Every election is a puzzle, and each candidate is looking to arrange the pieces so that when it’s finished it’s a picture of their victory. The current race for District Supervisor between appointed incumbent Vallie Brown and two-time challenger Dean Preston is no different. To understand what benchmarks Brown or Preston need to do to win, its illustrative to look at the last D5 race – when Dean Preston challenged then-incumbent Supervisor London Breed in her reelection.

In that race, Breed won rather narrowly for an incumbent (by 1,784 votes out of more than 40,000 cast, or a little more than 4%). That victory looked like this:


Breed (in blue) was most successful in the NE and SSW parts of the district. Dean performed most strongly in the center of the district, especially along the East-West axis.

Neighborhood By Neighborhood

Neighborhood voting patterns play an important role understanding how these elections shake out. I’m sure many of you will squabble over how I drew the lines, but I carved the district into seven neighborhoods based on their voting patterns. It looks like this:


The neighborhoods are: Japantown (or Lower Pac Heights), the Western Addition, Hayes Valley, NOPA, the Haight, Cole Valley, and the Inner Sunset.

Not all these neighborhoods have the same number of voters, and not all of them vote alike.

In 2015, the election was dominated by four vote-rich neighborhoods: NOPA, the Haight, and the Western Addition, and Hayes Valley in the geographic core of the district.


While the three smaller neighborhoods, Japantown, Cole Valley, and the Inner Sunset combined make about 30% of the vote.

Screen Shot 2019-11-04 at 8.16.23 PM.png

In 2016, Haight, NOPA, and Hayes Valley comprised 53 percent of the District’s voters and if the election was only those three neighborhoods, London Breed would have lost her re-election by about 700 votes. Luckily for her, the voters in the Western Addition had other ideas. Her massive 1200 vote win in the Western Addition more than made up for her losses elsewhere. She also won fairly decisively in Japantown and Cole Valley, and more modestly in the Inner Sunset.


Rather depressingly for Dean Preston, London Breed’s margin and combination of neighborhoods means she would have won even if the Western Addition were not in the District. Her combined margin in Cole Valley, Japantown, and the Inner Sunset was also enough to overtake Dean’s combined margin in NOPA, the Haight, and Hayes Valley.

Screen Shot 2019-11-04 at 8.16.50 PM.png

What does this mean for Dean Preston this time? It means he either has to 1) make inroads into communities that he previously lost or (2) further galvanize the voters in his core areas (Haight, NOPA) so that they turn out in numbers that dwarf those in the other parts of the district.

It seems unlikely that Preston is going to win over voters in those communities he lost. In 2016, turnout was incredibly high (over 80%), and so just about everyone who could vote did. This time he’s running with much the same message as he had last time, and massive losses in Cole Valley, the Western Addition, and Japantown suggest that it didn’t really resonate with them.

However, he has a really big opportunity to galvanize the people who previously voted for him. Turnout in this election will almost certainly be significantly lower than in 2016 (maybe half as many voters will turnout). His advantage is that he already has a whole base of people who previously voted for him and he just needs to make sure they get to the polls.

In response, Vallie Brown needs to turnout those Western Addition, Cole Valley, and Japantown voters. Dean Preston is likely going to focus all his energy on the Haight and NOPA.

Early versus Late Deciders

Turning out voters is not as simple as it sounds. Not every voter votes at the same time. There are roughly four time opportunities to vote:

  • Early Vote By Mail – folks who get their ballot in early October and fill it out immediately (the first votes counted)
  • Late Vote By Mail – folks who get their ballot in early October and then hold it until nearly election day (Counted throughout the week after Election Day)
  • Election Day – folks who vote in person at a polling location (counted at the end of election day)
  • Election Day Provisionally – folks who had issues with registration but are eligible to vote (these are the last to be counted)

You can see the vote time breakdown by neighborhood below.


Unsurprisingly, NOPA has large numbers of voters in Early VBM, Election Day and Late VBM voters. The Haight has a large election day turnout, and the Western Addition has a large number of Early VBM voters.

Voters who vote at different times vote very differently. As you can see from the map below London Breed would have won handily if only the Vote By Mail (both early and late) voters were counted.


She won almost every precinct among Vote By Mail Voters. Likewise, Dean Preston would have beat her if only Election Day (including provisional) were counted, which you can see in the map below:


These rival contingents of voters is likely how this race will be decided. The full breakdown is below. Preston’s biggest margins came from Election Day voters in NOPA, Haight, and all the provisional voters in the district.


Can Vallie Brown get enough of London Breed’s Vote By Mail voters to submit their ballots? If so then she’s probably already won the election. Conversely, Dean Preston needs to galvanize Election Day turnout in his must win neighborhoods.

Turnout or Burnout?

Turnout will almost certainly be lower in this election that in was when London Breed ran for re-election, which was during a presidential election and more than 80 percent of San Franciscans voted. London Breed’s very competitive mayoral election in June 2018 only saw a 53 percent turnout, by comparison. I don’t know what turnout will be, given that the competitive races are for District 5 Supervisor and District Attorney, it could be below 40 percent.

Turnout doesn’t drop evenly across every neighborhood. By comparing turnout levels in 2016 (high) to 2018 (low) you can get a sense of what parts of the district are mostly likely to see large drop off. The good news for Vallie Brown is that Cole Valley consistently has the strongest turnout in the district; the good news for Dean Preston is that the Haight (which has many more voters) is right behind it. The bad news for Vallie Brown is that Japantown is likely to see the largest drop off; the bad news for Dean Preston is that NOPA is also likely to see substantial drop off in turnout.


A Vallie victory and a Dean defeat mostly like looks like substantial early voting and turnout from the Western Addition, Japantown, and Cole Valley. If Dean dominates, he’ll do so with strong election day turnout in the Haight and NOPA. In this lower turnout election, we’ll see if Vallie Brown’s decades of service as a neighborhood activist and legislative aide win her any converts; or if Dean Preston’s year of campaigning and tenant advocacy has inured himself to voters who previously voted for London Breed.

One last thing; as we head into election day – it’s helpful to look at what voters might be looking at. Here’s a screenshot of San Francisco-based google searches for both candidates over the last 30 days.


Vallie Brown is certainly getting more search activity, but that comes with a lot of caveats – not all these googlers live in District 5, and the vast majority of voters are likely not captured here. However, the Brown campaign can’t feel bad about this.

As you watch returns come in, remember that London Breed was beating Dean Preston by 19 percent in the first returns, which was only early Vote by Mail . That narrowed considerably  once Election Day votes started to be reported. Those are helpful baselines to keep in mind as you gaze into your crystal ball on Tuesday night.

Sign up for the Bay City Beacon weekly newsletter! It's a free way of getting the best of political gossip and cutting-edge culture in your inbox every Friday.

comments powered by Disqus