Unexpectedly Intriguing!
April 16, 2009

The plus-minus statistic in hockey is one of the neatest statistics out there. Simply put, the plus-minus statistic considers how valuable a player is over a period of time by finding the net difference in points scored by their team and their opponents while they're on the ice. A player who consistently maintains a positive plus-minus statistic is one whose presence is one who will likely be greatly valued, while a player who regularly runs a negative plus-minus statistic is one who will be considered to be less valuable.

To be fair though, what counts most is a player's relative plus-minus statistic compared to those who might play in their place. In the case of when a star offensive or defensive player needs a break, the challenge for their coaches is to put players on the ice who can minimize any potential point loss during that time while the other team might have their best players going, assuming the coach doesn't have a deep bench full of star players.

Today though, we're going to apply the plus-minus statistic to something very different. We're going to use it to see how valuable we are when it comes to making predictions! Better still, we'll challenge our competition to put up their own numbers! We figure we'll find out two things: who's got the guts to be honest about it, and who's got game when it comes to forecasting!

Here's how we'll do it. We've arbitrarily gone back to January 1, 2008 and run through all of our posts where we've made a real prediction, as opposed to simply offering an observation. If we got it right, we score it as +1. If we got it wrong, we score it as -1. And if the outcome for the prediction in question is uncertain, for whatever reason, we score it as a zero (+0).

The cool thing about this approach is that if we're just making coin toss random predictions, over time, our plus-minus will drift toward zero. Alternatively, if we're better at this prediction game than not, then our plus-minus will drift higher and if we're worse, our plus-minus will fall.

How did we do? We found we made 25 predictions since 1 January 2008 related mostly to the economy or the stock market. The table below tells the rest of the story....

Political Calculations' Plus-Minus for Predictions Since 1 January 2008
Date Prediction Outcome +/- Score
2 January 2008 In December 2008, the S&P 500 would reach 1610. We dumped our prediction just two weeks after making it, recognizing that a significant disruptive event was overtaking the stock market. -1
16 January 2008 A recession would not be declared to have begun in 2007. Technically correct. The NBER later declared that December 2007 marked the final month in which the U.S. economy expanded in the previous business cycle before a recession began. In common word usage though, most people take the NBER's declaration and cite this month as the being the starting point for recession. +1
23 January 2008 The stock market will begin bottoming in September-October 2008. We're scoring this prediction as zero, or "mixed". Partly this is due to not understanding well enough how dividend futures worked at the time, which ultimately caused us to put the peak of market distress ahead of schedule (that happened in January 2009.) However, our error in interpreting the dividend future data wasn't that far off target at this point - the S&P 500 began did indeed begin its bottoming process in September-October 2008! 0
5 February 2008 The level of distress in the stock market will peak in July-August 2008. Hugely wrong, and directly due to our rookie error in reading dividend futures data! -1
6 February 2008 Distress will peak in the stock market in March 2008. Even more wrong. But not yet as wrong as we could get! -1
7 February 2008 The peak of distress in the stock market will occur between February and March 2009. As wrong as we could get (and got!) We finally figured out what we weren't getting nine days after this post! -1
27 March 2008 First quarter GDP for 2008 will come in at \$11,754.8 billion (real GDP - adjusted for inflation in chained 2000 US dollars). Real GDP in 2008Q1 came in at \$11,646 billion (chained 2000 USD). We were over target by 0.9%. +1
31 March 2009 The New York Times' weekday circulation would drop below 1 million in the next 12-18 months. Right on target. A year later, we believe the weekday circulation for the Grey Lady is occasionally dipping below the million mark, which we expect to be a very regular occurrence another six months from now. At this point, the New York Times' leadership could pull out the stops to make this prediction wrong, but only by sacrificing its revenue. +1
5 April 2008 U.S. unemployment levels for the year will peak in November-December 2008. The timing was right, but the composition of the job losses wasn't quite what we expected. Up through October 2008, most of the jobs lost throughout 2008 were the minimum wage jobs we anticipated (these also account for the majority of jobs lost for the year.) However, the summer oil shock spread the job losses into the automotive sector, which began massive layoffs toward the end of the year. +1
16 July 2008 "Until the erosion in dividends ends, we're afraid that the S&P 500 and other major stock indices will not be going up anytime soon. We would anticipate that the carnage in the market will continue until the powerhouse financials dragging the indices down get serious about building up their balance sheets and pull the trigger on dividend cuts. Then, and only then, will investors regain confidence in these companies." Dead on target. This is, in fact, exactly what's finally being played out after the first quarter of 2009, and is why stock prices are finally rising off their lows. +1
22 September 2008 We noted a unique correlation between when an uptick occurred in our preferred tool for determining the probability of a recession and when the stock market began plunging in mid-September 2008. We note that the next uptick in this measure would coincide with the week of 13 November 2008 through 20 November 2008. In retrospect, a creepily accurate prediction. +1
29 September 2008 We backnoted how our "modified limo" technique would have predicted the level of Real GDP for 2008Q2 (which we can't count as a prediction since we hadn't posted anything before-hand), and we predicted that 2008Q3's Real GDP would be \$11,781.4 billion (chained 2000 USD). 2008Q3's Real GDP came in at 11,712.4 billion, which means we overshot the mark by 0.6%. +1
1 October 2008 We anticipate that January 2009 would see the bulk of companies revising their forward business outlooks downward. We also anticipated that the driver for this outcome would be the spreading of problems in the financial sector to the rest of the U.S. economy. +1
22 October 2008 After rebuilding our confidence offline through most the year, we revisit the dividend futures data that had been the main source of our biggest failures in the prediction business and peg January 2009, give or take a month, as being the peak month for distress in the stock market, projecting the bottom for the stock market not long behind. Now with the advantage of 20-20 hindsight, we were very much on target with this call. The level of distress in the stock market, as measured by our price-dividend growth ratio, spiked for January 2009 at -111 (absolute value +111). +1
30 October 2008 We reviewed the biggest single financial transaction in presidential candidate Barack Obama's life and extrapolate that an Obama presidency might be characterized by a significant lack of fiscal discipline in pursuit of grandiose ambitions. We wish we were wrong and that President Obama's legacy would not be one defined by massive amounts of wasteful spending with little to show other than lasting, unwanted obligations. We dare President Obama to make us wrong. +1
5 November 2009 We project the bottom for the stock market "in or around January 2009." We're going to count this one as incorrect, even though you'll see that we're not far off…. -1
6 November 2009 Remember that "uptick" in our measure of recession probability that coincided with a massive point loss in the stock market? We go officially on record with our prediction that carnage would ensue in the stock market on 20 November 2008. We called a bottom in the stock market. There's no ifs, ands, or buts about it. +1
20 November 2009 In the middle of the day, nearly at the peak for the stock market for the day, we doubled-down on our prediction from 6 November 2009. In our update to the post later that day, we also hint that Round 3 for the stock market might begin on or around 21 January 2009. It's true. We achieve modern-day Garzarelli status! (and if there's any doubt that we did, you can see our pre-updated time-stamped post via our RSS feed.) Although unlike Garzarelli, we may not be a one-hit wonder.... +1
5 February 2009 We call *the* market bottom in a three-part prediction based based on past market performance, finding that the bottom would occur: 1. A 24.2% chance of happening by the end of January 2009 2. A 59.6% chance of happening by the end of February 2009 3. A 79.8% chance of happening by the end of March 2009 In a next-day update, we pick February 2009 as the most likely time for the market to bottom. We missed the absolute market bottom by nine calendar days (or six trading days.) We're scoring this as a -2, as we've presented four separate predictions, three wrong and one right, which nets out as -2. The only prediction that was right? The one that gave an 80% likelihood that we'd see the bottom by the end of March 2009 following the peak of market distress. -2
13 February 2009 We predict that significant changes in the U.S. income tax are in the works given the kinds of questions that members of the U.S. Senate (or their staffs) are asking us through Google. We'll see. +0
17 February 2009 We present a prediction that GM is heading toward bankruptcy that we had originally put forward back on 2 October 2008 at Econbrowser. Looks pretty likely. +0
3 March 2009 We put the bottom for the S&P 500 index value occurring within a range between 655 and 680. Well, what do you know? We do what Garzarelli never could: between our timing prediction from 5 February 2009 and this prediction for calling the range in values that would define it, we called a market bottom for the SECOND time! Calling it three days early doesn't hurt our cause either! +1
16 March 2009 With the bottom of the market behind us, we predict that the S&P 500 will move into a range between 815 and 840. The gold standard in being a good analyst isn't just making a bottom call, but also calling the turn in a market. You always have those people who make just bull calls or bear calls, and like broken clocks, they get to be right at least once if you give them enough time. We called the turn in the market seven days in advance of the market hitting our target. +1
2 April 2009 We up our forecast for the S&P 500 to hit values between 860 through 890. Although we didn't specify it on our own site, this prediction applies through the end of June 2009. Right now, there's been a very small bit of erosion in the dividend futures we use to create our forecasts, so if we made this prediction today, our methods would have us put the S&P 500 between 855 and 885 from now through June 2009. For now though, we're going to count this prediction as neutral. +0
14 April 2009 We predict, after stocks fell by a healthy margin on 14 April 2009, that they would close up by a healthy margin on 15 April 2009. There are two ways you can look at this particular prediction. Either we had a 50% chance of being right and we got lucky, or there's something to the kind of analysis that we do and luck wasn't much involved at all. If you think it's all random, at what point might you place your bets on the guy who keeps winning the coin toss? +1

Out of these 25 predictions, we find three that have yet to have their outcome determined. Totaling up the numbers of the remaining 22, we find we have a plus-minus of +8. Since we discovered what we were doing wrong in reading dividend futures back in mid-February 2008 for our stock market predictions, our plus-minus is +11.

How do you think Garzarelli might rank? Or Roubini? Or Krugman? Or any of those talking heads who make a living making predictions? Will they put up their own numbers? Ever?

Labels: , ,

is good for you

Welcome to the blogosphere's toolchest! Here, unlike other blogs dedicated to analyzing current events, we create easy-to-use, simple tools to do the math related to them so you can get in on the action too! If you would like to learn more about these tools, or if you would like to contribute ideas to develop for this blog, please e-mail us at:

ironman at politicalcalculations.com

Recent Posts

Applications

This year, we'll be experimenting with a number of apps to bring more of a current events focus to Political Calculations - we're test driving the app(s) below!

Most Popular Posts
Quick Index

Site Data

Visitors since December 6, 2004:

#### JavaScript

The tools on this site are built using JavaScript. If you would like to learn more, one of the best free resources on the web is available at W3Schools.com.