Sunday, October 19, 2014

With some tweaks to the model, it now special cases SNP and Green party contested seats to compensate for geographic concentration not reflected in national polling. This obviously reduces the chances for Labour to form a majority government. Again, this only matters if the election continues to be close enough that the small number of seats affected make a difference.

I've also run a short back-history of what predictions were for the last few months.

I've put the results graphics in a bar on the side, and will only be posting changes to the history graph in any updates.

Thursday, October 16, 2014

A little more tweaking to the internals, and adding in today's YouGov.




This also gives a rough prediction of the probability of either Conservative or Labour having a majority in the House of Commons, currently on 39% for Labour and 2% for Conservative.

Known problems with the model so far: SNP seat projections are problematic, and probably benefiting Labour a fair amount, but this is hard to fix without special casing SNP marginals. UKIP seats don't include any wins in ByElections, again this would require special casing to try and predict if they will retain them at the General Election. The Polynomal trend line can not be used in the election simulation for technical reasons, so Labour's vote share is probably a little low. This probably compensates for phantom gains from the SNP now, but probably not closer to the election. However, this is a tiny number of seats being predicted, and within variation of any close result.

Recent fixes to the model: Standard distribution now used for the election simulation, rather than a flat random variation. Reduced over-confidence that the trends are correct. Introduced a Consumer Confidence rating based limit to the weight of "return to mean" tightening of the result.

Tuesday, October 14, 2014

A few tweaks to the internals of the simulation (and this time ensuring all 2000 runs are put into the graph, not just 1000 of them) and here's another seat spread...


As you can see, the most common result puts Labour short of a majority. But again we see a second peak just after majority, that suggests a significant chance of a majority. The non-uniform nature of the distribution of seats does mean this kind of quirk can pop-up.



And here's a somewhat familiar trend graph from recent polling, which shows what rough area you could expect the vote shares to be in if trends continued. It's an important caveat that trends almost never continue without some change. It's somewhat difficult to identify if election polling follows a power or a polynomial trend, and some times a party's path towards election day can look like one or the other. So this graph has a Y split showing how the two different types of trend prediction diverge.

You'll note that the trend for UKIP looks pretty good for them, but I'm the model isn't showing them winning any seats. Which you might think is odd, because UKIP now have an MP from a by-election. There's a simple reason for this, the model doesn't know about by-elections. In my previous model, I attempted to use by-elections as a data point, this was a mistake. By-elections do not have a strong link to general-elections, and there's strong historical precedent for not applying swing calculations to a seat's by-election result, including many cases where a change due to by-election was over-turned at the following general election. As such, the seat predictions work based on how the seat was at the 2010 general election. I think this will better model a national general election, rather than using data from by-elections where an entire party and media circus is focused on a single seat.
I'm not committing to coming back to regular updates, but here's the output of a model I've put together for the 2015 elections. This runs 2000 individually simulated elections, using minor random variations on data from current polling. That way we can see what the possibility space for the election results look like.

Here we see that in the simulated election, both the Conservatives and Labour don't have their most likely positions in place to get a majority of seats. But we can notice that Labour have a second small peak closer to the majority needed, so may well be much better placed than they look, and it's not entirely ruled out that they could get a majority. However, a majority is still unlikely.

This result comes about because my model is very pessimistic for labour at the moment, as the underlying trends look weak after being disrupted by conference season.

Wednesday, April 20, 2011

Site Notice

I'm sorry that I'm going to have to announce again that the site will be moth-balled for an unknown amount of time, due to personal health issues cutting into the amount of time and effort I can spend on collating polling data.


My final thoughts on AV as we go into the campaign period, is that it's still too unclear to call it. Even after disappointing polling for the Yes campaign, they can't quite be counted out. It's now a matter of differential turnout, how the don't knows will split, and how many will not vote on the referendum even if they vote on their local elections. Such things are impossible to tell from the polling we have.

As for local and Westminster elections, it is now my belief that we are in a time of transition from one political landscape to a significantly altered new one. The polling data of recent times has had a slightly higher amount of 'outlyers' and somewhat pronounced differences due to methodology changes and variations. It may well be that we have back-slid in polling accuracy, due to changes in the electorate following the collapse of the Lib Dems that are not reflected in the representative shaping that polls use. Polling methodology can only ever really be tested by a new election, so it may be some time before we know if the poll firms have adjusted correctly to the new realities.

As such, it isn't that productive to keep producing a model, since the data being fed to it is suspect.

Best of wishes to all my readers.

Tuesday, April 12, 2011

AV, a case in Statistics.

A paper published in Significance, the statistics journal, collated election returns of the various government election systems. The aim is to show the various real-world properties of these election systems. It came to some unsurprising, and some surprising results.

Proportionality

We can identify how related the returns are to the proporition of the national votes. This identify if elections produce proportional results, something that smaller parties find is a problem with FPTP. The average deviation from the vote share shows an aproximate guage of this, with smaller numbers being better.

* UK, First Past the Post - 29.4
* Australia, Alternative Vote - 27.3
* Republic of Ireland, Single Transferable Vote - 12
* Netherlands, Open List Proportional Representation - 5.22

Naturarly, the Open List PR provides the closest representation. But despite claims that AV is no more proportional than FPTP it does show a slightly closer result. Some individual results under FPTP have been more proportional than AV. However, this only occurs for a third of the sample set. This suggests that AV is naturally more proportional than FPTP with some error overlap.

Probability of Coalition

Using probability statistics, it's possible to estimate the probability of certain events based on previous ones. Identifying the probability that random variations in future results, based on the known standard deviations from the data.

Here we see the probability, considering a rerun of the most recent election, of there being a coalition. For purposes of comparison, the Netherlands result was modelled as a race between the two largest parties and only one other minor party.


* AV Australia 2007 - 93% (Actual result - Single Party majority)
* FPTP UK 2010 - 87% (Actual result - Coalition)
* Ireland 2007 STV - 77% (Actual result - Coalition)
* Netherlands PRO - 48% (Actual result - Coalition)


The result here is a little muddier. If anything, it shows that the political situation is more likely to generate coalitions than the election system. The statistics suggest that AV has a higher probability of coalition, but the actual result in Australia was one that firmly established a single majority. They also show that FPTP can also produce a very high probability of coalition under the right circumstances.

Seat Changes

Another metric to compare elections is the amounts of seats than change hands in an election. This is somewhat a contentious statistic however, since the differing politics of the countries may produce more difference than the election system used. Numbers indicate more seats changed in the average election.

* FPTP - 17.3
* STV - 17.4
* AV - 19
* PRO - 21

As we can see, the surprising result is that AV does seem to make it slightly easier to unseat incumbents. Being only behind Party List based elections, while STV and FPTP are statistically tied for last.

Even assuming that the differences are more based on differing politics amongst the country; at the very least this indicates there is no incumbency advantage generated by AV, and it will not be harder to vote out unpopular governments.

Conclusion

The conclusions taken from this paper for comparison between FPTP and AV seem to be that AV does produce slightly more proportional results, that there is no clear link to coalitions being more likely, and that it may be easier to vote out unpopular governments under AV.

Tuesday, April 5, 2011

Disinformation being promoted for a NO Vote

There's a little bit of disinformation being passed around to support voting NO in the upcoming referendum on AV. It goes a little like this, "AV will result in the BNP getting more power. In France they have runoff votes, and the national front almost won the presidential election!" The problem here, is that it starts of with an outright lie, and supports it with a misleading truth.

In 2002 the Left-Wing and Left-Centrist vote in France had become fragmented over an astonishing 11 parties each proposing a candidate for the Presidency, eating deeply into the mainstream Left-Wing candidate's vote. At the same time, the mainstream Right-Wing candidate was hugely unpopular.

This resulted in an election where Jean-Marie Le Pen, leader of the National Front and regular fringe candidate for Presidency was able to get in contention for a First Past The Post win of the election. In actuality he came just shy of it and came second. But it does show the weakness of First Past The Post, in that it can elect a candidate the vast majority can not accept.

The French system however requires a 50%+1 Majority, so a runoff election was held. The French legislation however, only allows for the highest two to be on the second ballot. This lead to an overwhelmingly Left-Wing populace have to chose between an unpopular Jacques Chirac, and a despised Jean-Marie Le Pen. In the end they "held their noses" and voted for Chirac overwhelmingly, so in essence the runoff saved them from any chance of a fringe minority winning by 'default'. In fact, had it been run as FPTP, with all that entails, there's strong possibility that the confusion of tactical voting could have seen Le Pen elected 'by default'.

AV, or my preferred term "Instant Runoff Voting", preference voting allows runoff elections to be conducted on the same night. This allows for a 50%+1 theshold, and in essence works identically to a series of "ordinary" run off polls where all the voting had been conducted by sealed postal vote.

There is no way that this can benefit the BNP, unless by some magical transformation they are able to detoxify themselves enough to gain 50%+1. And this is actually why the BNP support the NO vote, and fear being sidelined entirely in votes conducted under AV.