On election night in San Francisco, Ed Lee ended up with 31% of the first-choice votes. John Avalos had 19%, and Dennis Herrera had 11%.  The remaining 39% of the votes were split up among the rest of em.  Here are the preliminary results.

Avalos has a shot to beat Lee in the IRV tabulation, but just barely. The only way he wins is if the second and third choice votes of all the other candidates transfer at a rate of at least 3-to-1 Avalos over Lee. I think this is unlikely because there were so many other qualified candidates in the race, and because Ed Lee is generally considered a competent Mayor. Very few people hate him, he got a lot of second- and third-choice votes.

In the Oakland Mayor’s race last year, the results were surprisingly similar on Election Night. Don Perata ended the first-choice tabulation with a commanding 11-point lead over Jean Quan. But when third-place finisher Rebecca Kaplan was eliminated and her votes transferred over, an unprecedented two-thirds of them went with Quan, sending Quan across the finish line. This happened because Quan and Kaplan were the only credible candidates in the race other than Perata, and because they ran a concerted IRV strategy, convincing the voters to rank each other second. Moreover, Perata was a polarizing figure – it was an “Anybody but Perata” strategy.

The same doesn’t hold true here. I don’t think Avalos and Herrera ran an IRV strategy together, and there were loads of other credible candidates in the race who will eat up each other’s second- and third-choice votes. Ed Lee himself will get a lot of those second- and third-choice votes. Because he’s no Don Perata.

AND – a lot of ballots are going to be exhausted. Which will be good for Lee. As ballots get exhausted –  because voters bullet-voted (i.e. voting for only one candidate), or because all three of their choices are eliminated – the total vote count goes down. As the total vote count goes down, the percentage of the total vote count for a candidate goes up, inching Lee up towards the goal of obtaining 50% of the total vote count. And that’s where he wins.

And so a ballot that includes votes for only candidates OTHER THAN Avalos or Lee… is ultimately a vote for Lee.

There were a lot of candidates in the race, and so a high percentage of ballots are going to be exhausted early. Because I didn’t see many IRV strategies out there, I’m guessing a lot of voters bullet-voted.  It’s all going to come down to one question: Who did the losing candidates’ voters choose as their second- and third-choice votes, if anyone?

The people who picked Ting, Dufty, Rees, Hall, and Alioto-Pier are more likely to give Lee their second choice votes based on their shared politics. That’s about 20,000 votes, not counting exhausted ballots.

The Baum and Herrera voters are more likely to have picked Avalos as their second or third, and that’s about 17,000 votes, not counting exhausted ballots.

Adachi voters? Forget about it. His votes will transfer to other Asian candidates, some conservative candidates, some to progressives. He’s the wild card. And that’s 9000 votes.

The Chiu and Yee voters are going to be squirrely.  Yee and Chiu worked hard to defeat Lee, and their voters are more progressive than Lee’s. Yet Chinese-American voters tend to be loyal to Asian candidates, and these seconds and thirds are more likely to go with Lee over Avalos. But it’s hard to tell where the Chiu and Yee votes came from just yet. I voted for Chiu, and my second went to Avalos. I also got 1400 hits on my voter guide, so I bet a lot of my friends did the same.

Together, Yee and Chiu had 23,000 votes. Based on the numbers above, these votes have to transfer to Avalos at an extraordinarily high rate (4-to-1?) to put Avalos over the top. And I just don’t see that happening.