« Statistical evidence is both powerful and limited: fraud in baseball and lotteries | Main | What did Grandma say? Nothing in life comes free. »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Merrick Usta

Nice proposal, would recommend using Bhattacharya distance (https://en.wikipedia.org/wiki/Bhattacharyya_distance) for distances within this probability simplex (https://en.wikipedia.org/wiki/Simplex) and finding same-distance divergences from 50-50.


Sorry, but no amount of math could tell you this — you don’t know how people would have voted if there had been only two candidates. This method is as much a fallacy as what the news sources do. The only way of knowing how people would have voted in a 2-way election is to hold a 2-way election. In the same way that the only way to know if someone is electable is to see if they’re elected.


Cris: Here's how I think about it. I am not trying to predict who would have voted for whom - which as you pointed out, is a counterfactual. What I'm doing is more descriptive statistics. Given the observed vote share distribution, what can we learn about the competitiveness of the contest? What is really happening is that I'm finding a principled way to order multidimensional vectors. It's easy to order a 2-dimensional race where the winner's share is all you need. Once you have multiple contestants, it's not clear how to order the results. The fact that there is no one perfect method does not mean that all methods are equally bad!


MU: Thank you very much for those suggestions. Will reach out once I have more to say.

The comments to this entry are closed.

Get new posts by email:
Kaiser Fung. Business analytics and data visualization expert. Author and Speaker.
Visit my website. Follow my Twitter. See my articles at Daily Beast, 538, HBR, Wired.

See my Youtube and Flickr.


  • only in Big Data
Numbers Rule Your World:
Amazon - Barnes&Noble

Amazon - Barnes&Noble

Junk Charts Blog

Link to junkcharts

Graphics design by Amanda Lee

Next Events

Jan: 10 NYPL Data Science Careers Talk, New York, NY

Past Events

Aug: 15 NYPL Analytics Resume Review Workshop, New York, NY

Apr: 2 Data Visualization Seminar, Pasadena, CA

Mar: 30 ASA DataFest, New York, NY

See more here

Principal Analytics Prep

Link to Principal Analytics Prep