Many a mickle maks a muckle - Long-shot Prediction Game

> Recent Entries
> Archive
> Friends
> Profile

June 30th, 2008


Previous Entry Add to Memories Share Next Entry
10:02 pm - Long-shot Prediction Game
New Year is probably my favourite time of the year to read blogs. I love it when people do year-in-summary reviews and build up a string of sets of results from year to year so that you can compare things easily. I also enjoy Iain Weaver's annual predictions; Iain makes interesting predictions of relatively unlikely events, and then - with admirable integrity - goes and marks the success or failure of the previous year's predictions. He marks himself harshly; even though his raw score is not high, considering how long many of his shots are, I consider his record to be remarkably good. Even many of his misses are near, interesting and insightful. The only other similar prognosticator I can compare Iain to is Robert X. Cringely, who has predicted annually in a similar way.

I love reading interesting predictions. Anyone can make safe short-odds claims, but I'd take a handful of unusual and correct pot-shots over stating the bleeding obvious any day. (Or perhaps what is the bleeding obvious for some well-informed commentators may be an unusual pot-shot to myopic ol' me.) Accordingly, while we've played prediction games in the past - Iain's even run a couple, though rather intricate ones - there is often a tendency for the games to reward conservatism and selection of relatively likely options. There are things you can do with scoring systems, but they don't make the underlying prediction targets more interesting. I would be more entertained by a prediction game dealing with plausible long-shots - ideally ones where players make the plausible long-shot predictions, but a cunning mechanic would be required to ensure, er, length of shot.

This is not quite that game. This is an attempt to see whether my Friends list can produce a prediction market worth a damn by trying to form a consensus opinion of just how unlikely some moderately unlikely events are.

Below, please find twelve moderately unlikely events, all of which would be slightly surprising, hopefully none of which are completely implausible. Unless otherwise stated, I am interested in whether they might happen after July 31st, 2008 and before February 1st, 2009, excluding these boundary dates, considering local time zones as appropriate. Hopefully these are moderately international, or at least skewed towards the predicted audience.

Finance:
* Oil to trade at $200/barrel or higher.
(Any of Brent crude, New York light sweet crude or West Texas Intermediate crude, which are trading at around $143.)
* BBC Global 30 stock market index to drop below 4750.
(It is currently valued at 5230 and was based at 5000 in September 2004.)
Politics:
* John McCain to win at least 336 Electoral College seats in the 2008 US Presidential election.
(George Bush won 286 Electoral College seats in the 2004 US election.)
* Gordon Brown to stop being Prime Minister of the UK, Labour Party leader or both.
Sport:
* The USA to win at least one more gold medal than any other single country at the Beijing Olympics.
(At the 2004 Olympic Games, the USA won four more gold medals than second-placed China.)
* The New England Patriots to win Super Bowl XLIII, expected to take place on or around February 1, 2009.
(They won their first 18 matches in the previous NFL campaign, only to lose Superbowl XLII. See also the so-called Super Bowl curse.)
Internet:
* Microsoft to make a takeover bid for (or merger bid with) Yahoo that Yahoo's board of directors accept or recommend acceptance.
* Firefox usage to reach 33% or higher.
(Source: any month or any quarter reported by W3 Counter global stats, TheCounter global stats or Net Applications stats. See recent stats for comparison.)
Entertainment:
* The Harry Potter and the Half-Blood Prince movie, i.e. Harry Potter 6, to gross at least $285,000,000 at the US box office by (and including) the weekend concluding Sunday February 1st, 2009.
(Counterpart figures for the previous five Harry Potter movies.)
* One or more of Prince William, Prince Harry, Princess Beatrice and Princess Eugenie to marry publicly.
Other:
* USA and/or Israel to execute an overt air strike against Iran.
* There will be no more than two major hurricanes (category 3+) in the 2008 Atlantic Hurricane season.
(Current predictions are for two to five.)

My questions: how many of these will occur? Which are most likely? Which are least likely?

Formally, let's incentivise you to predict rationally through a scoring system whose incentives are hopefully not too skewed. Please predict how many of these twelve events will happen; earn 10 points for being exactly correct, 9 points for being incorrect by one in either direction, 8 points for being incorrect by two in either direction and so on. Please select which four of the twelve are most likely to occur; putting my Bayesian spectacles on and conflating probability with results from single trials, three bonus points will be awarded for each of your "most likely to occur" selections that do occur. Please select which four of the twelve are least likely to occur; two points will be deducted for each of your "least likely to occur" selections that do occur.

Also, shout up if it is not clear how any of these will be assessed. Should you spot a reasonable way in which any of these possibilities could be described as being neither true nor false, or where it is not clear how the claims of truth or falsehood will be authenticated, please comment sooner rather than later. You can safely assume that I am prepared to interpret these moderately liberally - or, at least, that I am unlikely to be swayed by pedantry. On the other hand, if what you consider an important distinction that I may not have taken into account is what you fear I consider pedantry, I will appreciate it if people quibble early rather than waiting to see if it'll make a difference in the scoring.

Predictions are visible to all and can be changed up until the point at which I close the poll. The closing date and time will depend on my real-life demands, but I intend it to be before I go to bed after my day shift on Tuesday 8th July.

Poll #1213908 Long-shot Prediction Game poll, part one
This poll is closed.
Open to: All, detailed results viewable to: All, participants: 18

How many of the twelve events will occur?

View Answers
Mean: 4.18 Median: 4 Std. Dev 1.62
0 (0.0%)
0
0 (0.0%)
1
0 (0.0%)
2
1 (5.9%)
3
7 (41.2%)
4
4 (23.5%)
5
1 (5.9%)
6
2 (11.8%)
7
1 (5.9%)
8
1 (5.9%)
9
0 (0.0%)
10
0 (0.0%)
11
0 (0.0%)
12
0 (0.0%)

Which events are the four most likely to occur?

View Answers
Oil to trade at $200/barrel or higher.
0 (0.0%)
BBC Global 30 stock market index to drop below 4750.
0 (0.0%)
John McCain to win at least 336 Electoral College seats.
0 (0.0%)
Gordon Brown to stop being Prime Minister of the UK, Labour Party leader or both.
0 (0.0%)
USA to win most gold medals in Beijing.
0 (0.0%)
New England Patriots to win Super Bowl XLIII.
0 (0.0%)
Microsoft to takeover Yahoo.
0 (0.0%)
Firefox usage to reach 33% or higher.
0 (0.0%)
HP 6 to gross at least $285 million at the US box office.
0 (0.0%)
Prince William, Prince Harry, Princess Beatrice and/or Princess Eugenie to marry publicly.
0 (0.0%)
USA and/or Israel to execute an overt air strike against Iran.
0 (0.0%)
No more than 2 major hurricanes in 2008 Atlantic Hurricane season.
0 (0.0%)

Which events are the four least likely to occur?

View Answers
Oil to trade at $200/barrel or higher.
0 (0.0%)
BBC Global 30 stock market index to drop below 4750.
0 (0.0%)
John McCain to win at least 336 Electoral College seats.
1 (5.6%)
Gordon Brown to stop being Prime Minister of the UK, Labour Party leader or both.
0 (0.0%)
USA to win most gold medals in Beijing.
0 (0.0%)
New England Patriots to win Super Bowl XLIII.
0 (0.0%)
Microsoft to takeover Yahoo.
0 (0.0%)
Firefox usage to reach 33% or higher.
0 (0.0%)
HP 6 to gross at least $285 million at the US box office.
0 (0.0%)
Prince William, Prince Harry, Princess Beatrice and/or Princess Eugenie to marry.
0 (0.0%)
USA and/or Israel to execute an overt air strike against Iran.
0 (0.0%)
No more than 2 major hurricanes in 2008 Atlantic Hurricane season.
0 (0.0%)


(11 comments | Leave a comment)

Comments:


[User Picture]
From:sbisson
Date:June 30th, 2008 09:20 pm (UTC)
(Link)
You should probably look out Mark Anderson's predictions too (check out his SNS mailing list and web site for these). He runs the annual Future In Review conference, which is one of the highlights of our year...
[User Picture]
From:theoclarke
Date:June 30th, 2008 09:55 pm (UTC)

UK futurologists

(Link)
Ian Pearson is pretty amazing. His erstwhile colleague Graham Whitehead is also fascinating.
From:songmonk
Date:June 30th, 2008 10:08 pm (UTC)
(Link)
I need to read your posts as well as the other posts more carefully later. I only had a chance for a quick glance just now.

I would say (pending further rereading to make sure I'm not misinterpreting) that I don't find a bunch of longshot predictions that interesting. That's like throwing a bunch of darts and hoping one of them hits the bullseye. If you throw enough darts, something will hit.

I'd find it more interesting to attach probabilities to those predictions and seeing how it matches with the view of others. (Which is kind of what you're doing here.) If you put something at 100-1 and everyone else thinks it's 10,000-1, that's interesting if your assessment is right, even if your horse doesn't come in. (Of course, for a lot of these, it's difficult to say "you were right" if your horse doesn't come in. :-) )

But bores the hell out of me on sports shows is when the pundits make their prediction for champions at the end of the season. It's what the dynamics of the situation demands, so I don't blame them (no one wants to hear, "Well, the Patriots are the favorites to win the Super Bowl, but they only have a 30% chance"), but I find it utterly uninteresting.
[User Picture]
From:titanic_days
Date:June 30th, 2008 10:35 pm (UTC)
(Link)
Chris, disregard my last choice on the most likely section. Sorry. If this skews your results then direct all subsequent flames to my inbox. Looks an interesting project though.
[User Picture]
From:undyingking
Date:July 1st, 2008 08:15 am (UTC)
(Link)
You can revise your selection by clicking the link that says "Poll #1213908" above, and then "Fill out poll".
[User Picture]
From:hedwig_snowy
Date:July 1st, 2008 01:50 am (UTC)
(Link)
Only 4: Oil, <2 Hurricanes, Gold Medals, Microsoft.

1) McCain will not get over 230 electoral votes.
2) The NE Patriots won't even make it to the Super Bowl
3) Sorry, but Firefox - which I use - will not make that bi a jump.
4) Even with movie cost inflation, the HP movie will be doing well to break $250 (remember the books are over and I know that people will want something, anything HP but I don't think that will be enough to push it past POA in terms of box office)
5) Have no idea about Gordon Brown. Would depend on 1) Whether there is an Iranian air strike - which I wouldn't put past Bush, Cheney, and the Israelis to get done because they fear the next administration would...do precisely what they did with N Korea - and how things go in Afhganistan. Oh, and there are always the human concerns like health, etc... But is there some idea that they might call an election before then?
6) Princes are having too much fun.
7) US stock exchange might fall below 10,000. ~13% drop. If FTSE follows suit, it could be close...watch oil. If it actually hits $200 (can you say Strait of Hormuz?) then it may be worse...

And there are my predictions. Absolutely bankable. Bet the house. I wouldn't, but that's just me....
[User Picture]
From:mrwarfrog
Date:July 1st, 2008 07:25 am (UTC)
(Link)
I would be incredibly surprised should any member of the Royal Family be allowed to marry secretly!
[User Picture]
From:undyingking
Date:July 1st, 2008 08:18 am (UTC)
(Link)
Excellent, I think you've structured this really well. I won't fill it out just yet as am busy, but I'll return shortly.

Pain that LJ's poll reporting thing doesn't allow the easy assessment of current prediction aggregates.
<*feels guilt that he hasn't yet got round to writing the long-promised tool that would do so*>
From:(Anonymous)
Date:July 1st, 2008 10:15 am (UTC)

From sir_quirky_k who can't log in anymore, annoyingly!

(Link)
I predict: oil, stock market, Patriots (going wildly against the grain here), strike on Iran. Least likely: McCain, Olympics, Firefox, hurricanes.

Of the others: Brown will do anything to hold on until 2010, as proven by not calling a 2007 election, but the prospect of internal revolt pushes it off my 'least likely' list; I think MS will find some other way of expanding their presence in search; I nearly predicted Potter, but am thinking that in a recession it won't quite make it; I don't care to predict anything with the Royal Family.

Might run something along the lines of Weaver's in 2009...
[User Picture]
From:jiggery_pokery
Date:July 2nd, 2008 11:53 am (UTC)
(Link)
A highly welcome off-LJ player adds his votes as follows:

Prediction: 2 (two).

Most likely:
* Oil over $200
* Stock market drop
* Potter VI
* Iran

Least likely:
* McCain
* Microsoft-Yahoo
* Royal wedding
* Hurricanes

Notes and queries:

Finance:
* Oil to trade at $200/barrel or higher.
* BBC Global 30 stock market index to drop below 4750.

It may be useful to declare if you'll be using end-of-day figures or intra-day trading for each of these.


Should intra-day trading figures be published, I would use them, but I wouldn't expect the intra-day and end-of-day answers to be different. (Though I guess there was a gap of a week or two between the first $100/barrel oil trade and the first $100/barrel end-of-day. I regarded the announcement of the first trade as credible and would do so again for a $200/barrel oil trade.)

* The USA to win at least one more gold medal than any other single country at the Beijing Olympics.

How will post-event disqualification play here?


I suppose it would be appropriate to settle based upon consensus medal table standings as of game conclusion. Should there be a gold medal in abeyance and should that gold medal make a difference, we will regard the game as representing the quantum superposition of a universe in which the gold was awarded and a universe in which the gold was not awarded. Not very elegant, I grant you.

* One or more of Prince William, Prince Harry, Princess Beatrice and Princess Eugenie to marry publicly.

How are we defining "publicly"? A wedding to which television cameras are invited? The photographers of a weekly supermarket magazine? That is announced as an engagement in that day's Court Circular?


Yes, yes and no respectively, but note there are many other ways in which a wedding could be considered public (and, indeed, very few ways in which a marriage could take place and not be public, though you can't rule anything out and I'm not going to believe the National Enquirer or the Daily Sport if vanishingly few other people do). I consider your notion that a marriage starts when a proposal is accepted, and thus the engagement may be announced, to be highly romantic but not in keeping with generally accepted usage.
From:enfarcer
Date:July 5th, 2008 08:16 pm (UTC)
(Link)
Senator McCain hasn't a ham sandwich's chance in Rosie O'Donnell's fridge of even winning the election, let alone garnering 336 electoral votes. Americans have wised up considerably over the last 3½ years. Or at least I hope they have. How I hope they have...

> Go to Top
LiveJournal.com