jackman.stanford.edu/blog
• Bayesian Analysis for the Social Sciences Wiley; Amazon; errata as of 5/23/13

• 113th U.S. Senate
• ideal point estimates pdf csv 4/15/14
• scatterplot against 2012 Obama vote share pdf
• roll call object: RData
• 113th U.S. House
• ideal point estimates pdf csv 4/16/14
• scatterplot vs Obama vote share pdf svg
• roll call object: RData

## Friday August 9, 2013

Filed under: Australian Politics — jackman @ 6:47 pm

Nielsen’s 52-48 (n=1400) poll (png from GhostWhoVotes) bumped my Guardian Australia poll average a little, Labor moving from 48.2% TPP to 48.1% TPP, a new post-Rudd-return low.

Centrebet’s prices have moved this morning, Labor now out to 4.90 5.50, the Coalition in to 1.18 1.15, for an implied probability of a Labor win of less than 20% (19.4% 17.3%). Galaxy must be due to drop some numbers soon, perhaps in the Sunday News tabloids. We’ll also have Morgan’s mega-multi-mode melange and Essential’s two-week rolling average released early next week. The leader’s debate might also bring out some other polling.

I’m also publishing the GA poll average here on my own blog: see the growing list of Australian election stats and links appearing in the header of my blog, above.

The model-based average gets updated as we enter new polls into the database, along with a graph showing the trajectory of the model’s estimate of ALP 2PP voting intentions over the last 90 days:

## Wednesday August 7, 2013

Filed under: Australian Politics — jackman @ 5:21 pm

Labor’s price has come in from 1.75 (sportsbet) and 2.05 (Centrebet) to 1.20 and 1.25 respectively.

The implied probability of a Labor win in Forde is 68% and 71%, up from 47% and 42%. Meh.

Labor expected seat count: 61.

Forde’s really changed its status in my probability by margin plot.

Three probable Labor pickups in the top left quadrant of the graph: Forde, Melbourne (from the Greens), Brisbane.

The betting markets still tip 11 probable Labor loses, lower right quadrant.

## Tuesday August 6, 2013

Filed under: Australian Politics — jackman @ 6:20 pm

Labor’s out to 4.50 4.80 at Centrebet, the Coalition in to 1.20 1.18, for an implied ALP win probability of 21% 19.7%.

That just about takes us back to June 27, when Labor’s price came back in from beyond $5.00 on the back of Rudd’s return to the leadership. Just one week ago (noon July 31) Labor was at 33% (2.90 to 1.40). Labor’s post-Rudd peak at Centrebet was around 35% on July 16 (2.70 to 1.45). ### Newspoll the market mover (redux redux redux) ## Monday July 22, 2013 Filed under: Australian Politics — jackman @ 10:22 am 52-48 TPP to the Coalition overnight from Newspoll (The Australian). The betting markets haven’t wasted anytime responding. Indeed, Labor’s price could ease further through the day as news of the poll (and spin re poll) circulates. As a result, the 1st digit is a “2″, not a “3″, on the probability of a Labor win, implied by the prices; see the graphs (links at the header of the blog), at least for now. Thats still a long way up from pre-Rudd levels, but a long way short of 50% or better. Implied ALP win probability over the last week, at the 3 bookies I’m following closely (click for larger view): Newspoll’s influence on Australian political discourse and indicators like the betting markets is really something to behold. I’ve blogged on it numerous time over the years here (e.g., Nov 2011; May 2010; Nov 2007; Sept 2007). This topic is something I’ll try to expand on my piece this week for the Guardian Australia. Comments Off ### polls and markets, ups and downs ## Friday July 19, 2013 Filed under: Australian Politics — jackman @ 9:29 pm The last two weeks have produced a similar pattern in Centrebet’s betting market. (1) Good news for Labor in polls released early in the week sees some support for Labor. (2) The Coalition’s price drifts out to levels that attract some punter interest (e.g., 1.35 to 1.40), and the market comes back to the status quo by the end of the week. It is interesting to consider just what kind of signal from the polls the betting markets would want before swinging close to even-money. The market fluctuations of the last month week are in the attached graph. We’re getting close to an arbitrage opportunity between the 3.30 Labor price at Centrebet and the 1.42 for the Coalition at Tom Waterhouse (the prices as I type this). Just a 1.007 over-round there. ### watching the Liberal leadership betting markets ## Thursday July 18, 2013 Filed under: Australian Politics — jackman @ 5:45 pm Per my Guardian Australia column this week, I’m just keeping an eye on this one… I’ve added a current snapshot of the Centrebet and sportsbet Liberal leadership betting markets to the header of my blog. The sportsbet market has 12 contenders, including Peter Costello at$81, Wyatt Roy at \$1001, etc. The Centrebet market has just 6.

No time series graphs to report just yet. We’ll see how it goes.

## Monday July 8, 2013

Filed under: flight nerdery,general — jackman @ 4:36 pm

So the Asiana crash landing at SFO reminded me of the following weird coincidence.

One week ago, on July 1, I flew SFO-LHR on UAL 930. The pilot piped through ATC on Ch 9 (one of the nerdy things I like about flying United).

Just before we were given our takeoff clearance, SFO tower alerted an incoming flight to 28L that it was too low. The flight in question was UAL 1601 (IAH-SFO).

28L is the same runway that Asiana was trying to land on, with ILS out this summer; they’ve being doing some work on the runway there.

I went to the liveatc.net audio archive to download the KSFO tower comms from that day and time (mp3). Note that liveatc.net timestamps in Zulu time and stores 30 minute archives.

At about the 23:38 minute mark comes the low altitude alert from SFO tower to UAL 1601:

TWR: “Low altitude alert United 1601, check your altitude. San Francisco altimeter two niner eight three”.

A brief response follows, presumably from UAL 1601.

At the time I thought “wow, you don’t hear that often”. And I forgot all about it until the Asiana crash.

Pretty cool you can go recover that from liveatc.net.

## Sunday June 30, 2013

Filed under: Australian Politics — jackman @ 7:36 pm

Gillard was Prime Minister for 1100 days, from June 24 2010 until June 27 2013 (inclusive).

Rudd’s 1st term as PM, December 3 2007 until June 24 2010, was 935 days (inclusive). His 2nd term as PM started on June 27 2013.

December 8 2013 would be Rudd’s 1100th day as PM.

If there’s going to be a change of government, its likely to have happened prior to this date.

Just saying.

### Newspoll Labor 35 (+6), sportsbet 25% chance of winning

Filed under: Australian Politics — jackman @ 3:35 pm

Newspoll out this morning has Labor up 6 on both 1st prefs (29 to 35) and TPP (43 to 49).

The betting markets are reacting: sportsbet has Labor in to 3.75, the Coalition out to 1.25, an implied 25% probability of a Labor win. This is Labor’s best position in the betting markets since, well, the spill-that-went-nowhere in March, and prior to that, late January, early Feb this year (time series pdf).

## Friday June 28, 2013

Filed under: Australian Politics — jackman @ 10:44 pm

The graph accompanying my last post for The Guardian appeared as a low-res PNG. Here is the original, click on the in-line version for a standalone version.

I got some Twitter traffic as to why I’m modeling 1st preferences. The short answer is “first things first”. My strategy is (1) do my model-based, dynamic, poll-average of the 1st preferences, then (2) overlay assumptions about preference flows for 2PP estimates. That is, each pollster is allocating preferences to their 1st preference estimates; each pollster has only so much sample underlying their 1st preference estimates, so the 2PP estimates inherit that. My thought is to get a more precise estimate of the 1st preferences from my statistical model, *then* assign preferences. The aim is to wind up with a more precise 2PP estimate at the end of the process.

The easier thing to do is to simply average the pollster’s 2PP estimates (and I may yet wind up doing that). And it could well make very little difference.

An observation: if Green + Other etc is relatively constant (and this seems plausible, at least for short time frames) and the preference flow from those voters to Labor (conversely, the Coalition) is also rather steady (and everyone assumes it is, by the way), then Labor 2PP = Labor 1st preferences plus a term that is approximately constant.

Thus, variation in 2PP is almost all driven by variation in Labor 1st preferences, if what I’ve asserted in the previous paragraph is correct. The implication is that you don’t lose much by looking at 1st prefs vs 2PP, at least over the short term.

One the other hand, if Labor is shedding/gaining votes to/from its Left (the Greens) and not the Coalition, then no, Labor 1st prefs won’t track Labor 2PP particularly well, since the vast bulk of those Green votes flow back to Labor via 2nd preferences. I’ll have more to say on this once I’ve done some more work with the modeling and the data. But it is my intention to produce 2PP numbers one way or the other.

I’m back in the USA now, having spent June in Sydney and Brisbane. Our last night in Australia was in Brisbane with family and friends on Wednesday night, watching politics and State of Origin (and then politics again).

I tweeted out Thursday morning’s Courier Mail cover. We’re going to frame that one.

« Previous PageNext Page »