The Upshot’s model shows good news for Jones, but it has also been swinging around a fair bit. A few minutes ago, they had him with a 90 percent chance of winning, but that probability is now back down to 67 percent.
I really admire what the Upshot folks are trying to do, but it’s also worth noting that building a live-updating model based on partial returns is a really hard problem. Among other things, its hard to calibrate your estimates of uncertainty based on empirical data, because (to put this in a nontechnical way) every state and every election is different. So I’d keep a watchful eye on the Upshot’s dials to get an overall sense of how things are trending, but I wouldn’t take the probabilities too literally, at least until later in the night.
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
no subject
Date: 2017-12-13 02:50 am (UTC)The Upshot’s model shows good news for Jones, but it has also been swinging around a fair bit. A few minutes ago, they had him with a 90 percent chance of winning, but that probability is now back down to 67 percent.
I really admire what the Upshot folks are trying to do, but it’s also worth noting that building a live-updating model based on partial returns is a really hard problem. Among other things, its hard to calibrate your estimates of uncertainty based on empirical data, because (to put this in a nontechnical way) every state and every election is different. So I’d keep a watchful eye on the Upshot’s dials to get an overall sense of how things are trending, but I wouldn’t take the probabilities too literally, at least until later in the night.
http://fivethirtyeight.com/live-blog/alabama-senate-election-results/?lpup=14604987#livepress-update-14604987