From Richard Winger at Ballot Access News:

From the article cited:
With the 2014 cycle in the rear view mirror, political parties and candidates are already gearing up for 2016, with California Attorney General Kamala Harris’ announcement she would run for the state’s open U.S. Senate seat the big news of the week.
While the top-two primary system has effectively shut out the Libertarians in the Golden State, the country’s third largest political party has much reason for optimism heading into 2016 in races across the country.
Although many Libertarian U.S. Senate nominees fell short of their pre-election polling numbers (as conventional wisdom holds), the party nonetheless enjoyed its best cycle ever in races for the nation’s upper legislative chamber.
A Smart Politics analysis finds the Libertarian Party set records for the best showing in a U.S. Senate race in 10 of the 20 states in which it fielded a candidate in 2014.
Overall, voters had the option of voting for a Libertarian candidate in 20 of the 36 general and special elections held last November.

I can’t remember which publication it was in since I get so many of them. Also because my brain is full of swiss cheese holes 🙂
Maybe try contacting Carla.
I’d like to read that if it’s available. Sounds interesting.
Good point. Carla Howell recently did a statistical analysis of this with LP candidates and found we actually do better in close races between the establishment parties.
It was in a recent LP News, Liberty Pledge News or Staff Report to LNC or maybe more than one of these.
Conventional wisdom often holds that a close major-party race is bad for third-party vote totals.
I see something of the opposite effect going on here: when two major-party nominees are in a bitterly divided race (particularly if neither was that popular to begin with), then their willingness/need to go negative on each other tends to drive up negative perceptions of both of them. When both major-party nominees have low approval ratings, more people are willing to vote Libertarian, etc. You don’t see quite so much of that when one of the major-party nominees is popular, and is running a more positive campaign because they’re already expected to win in a landslide.
Of course, the wasted-vote thinking is also stronger in close races, so there are probably multiple factors at play here tending to counteract each other. But it’s just something that jumped out at me in this data set.