2009 Prognostication: Post 0 of 14

Entrant Statistics

After a dedicated media blitz including much browbeating, I managed to secure a field of 43 entrants for this year's quiz, up from 27 last year. I recruited two aunts, an uncle, and a cousin along with my immediate family. Grant also helped out to to form the second-largest family contingent, recruiting his two parents. Ekrem helped out by convincing his wife to join this year, followed by his co-worker, sister-in-law, and friend of sister-in-law. Welcome to all!

Of the 43 entrants:

26 are new and 17 have returned from the 2008 quiz. (The latter group will be competing for grand champion 2008-2009.)

10 are from Katie's sphere of influence, and 33 are from my sphere. (Who will have the best average score?)

21 are from Minnesota, and 22 are from elsewhere in the US.

Three people made patterns from their answers:

Best scream:
AAAAAAAAAAAAAA
Tiebreak: A
-Ekrem

Best choke:
CCCCCCCCCCCCCC
Tiebreak: C
-Greg

(In hexadecimal, A=10 and C=12.)

Tina wrote a little poem:

BAD CAB.
FAD DAB.
AD.

The rest of the answers seemed to be straight-up guesses.

Ignoring the hexadecimal outliers, the out-of-staters demonstrated their sadistic tendencies by predicting a significantly higher Minnesota snowfall (56.39 inches) than the Pollyanna Minnesotans themselves (53.06 inches). Honorary medals of the Marquis de Sade go to Northwestern Ben (72.4 inches), Southern Adrian (81.8 inches), and Northeastern Leanne (83.0 inches).

Predicted Score

As I wrote last year, the "mass intelligence" of the 43 entrants can be used to predict everyone's final score. By this method, the number of times a particular answer is chosen is correlated with the probability that answer will be correct. (If A is chosen 10% of the time and B is chosen 25%, then those who choose A get a 0.10 predicted score for that question, and those who choose B get 0.25.)

Here is the current leaderboard, adding predicted scores for all 14 questions:

1. Katie (4.51)
2. Liz (4.47)
3. Kevin (4.33)
4. Pete (4.28)
5. Jeremiah (4.21)
6. Ted (4.21)
7. Jan (4.07)
8. Janette (4.05)
9. Gloria (4.05)
10. Michael (4.02)

Yes, my wife is leading the pack, while I barely eked into the top 10. Read on to hear why this may be contraindicative of success.

Unusual Correlation

One bizarre trend present in the data is that Predicted Score correlates quite significantly with Entry Order:
This correlation was so shocking to me that I thought I must have made a mistake in my spreadsheet to explain it. After careful checking and rechecking, though, my spreadsheet is correct.

I could find no credible explanation for this odd correlation. Covariates that might impact Predicted Score (participating in the 2008 quiz vs. not, Michael's friends vs. Katie's friends, MIT alum vs. not, etc.) did not correlate with entry number.

Last year, I led the field in Predicted Score at the start of the year, and yet ended alone in last place. This suggests to me that Predicted Score might be less of an indication of what is likely to take place and is more a prediction of who is most swayed by the amount of "Michael bias" inherent in the wording of the questions. It might make sense to me that the fastest entries to the quiz were either (a) the least researched or (b) the most susceptible to question-writing bias. I guess we'll find out as the year unfolds.

Answers Explained

Finally, I know I'm running long, but I found it very interesting that Grant included analyses for his answers. Others indicated that they had spent some time researching their answers (notably Matthew, who wrote "I have spent way too much time researching this. Begone!" as he sent in his answers), but Grant included the most thoughts in his entry email. Here they are in full, with my comments in brackets:

1 [Freedom]. B. Minus one. Improvements have been made in many countries, notably Pakistan, Liberia, and Ghana, but there have been downturns in Ethiopia, Venezuela, Angola, and the entire former Soviet Union with the exception of Ukraine and the Baltics. The global economic turmoil has contributed to net losses for freedom around the world in 2008, (and I would expect things to get worse in 2009, although I'm not filling out my Prognostication Quiz entry for 2010 yet).

2 [Beauty]. C. Dance. I have no idea. There seem to be more dance shows on TV than there used to be. [Interesting to think that TV shows might give dance more subliminal "legitimacy" or some such on the part of the judges.]

3 [Football]. A. the Giants. This is a wild guess; this is one of the closest playoffs in a while, and I think almost everyone still in (not Arizona or San Diego) has a legitimate shot this year.

4 [Film]. E. Slumdog Millionaire in a landslide. I should probably go see it.

5 [Mushing]. B. Uh. I can see Russia from my house. [All of the named options are, in fact, from Alaska. If seeing Russia is an advantage, it's actually C. Martin Buser who is from Big Lake, a little closer to Russia than B. Jeff King's hometown of Denali.]

6 [Writing]. C. 2008 was full of so many financial stories that WSJ must have written some of them up right. [A good thought!]

7 [Hockey]. D. Detroit. Never bet against the Red Wings.

8 [Hoops]. C. Lakers. Hopefully if I keep betting against the Celtics they'll keep winning.

9 [Golf]. E. Nothing. Betfair has this at almost 50% probability, which is a lot higher than any specific major. [I admit that I picked E, too, though I'm surprised that the probability this high on Betfair, given how impressive his game seemed in 2008, despite the pain from the injury. Recovering from surgery can be very tricky, though.]

10 [Peace]. C. Who the heck knows what those crazy Swedes will do.

11 [Baseball]. F. It's tough to bet against either the Yankees or Red Sox, but offering me two different series lengths more than compensates for the NL's incompetence. [Despite the precise 2-1 edge over the past 25 years? You think the days of AL dominance is over? As I type this, I'm now noticing the "25 years" in the question is actually 27 years, but still!]

12 [Storms]. B. Global warming is a big long-term trend. People trying to predict individual hurricane seasons on the basis of their beliefs about global warming are deluding themselves. [It's a big long-term trend, but a "tipping point" is possible in such complex systems. Think of a finger gradually adding pressure to a lightswitch. Even if you add pressure very gradually over the course of an hour, the onset of light happens very suddenly, occurring completely in just one of the 3600 seconds. The long-term average number of hurricanes is 6, but the recent average is clearly higher than that, averaging 7.5 over the past 12 years. Check out this page, which includes this graph with the number of named storms (some percentage of which turn into hurricanes) over the last 100 years:
I think the trend is clear.]

13 [Stocks]. D. Economy won't recover all that much, but the market is a leading indicator and should start to turn around mid-year and end modestly up on 2009.

14 [Snow]. B. Not much snow yet this year, unlike here.

Tiebreaker: 51.5 inches. Yeah, I know that's outside the range I guessed, but I don't see any rules against putting my tiebreaker in contradiction to my guess on 14, and I think it's the right strategic move. [An interesting ploy, picking the mean snowfall for the tiebreak but not for question 14. We shall see if it works!]