Saturday, December 31, 2016

2016 Site Review

Another year done for a site that is now just over four years old.  When I started this site I intended it just as an interim site while I explored other options, but in spite of some very minor frustrations (like difficulty in stopping spammers from even submitting comments) I've seen no compelling reason to move.

This year was a federal election year and the activity pattern for the year looks like this (the units are sessions per week):


The federal election and its post-count events basically swamp everything else.  Senate reform performance reviews maintained a high level of reader interest through to mid-August and after that the rest of the year was quieter, with little spikes for the ACT and NT elections. The massive federal election spike meant that this year had about 128% more traffic than last year, and about 60% more than the previous busiest year (2014).



For the third year running I released exactly 86 articles.  Of these 18 were federal poll roundups. 10 pieces covered Senate reform and nine were federal election post-count threads.  As usual not quite everything I started got finished.  One of the pieces I didn't finish is a non-pseph work in progress that may well appear sometime, and I don't want to reveal what it is in advance (I was going to write it on Christmas Day but the humid weather in Launceston put paid to anything involving actual thinking).  The remaining incomplete drafts were titled:

Enough Of All This Alan Lichtman Nonsense!
Women's Chess Champs Hijab Controversy
Margin of Error Myths (someday this will probably get written; there is so much nonsense written about MoEs that I ran out of time and moved on before I could list it all)
Hobart City Council: Suzy Cooper Recount (recount was held before I could finish the article, which was to correctly predict the Greens would win, though fence-sit on which of two Greens contenders would do so)
Hot Air in Tasmanian Bushfire Politics

(The Lichtman piece was to concern a professor who was being widely hailed for predicting that Donald Trump would win the US election, and for predicting every other POTUS election for many years with his magic "keys to the White House" system.  In fact, after Gary Johnson's vote dropped below 5%, Lichtman's formula only correctly predicted that Hillary Clinton would win the popular vote, and said nothing about Trump winning the election.  Moreover in 2000, Lichtman had first predicted Bush would win, then switched to Gore, and only later retrofitted by claiming his model predicted the popular vote.)

Top of the pops

The following were the ten most popular articles for the year by number of unique readers:

1. How To Best Use Your Vote In The New Senate System

This FAQ-style resource piece giving advice on how to vote effectively in the new Senate system was my second most-viewed article ever.

2. List Of Senate How-To-Vote Card Preferences

A surprise second place (and nearly topped the list too).  This piece was just a list of known Senate how-to-vote cards issued by parties, which I posted mainly for those who would find them useful in trying to predict results.  It turned out that Senate HTV card follow rates were way lower than anyone expected.

3. 2016 Tasmanian Senate: A Model of What Might Occur

The second of three Tasmanian Senate articles, this one presented the results of actual scrutineering sampling I had used to try to forecast the Tasmanian Senate outcome.  The count was complicated by candidate effects in both major party tickets, and by confusing AEC presentation of information.  The article correctly predicted all 12 winners although there was a very close shave between the Greens and One Nation for the last seat.  Easily the most commented article in site history with 87 comments.  This is among the work I am most pleased with this year because an empirical model based on (seriously) hard work outperformed - as it generally should - a number of theory-based predictions doing the rounds in the media.

4. 2016 House of Reps Postcount: Vanilla Reps Seats

Post-count thread for the several Coalition-vs-Labor two-party preferred seats that were clearly "classic" (Melbourne Ports and the non-classic seats had their own threads).

5. House of Reps Postcount: Melbourne Ports

In a post-count that was very badly covered by the media, Michael Danby (ALP) defeated Owen Guest (Lib) after staying ahead of Steph Hodgins-May (Green) by just 953 votes.  The possibility that Hodgins-May might catch Danby on preferences from Green-leaning micro-parties, and the impact of post-count votes on that chance, was generally ignored.  It mostly didn't look likely Hodgins-May would quite do it and in the end it didn't happen.  One to watch next time!

6.2016 Senate Postcount: A Very Long Way Still To Go

A general overview of early stages of Senate post-counting that mainly acted as a bucket of cold water for those trying to predict close micro-party fights in a system they hadn't sampled before all the votes were in.

7. 2016 Late Postcount And Expected Recount: Herbert

Covered the recount in the closest House of Representatives seat, where the LNP's Ewen Jones looked like he would probably just hang on until two significant counting errors in Cathy O'Toole's favour were found and sent the seat to the wire.  O'Toole ended up winning by 37 votes to cut the government's majority to one, and rumours of a court challenge came to nought.

8. Postcount: 2016 Tasmanian Senate

Part 1 of the Tasmanian Senate postcount series (see #3 above).

9. Tasmania Senate 2016: Prospects and Guide

Both a candidate guide for the 2016 Senate (doing some sometimes lighthearted due diligence on a number of the more eccentric parties and candidates) and an attempt to predict the outcome.  In the last respect the article did somewhat better than its 2013 equivalent, but underestimated the state swing against the Liberals (because the polls did - severely) and completely ignored One Nation who nearly won a seat (because they were not doing anything!)

10. Tasmania Senate Button Press And Analysis

The final installment of the Tasmanian Senate series, this dealt with the final results and analysis of the count, which was to be useful in seeing what to expect in other states (though in many respects Tasmania was an extreme case.)

The top ten list is the same if counted by pageviews, though the order changes slightly.

Some other stats

The ten biggest days of the year in terms of pageviews were mostly around the election day and immediate aftermath - in order July 5, 3, 6, 4, 1, 2, 27, 7, June 30, July 8.  The outlier July 27 was Tasmanian Senate button press day.  The lull on the day itself was because I was off at the Mercury site on the night.

The most popular pieces hailing (mainly) from a previous year were the old edition of the Field Guidethe bio page, the 2013 best and worst pollsters review, A Same Sex Marriage Plebiscite Is A Terrible Idea and Why Preferred Prime Minister/Premier Scores Are Rubbish.

The most clicked-on tags were pseph, Senate reform, Xenophon, silly greens, Tasmania, Senate, Greens, Legislative Council, how-to-vote cards and Jacqui Lambie.

The top ten visiting countries (as defined by Google Analytics) were Australia, the USA, UK, NZ, Canada (+5), Germany (+1), France (+2) Japan (-2), Singapore (re-entry) and Brazil (new entry). 133 countries dropped in.  169 countries have now visited and with Sudan now joining the party, the most populous nations never to see this site are the Democratic Republic of Congo, North Korea and one of my favourite places I've never been to, Madagascar.  The least populated units to score hits have been Norfolk Island, Anguilla and Turks & Caicos Islands.

The most visiting units by population (excluding very small countries) appear to have been Australia, NZ, Singapore, Iceland, Fiji, Ireland, Solomon Islands, UK, Hong Kong and Norway.  A single chess-politics article written while in Norway continues to mess up the stats here. 

The most visiting cities were Melbourne, Sydney, Hobart, Brisbane, Canberra, Adelaide, Perth, Launceston, Devonport and London. 

The most common search terms for the year - my name excluded - involved how to vote cards and preferences (mostly for various seats).  For the number of hits this year the number of eccentric search terms winding up here was remarkably small; here are some feeble examples:

affectve line for ganing vots

alderman mao.ding (He wasn't elected - missed out by one vote in 2014 - KB)

australia's best poll forecasting couple

bonham what is this did you not putting any

"glistens and stinks" mackrel

mhas support mp sex

so liberal iam an independent

what does lapoinya look like now 

Of those searching for my name only 54 (3.6%) misspelled it, in a total of 25 different ways (many of them apparent mistypings), with "Kevon" (11) most common.   "Bonhapp" was the most impressive mis-spelling for the year.  

The biggest hit sources for the year were Google (+1), Twitter (+1), Crikey blogs (mainly Pollbludger) (-2), Facebook (+5), Tally Room (-1), Tasmanian Times (-1), The Guardian (new entry), Mark The Ballot (-2), Bing (+1) and Reddit (new entry).

Thanks again for all the support from you the readers through the year, especially those of you who have donated $$$ to reward my efforts.  It's been a crazily busy year for me with nine weeks on remote fieldwork and two overseas but I still managed to post as much stuff as I normally do.

Orders of the year

In 2017 there will be a WA state election.  There's a pretty high chance of a Queensland state election and the Tasmanian Premier has recently failed to rule out an early Tasmanian state election (it's "due" in early 2018).  There will certainly be Tasmanian Legislative Council elections (with the ability of the Hodgman government to get major bills through the Upper House hanging by a few threads in what could be three very interesting contests). 

But before that - hopefully within the next week or so - will come the Ehrlich Awards dedicated to those with the most predictive egg on their face after the topsy-turvy year that was 2016.  What a rich crop of wallies have lined up for this gong this time around ...

No comments:

Post a Comment