Posts by Michael Homer
Last ←Newer Page 1 2 3 4 5 Older→ First
-
That last link is now broken, but the media release about it is here.
-
Being picky, I'm not sure what this is trying to say:
I arrived at the proposal to have the matter referred to the Court of Appeal arose because of that Court’s powers in respect of election petitions respecting the allocation of list seats.
It looks like two sentences mashed together there and I don't know which you meant. There is also a "candidate who’s nomination" that should be "whose".
-
Hard News: #JohnDotBanks and all, in reply to BenWilson,
ACT getting Epsom gives National an effective bonus seat, even if they got no further seats. Why? Because ACT’s seat does not count against the National total seats. If National won the seat, they would get one less list MP (so their numbers would be unaffected)
Not really true. ACT only got the party vote total it did because it had a good chance of getting in through Epsom. If they hadn't thrown them the seat and only 2,000 people switched their vote to National instead of ACT they would have had their 18th list seat anyway.
It's true that it's hard to pick that in advance, but given how Sainte-Laguë works they would be pretty likely to get it. More votes means more seats in general, and only in odd edge cases do you get more seats than you "should" by splitting votes across parties (and not more than one in any case). It's only if you assume that the party vote tallies are the same that National loses a seat.
-
Legal Beagle: MMP Review #1: The Party…, in reply to Sam Hill,
That's supplementary member. It's not in scope for the review.
-
Does he have a defamation case now?
-
OnPoint: Dear Labour Caucus, in reply to BenWilson,
At the moment, I think what can be gleaned from the long range predictions is the sentiments of the likely traders. That was all I commented on – why do they feel Shearer is so likely, and yet at the same time a Labour government unlikely?
People have said this a lot, but just to be clear - PM.2014.LABOUR is over 50¢ right now, and has been for a couple of days, as it was for almost a week before that. It dropped under briefly, caused by a few small trades, and that made the news. It's currently at 53c, but moves up and down pretty arbitrarily within a ten-cent-or-so band and trying to make news with that is absurd. If anything, somebody spent a few dollars to bring the price down just before ipredict's weekly blog post.
Gleaning the sentiments of traders is more interesting though. I don't know what you should or could take out of it all.
-
OnPoint: Dear Labour Caucus, in reply to BenWilson,
Arbitrage it then. As people get more interested (as the event approaches), the volumes go up, and the arbitrage opportunities get smaller and smaller.
I did. I will at this point make a (small) profit regardless of who the leader is, but that doesn't mean I have any confidence in the predictions it makes of either that or the 2014 election as being based on anything meaningful at all, yet.
But you don’t need to worry about the sampling method in iPredict.
I don't. The sampling method is fine, but the sample, right now, is inadequate. That's all.
I think we actually agree here, from this post - it will probably converge on something fairly accurate in time, but on these questions it's far from that point. On others it has been good, especially when the event gets closer. To counterbalance that, though, at the 2008 election there was a "David Benson-Pope to run as a candidate other than for the Labour Party" contract that launched at 80¢ and stayed thereabouts up until the end. That was not an accurate prediction of anything, it was David Farrar's wishful thinking. It's not an oracle.
-
OnPoint: Dear Labour Caucus, in reply to BenWilson,
Except that they have been extremely accurate? I don’t see how it’s possible to be sure there has or hasn’t been insider trading, that’s an incredibly difficult thing to prosecute.
They haven't been extremely accurate (yet), because the events these contracts cover are still far from being determined. Some of the contracts (I mean PM.2014.X and the like) are averaging ten shares traded a day and jump around markedly on single trades, because seemingly nobody is really interested in them. Because of all that market manipulation would be both cheap and easy, so to the extent that there are "insiders" on them it's probably to do that. It's working, too; somehow "X drops on ipredict" is now a legitimate story, and you can spend a few dollars to make it happen for your camp.
More to the point though, I don't trust them as predictions here because the various contracts don't match up or trade in line with each other. When a Labour PM is less likely than Shearer or Cunliffe personally the market simply isn't working. The shorter-term predictions have been pretty good.
On insider trading, they explicitly encourage you to do that. There's nothing to prosecute.
-
The other thing to note on ipredict is that the trading volume of most of the stocks in question is pretty low (very very low for some of them), and they've all been priced out of whack with each other for a long while. I think the real secret is that there hasn't been any insider trading on most of them at all, and I wouldn't take them as having much predictive value because of that on top of all the other reasons.
-
Legal Beagle: Election '11 -…, in reply to Geoff Pritchard,
Open lists sound great in principle, but I’ve yet to come across a concrete proposal that seems workable.
I am a fan of the Dutch system - print the whole list, one vote for an individual who can be elected in their own right, and distribute the excess votes down the list for proportionality. If people want to take the list the party chose they can just tick the name at the top and be done, or they can pick someone particular who represents them best. They've used that system for quite a while now and seem to find it workable.