Ping (zestyping) wrote,

I met two extraordinary people today.

One: Bear McCreary, composer of the incredible music for Battlestar Galactica and now for several other television series. (Another member of the audience got to sit at the piano and play with Bear, an experience for which I would have given my left arm. (I would have needed to keep the right arm to play.)) Hearing him talk about his art made me spend some time thinking about doing music more seriously.

Two: Nicolaus Tideman, creator of the Tideman Ranked Pairs voting method, arguably the fairest ranked voting method ever invented (along with the Schulze method, which Kingman uses in its elections). Out of the blue, the president of the Public Choice Society invited me to their annual meeting to present this stuff I posted on the Web years ago. (It's crazy what ranting on the web can do!)

I'm presenting tomorrow afternoon, with a new interactive visualization.
(What's going on in this Flash doohickey requires some explanation, but feel free to play with it if you'd like to try to figure it out yourself, and post your questions or theories here.)
  • Post a new comment


    default userpic

    Your IP address will be recorded 

I like the idea of the applet, but some concrete examples would be helpful. Perhaps working with the fivethirtyeight guys to load in some recent electoral data?

The most confusing part for me was the moving of the median, with none of the voting systems selected, which moved the distribution left or right without changing shape. When we talk about election results, we're never using a distribution, but always percentages. I realize you report percentages at the top, but I couldn't read them off the normal distribution. It's not clear what the lesson of the moving median is supposed to be.

The clear part was the difference between each of the voting systems in terms of who would win and who would lose. But, having the median tied to the selection of the voting system was confusing, because it added an additional manipulation variable to the answer to the question Plurality vs Condorcet. Being able to answer that question directly is important.

Perhaps having all of the voting systems selected as a default?

I get a HTTP 403 Forbidden when I try to load the applet.
Fixed! Thanks.
Okay, so... I didn't understand why the median vote dot was in the plurality bar until I accidentally clicked in part of one of the other bars, intending to toggle it. (I was constantly clicking just barely to the right of the area that toggles a vote method.)

I still don't get what the triangular sliders do.
Check out the help text — there's a "Help" link at the top.
It occurs to me that the Hugo Awards have used a slight variant of IRV for some time. If you wanted some real-world examples, it might be possible for you to get your hands on aggregate raw ballot information (so many A > B > C, so many B > C > A, etc) and rerun the contests using other voting methods.


March 15 2010, 15:57:17 UTC 7 years ago

The problem there would be that people vote tactically based on the system in play. Ideally, we'd know everyone's sincere preferences and how they'd vote under each system.

One thought: it may be nice to be able to type in percentages as well as sliding the sliders, since some examples rely on specific race conditions that are hard to hit exactly with the sliders.

Tufte's new book on data and politics.