voted for Cooley in aggregate. If you assign them estimated uncounted totals in proportion with their turnout, then Cooley picks up 4,322 votes.
Either way, Harris' holds on.
In its memo last night, the Cooley campaign noted that there are more uncounted ballots in counties that favored Cooley than there are in counties that voted for Harris. That's true.
But that doesn't go far enough. You also have to look at the margins of victory in each county.
For instance, Cooley won Sacramento County, where there are 120,000 outstanding ballots. Harris won Santa Clara County, where there are 108,000 uncounted votes. Advantage Cooley, right?
Wrong. Cooley won Sacramento by 2 points, while Harris won Santa Clara by 18. Figure it in proportion to their margins of victory, and Harris nets 17,000 votes from those two counties.
When you do that arithmetic for the whole state, Harris' lead appears safe.
Now, this analysis assumes that the uncounted votes will split for each candidate in the same proportion that the counted ones did.
But that assumption may be too favorable to Cooley.
Harris surged toward the end of the race, and won among votes cast on election day by about 2.5%. The uncounted votes consist of late vote-by-mail ballots (73.9%), provisional ballots (22.6%) and damaged ballots (3.5%). The Harris campaign argues that all of those ballots should reflect her late-breaking surge.
If you go with that assumption, then Harris picks up 58,349 votes.
But wait, you may be saying, aren't vote-by-mail ballots traditionally more conservative than votes cast at precincts? Wouldn't that give Cooley an advantage?
Well, Bay Area Republican David Harmer doesn't think so. Harmer, who trails the Democratic incumbent by just 400 votes, is suing in Contra Costa County to block the counting of late vote-by-mail ballots.
Harmer won that county, albeit narrowly, so you'd think he'd want more votes counted there rather than less. So what's up? Well, Democratic Rep. Jerry McNerney surged toward the end of the race. So the Harmer campaign is betting that late vote-by-mail ballots will reflect that surge.
Add it all up and it's hard to see how Steve Cooley wins. With all that said, obviously anything could happen between now and the certification deadline.
Update at 6:49 p.m. Hope springs eternal, as Cooley draws to within 8,000 votes statewide. The updated figure includes a big batch from L.A. County, and among those votes Harris only beat Cooley by a narrow 48-to-45 margin. The L.A. registar's office is making a priority of counting late absentees first, so it's possible those are trending much more conservative than the election day vote. If so, very good news for Cooley.
Update at 11 p.m. Some numbers now in from San Diego, Orange and Riverside put Cooley in the lead. Stay tuned.
*Corrected. The original post had a mistaken figure of 1,722 votes in Harris' favor. Other figures and analysis corrected.