The forecast summary is as follows:
And the detailed summary is as follows:
Also, tonight, I will have a look on a state by state basis the change in votes in the last 24 hours. This will become a regular daily feature.
NSW:
Yesterday: 69.29%
Today: 69.63%
Change in votes:
LNP +0.05%
ALP -0.05%
Change in election:
Statistically indifferent
(Note that my model is effectively taking Monte Carlo samples around a mean - expect an MOE of approx 3-4% @ the 95% confidence level)
VIC:
Yesterday: 62.18%
Today: 65.68%
Change in votes:
LNP +0.34%
ALP -0.14%
GRN: -0.09%
WIKI: -0.02%
SXP: -0.03%
All else: -0.01% or 0.00%
Change in election: No change - my model is still spitting out 100% to all six senators.
QLD:
Yesterday: 66.29%
Today: 66.38%
Change in votes:
+/-0.01%, or 0.00% for all
Change in election:
No change
WA:
Yesterday: 61.75%
Today: 61.75%
Change in votes:
Oddly, LNP down 0.01%, all else the same
Change in election:
None (all changes are easily within the MOE)
SA:
Yesterday: 71.36%
Today: 71.46%
Change in votes:
LNP +0.02%
XEN -0.02%
All else +/-0.01% or 0%
Change in election:
No change - all within MOE.
TAS:
Yesterday: 78.45%
Today: 79.46%
Change in votes:
LNP +0.06%
ALP +0.03%
PUP -0.06%
All else +/-0.01% or zero
Change in election:
The undecided seat seems likely to be won by PUP, LNP3 or SXP. The decline in PUP vote, in isolation, does negatively affect PUP's chance of winning the undecided seat by reducing the margin at the penultimate elimination to 0.27%.
Does your model for Tasmania take account of the BTL votes, and if so with what estimate? My view is that the Liberal Democrats are very exposed to below the line leakage as they receive nearly 20,000 votes from other micros on their way up to that likely three-way nexus point of Lib, PUP, Lib Dem.
ReplyDeleteTherefore I think that PUP's position compared to the Lib Dems is actually better for PUP than it appears.
Yes, the BTL % is much higher for micros than majors. See the 2010 situation in TAS.
Deletehttp://results.aec.gov.au/15508/Website/SenateUseOfGvtByGroup-15508-TAS.htm
No adjustment for BTL, at this stage. I have done a (now redundant) quantitative analysis for BTL in ACT which I am writing up now and will post very soon.
I have also put some thought into a more generic model for estimating BTL leakage. But busy week at work so may not get time to finish in the next week....
I am going to have a bit of a look at this for Tas if time permits tonight. Have heard from the AEC second-hand that the Tas BTL rate is 11%. Pleased it is still that high if so.
ReplyDeleteKevin, unless it is higher elsewhere, there has been no chatter of BTL levels that high - even at a single booth up north - that I have heard.
DeleteNow it may well be higher down south due to the higher Green / Palmer vote that produces far more BTL than the Lib/Lab vote.
SA results have now been updated. There is a Group L that did not have any results suddenly have 413 votes that have largely fed through to the Liberal Democrats. You wonder if something shonky is going on here.
ReplyDeleteCurious...
DeleteMy AEC data snapshots show that they had 403 votes on 8/9/13, so I don't see anything shonky.
Where did you see that they had none?
Group L's prefs do go to LDP before any of the other contenders.
I have a snapshot from the ABC Senate results page when 71.36% of the vote was counted (I think that was the final post on election night). I thought it could be significant as count 28 has only 2771 vote difference between FF and LDP even though LDP does not feature in your probabilities.
DeleteI will review my snapshots tonight. Also I will review incoming ldp preferences which may not be as favourable in sa as other states.
Delete