Mapping Strategy

Predicting and Dealing With the Sun and Space Environment

Because scientific understanding of solar phenomena is far from comprehensive, the wide range of predictions and methods used to inform this visually delightful graphic isn't all that surprising.

I wonder though, whether a prediction market on the same topic would come up looking prescient... or foolish? This attempt from 2007 looks like the answer to the question, "What if they declared a prediction market and nobody came?" Maybe I missed something, but I don't think so.

The almost unprecedented state of solar phenomena at the moment, the increasing dependency of the world on various kinds of satellites and the wide range of predictions would suggest it might be worth giving prediction markets another try. (I can find no other prediction markets on this, either active or defunct.)

Interestingly, this hefty pdf document, from 1996, pegged the market for space environment predictions at $100M and expected it to double in size by 1999 (see p. 29, aka 189 according to the document footers). Anyone have any idea how big it is now? Even 1% of that would be fine.  :)

This scary piece on the same subject  might lead some to wisely consider scenario thinking for business continuity and resilience as useful complments. Sometimes the unprecedented simply... happens, at which point the question quickly turns from prediction to how to cope and thrive in the messy, uncharted aftermath.

Prediction Markets and the Election

Four years ago, I grabbed some numbers off Intrade in the days leading up to the U.S. presidential election, ran some smoothing functions on them, and discovered a strong correlation with the vote in each of the fifty states (R-squared value of 0.77, and every state 'called' perfectly hours, even days before the actual voting).

This year several things were different, including no incumbent candidate, larger margin -- and expected margin -- of victory, new dimensions around polling and voting blocks that seemed hard to predict, tumultuous external events, etc.). Nonetheless, the state-by-state markets on Intrade coughed up predictions that were almost as good as 2004. Using my extremely unscientific sampling method, the state-by-state markets produced an R-squared value of 0.72 for data taken around Monday Noon).

In 2004, the markets were (as I recall) run as mirror images of one another. If you were going long on Bush, you were, by definition shorting Kerry, buying (at least in theory) from someone who'd lost confidence in or perhaps wanted to take profits and reduce losses in the Kerry contract at the market price (even if, perhaps, said trader still hadn't lost confidence in the candidate himself -- an entirely separate question we're not even going to approach).

This year, by contrast, Intrade ran separate markets for each state (plus DC) in: McCain (i.e., to win a particular state), Obama, and 'Other', or 153 markets in all. That's a lot on which to maintain liquidity, but on the swing states (e.g., Missouri, Indiana, North Carolina, Virginia, etc.) the markets appeared to be quite deep indeed.

Now, here''s the interesting part...

As the chart below depicts, the McCain and Obama markets as of Monday Noon (12:45PM EST, actually) -- roughly 18 hours before the polls opened -- were both trading at what ended up being an R-squared value of 0.72* (the square of the correlation coefficient, shaking out poor correlations from good ones) relative to the near-final results I took off CNN yesterday (Thursday). *Actually, they differed by two one-thousandths -- hardly significant.

As of Tuesday (Election Day) morning however, the correlation on the Obama market had dropped significantly (to 0.59). It dropped further through Noon on Tuesday then bottomed out. By contrast, the correlation of the McCain markets to the final results (again, state-by-state, not overall) maintained a relatively high correlation throughout the day.

Did Obama market participants lose interest? Were some of them trading on a non-rational basis for other (perhaps partisan, utilitarian or sentimental) reasons? I'm at a loss to explain the difference, since nothing of any note happened to the 'Other' markets (all except Wyoming). They traded -- if they trade at all -- with paper-think liquidity at values of $0.10 on a contract that paid $100.00. (The contract for a Wyoming third party victory was trading at $1.00)

Butterflies for Dummies

Check out this succinct review in last Sunday's Boston Globe of an important but oft-misunderstood concept critical to any forecasting or planning endeavor [emphasis and link added]:

The butterfly effect is a deceptively simple insight extracted from a complex modern field. As a low-profile assistant professor in MIT's department of meteorology in 1961, [the late Edward] Lorenz created an early computer program to simulate weather. One day he changed one of a dozen numbers representing atmospheric conditions, from .506127 to .506. That tiny alteration utterly transformed his long-term forecast, a point Lorenz amplified in his 1972 paper, "Predictability: Does the Flap of a Butterfly's Wings in Brazil Set Off a Tornado in Texas?"

In the paper, Lorenz claimed the large effects of tiny atmospheric events pose both a practical problem, by limiting long-term weather forecasts, and a philosophical one, by preventing us from isolating specific causes of later conditions... It is extremely hard to calculate such things with certainty... Realistically, we can't know. "It's impossible for humans to measure everything infinitely accurately," says Robert Devaney, a mathematics professor at Boston University. "And if you're off at all, the behavior of the solution could be completely off." When small imprecisions matter greatly, the world is radically unpredictable.

Moreover, Lorenz also discovered stricter limits on our knowledge, proving that even models of physical systems with a few precisely known variables, like a heated gas swirling in a box, can produce endlessly unpredictable and non-repeating effects.

"Lorenz went beyond the butterfly," says Kerry Emanuel, a professor in the department of earth, atmospheric, and planetary sciences at MIT. "To say that certain systems are not predictable, no matter how precise you make the initial conditions, is a profound statement." Instead of a vision of science in which any prediction is possible, as long as we have enough information, Lorenz's work suggested that our ability to analyze and predict the workings of the world is inherently limited.

What few articles on this subject touch upon, but I find endlessly fascinating, is how this undeniable, well-grounded scientific insight interacts with human nature -- specifically our ingrained need/desire to feel knowledgeable and in control. (I use the term 'we' in both the individual and corporate sense here.)

It has been my experience that few individuals and very few (if any) organizations ever err on the side of humility in this regard. Perceptual failures are seldom in the direction of assuming less than is actually predictable. More often, the response is something like this:

yes I know about the butterfly effect, chaos theory and all that... but it's kinda academic... and you don't know my boss... you see, I/he/she/they/we need to predict XYZ anyway...
because my/our future depends on it... so just give us a number... give us your best guess about the probability... please?

People fall back on a prediction/probability paradigm either because it's what they know (or have the planning tool-set to address) or because, even knowing that approach is flawed, they find it more comfortable... and comforting (or politically expedient in a particular organizational culture) to pretend that what is fundamentally uncertain is perhaps predictable after all. The chance of rain is 62.5%...

Predicting Armageddon

Sadly, In light of this, this seems undervalued [emphasis added].

"If Iran continues its nuclear weapons programme, we will attack it," said [Deputy Prime Minister] Shaul Mofaz... "Other options are disappearing. The sanctions are not effective. There will be no alternative but to attack Iran in order to stop the Iranian nuclear programme."

One thing notable about this graphic is how little reaction there has been to this news. To my reading (and I've followed this closely the last several years) it marks a significant step up in rhetoric by the top echelons of the Israeli government. The second thing I find remarkable is how the market has trended steadily upwards over the past four months or so.

This is but one of many 'quarterly' prediction markets Intrade has been running on an airstrike on Iran. All of them, obviously (at least so far) have rewarded the skeptics, however most have trended downward as the market close approached. This one isn't doing that.

I'm of two minds on how much stock to put in this market. (Sorry, pun intended).

On the one hand, military missions are, or at least ought to be, closely guarded secrets. And this would be no ordinary military mission. It is truly existential. Sure a few thousand people might be privvy to some kind of knowledge or direct insight into what was going on, but I doubt any of them would risk being shot in order to make a little extra on the side. So, on that basis, this has many of the characteristics of silly/failed markets in things like papal elections and supreme court nominations.

On the other hand, this is not as surprising as a terrorist attack (a dumb thing in which to make highly specific markets IMHO, even though there are many related indicators that might be useful).

Most nations don't want to go to war. (Iran may be an exception, but we're talking here about an attack on them, not by them). Most nations 'do the (diplomatic) dance'. They do relatively rote, predictable things in the lead-up to any conflict. And those things tend to follow patterns that are not random.

They may not be easily predictable, but there is information out there, held by people not constrained by an oath of secrecy, that is meaningful in discerning whether this will go down or not. Sadly, if I were to bet on it (and I don't -- at least not on real money markets) I would have to say that this market is way undervalued.

UPDATE (July 1, 2008): Almost a month after I first made this post, the trend continues (see below). Despite Israeli moves that make such a strike seem more likely (whether for real or for bluff we can't know), the market hasn't had a very high 'beta' (volatility). This despite (or perhaps because of) significantly higher trading volumes in June. A good summary of recent events around this can be found here.

Overt Airstrike on Iran on or before Dec. 31, 2008

Obstacles to Prediction Market Adoption

Harvard Business School Associate Professor Andrew McAfee asks, in a recent blog post, why prediction markets haven't been taken up more readily in the corporate world. He writes [links added]:

After writing the [Google prediction markets] case, teaching it a few times, and spending some time understanding the mechanics and utility of prediction markets, I share the puzzlement articulated by James Surowiecki in his book The Wisdom of Crowds:

". . . the most mystifying thing about [prediction] markets is how little interest corporate America has shown in them... companies have remained, for the most part, indifferent to this source of potentially excellent information, and have been surprisingly unwilling to improve their decision making by tapping into the collective wisdom of their employees."

Why is this? It’s not because the technology is hard to acquire... what is the real stumbling block? Is it that companies don’t really want the most accurate information about future events to come out and be widely known?

Sean Silverstone Silverthorne, the editor of HBS Working Knowledge expands on the question in a post last week, outlining what he sees as "three initial roadblocks":

• Unfamiliar with the Concept -- Most managers, especially the veterans, grew up hearing not to trust crowds, to avoid mob rule. Instead, they were encouraged to value expertise and be ready to pay consultants. It’s a major brain-bender to suddenly believe crowds can be smarter than experts.
• Internal Friction -- Corporate prediction markets rankle folks like product managers when company employees start betting on a bad outcome for their product.
  
  
• Implementation Issues -- Until recently, there wasn’t a whole lot of best practices around how to run efficient prediction markets. Results can be seriously skewed by having too few participants, asking the wrong type of questions, or by running a market too long or too short a time.

From my experience working directly with prediction market vendors and corporate implementers, and having watched, thought about and pontificated on this phenomenon for over a dozen* years, Silverstone's points seem reasonable as high-level headings. Several caveats and elaborations may be useful on this rapidly-moving target, however. *(I first heard about them in a pitch by Ken Kittlitz to the Digital Commerce Society of Boston circa 1995/96 -- shortly after I nominally helped co-found that late and loosely constructed organization, and shortly after Kittlitz founded the Foresight Exchange)

First, the number of managers completely unfamiliar with the concept of prediction markets has dwindled markedly over the past five years -- and at an accelerating rate. Silverstone is, I'm sorry to say, somewhat behind in writing that:

A relatively small number of companies including Google, HP, Intel, Yahoo, and Eli Lilly have jumped on this idea by creating prediction markets to aggregate the opinions of employees...

That was true five years ago. It is not the case today.

Doing some research into this recently, I was pleasantly startled to discover that there are at least fifty large corporations running prediction markets 'for real' now, and countless others running small, informal and for the most part un-trackable pilots. (Virtually all of the major vendors offer such trials, though with varying degrees of publicity and enthusiasm as well as varying degrees of ease in setting them up -- though none are terribly difficult). And that's not counting the number of academic institutions and news media organizations that run prediction markets for other reasons (e.g., experimental learning and securing reader attention, respectively).

Add to that the fact that HBS and HBR are covering this topic, conferring a halo of corporate legitimacy, that Surowiecki's book continues to thrust the key concepts into business management if not lay-public consciousness, that barely a week goes by without a major publication giving some ink to the subject (I've stopped even blogging them individually, there are so many) and that the proportion of people in the management class who trade various financial instruments is very high. I.e., they recognize, at some level, that prices represent crowd wisdom -- a body of information to which they may or may not be privy, but which they ignore at the peril of their 401K.

So, given that backdrop, I would turn Silverstone's point around and ask: What excuse can a manager offer, in May, 2008, for being totally 'unfamiliar with the concept' of prediction markets?

Wait. Don't answer that. If your boss isn't reading this, your CEO or your board very well may be, and that's not an idle quip. Some of the most important prediction market initiatives I've seen have come down from the very top, rather than from middle management...

...which leads me to Silverstone's second point: "Internal Friction". That would be putting it mildly.

Some cultures palpably chafe at the idea of unleashing democratic forces within their walls. This is not a scintillatingly deep or novel insight. We know this to be true for countries, and families, as well as for bowling leagues and PTOs for that matter. (I've also observed this when running scenario workshops within certain corporate and, treading lightly here, societal cultures).

Humans almost universally seek some kind of power or at least stability. They work hard for it. And they're loathe to give it up to some impersonal, anonymous and (it can sometimes appear) fickle and BS-penetrating tool (however powerful) that doesn't even have to justify its reasons. (Imagine an election in which each voter had to justify his or her reasons -- to someone -- before casting a vote and you'll understand what I mean.)

Which is all to say that the issue goes well beyond friction to questions of culture, organizational structure, leadership, and the role or perceived role of middle managers. Classically the role of 'middle management' (even the term has taken on a derogatory ring) has been to act as filters and conveyors of information moving between senior management and workers.

Prediction markets usurp some of that role and that can be, quite understandably, unsettling to them. Nobody ever said business was fair. But, bottom line: it is a real reason why prediction markets need to be contextualized carefully in order to succeed. Repeat after me: it's not about the software (mostly).

Which brings me to Silverstone's last point: Implementation Issues. Let's be clear about one thing: 'implementation' here does not refer to the kind of thing one thinks about when installing a big new software system. The software for these things is pretty darned simple in that regard.

The much bigger questions have to do with organizational issues. And those can be sweeping.

For example:

What questions is an organization, a) truly interested in knowing the answer to (strategic), b) willing to tell the employees they might not know the answer to (potentially humiliating) and, c) willing to feed back information on how and why we did or did not act on that input (empowering of morale and loyalty... or cynicism... depending on how it is handled). Just the first part of that is a big, hairy strategy question that goes right to the heart of what a company wants to be when it grows up and where it thinks it is currently 'stuck'.

Bottom line: prediction markets harbor all the paradigm-changing, employee-empowering breakthrough-idea-generating, process-improving, cost-saving potential of the suggestion box... as well as all of its potential as a reason to deride management.

Here are the twists that can and do cause managers to think carefully about the transparency they're prepared to unleash: prediction markets are real-time, they're at least nominally public, and the information they generate isn't really a suggestion (casual) so much as it is a serious investment of time, thought and expectation on the part of those asked to participate. One has to ask then: With all of that potential for discomfort, why have so many big companies begun using them in earnest?

From Risk to Uncertainty

I don't agree with everything in this piece by Thomas Homer-Dixon that appeared last week in the Toronto Globe and Mail, but this quote is an absolute gem (emphasis added):

Our global financial system has become so staggeringly complex and opaque that we’ve moved from a world of risk to a world of uncertainty. In a world of risk, we can judge dangers and opportunities by using the best evidence at hand to estimate the probability of a particular outcome. But in a world of uncertainty, we can’t estimate probabilities, because we don’t have any clear basis for making such a judgment. In fact, we might not even know what the possible outcomes are. Surprises keep coming out of the blue, because we’re fundamentally ignorant of our own ignorance. We’re surrounded by unknown unknowns.

It's something I've said for a long time:

It's tempting to think that all things are predictable given enough information, enough minds, enough time and enough computing power. It's just not true. (Which is not to say that some things are not predictable... and with incredible precision... a phenomenon that leads to overestimating the scope of problems and questions that lend themselves to such methods.)

Telling which is which is the trick...

I would go even further to say that really smart people who, by life experience know that some things are fundamentally unpredictable still draw an unvoiced sense of emotional comfort in their business life from the idea that some wise expert somewhere has been to the future (for all intents and purposes) and if we could just find him or her things would be OK... and/or that a really sophisticated computer model or prediction market (the 'collective mind') can provide crystal ball-level insights.

Sometimes yes. Often, no.

I liken Mr. Homer-Dixon's observations to those tragically massive car pile-ups that happen a few times of year in fog-prone areas like the Central Valley of California. Everyone is driving along at a reasonable speed, with reasonable spacing between vehicles. People are sipping coffee, tuning radios, maybe talking on cell phones. Slightly distracted, but mostly responsible. All is normal.

Then the first guy hits a fog bank and can't see squat. He taps his brakes. The second guy sees red lights and fog coming up fast and taps his brakes just a little bit harder, and so on. In just a few seconds, hundreds of cars end up in a tangled heap and people die. All because the guy in front was convinced by every one of his senses and not without justification based on experience that the visibility on the next 100 yards of road would be the same as on the last 100 yards of road.

Popping up Like Mushrooms

Not clear that there's anything new here conceptually but the presence of another news-oriented prediction market during a U.S. election year is further evidence of the power of a good idea whose time has come.

Hubdub - a new Web site [Nigel] Eccles and three colleagues in Edinburgh, Scotland, assembled - where customers will bet for fun, not money, on the outcomes of real news stories. The site launches Monday [1/28]...

After signing up, you'll receive 1,000 "Hubdub dollars," play money that works only on the site. You can look at stories about, say, whether Gregg Williams will be named the next head coach of the Washington Redskins or who will win the Florida Republican primary.

Guess right, and you'll win more Hubdub dollars. Lose, and your account will draw down. In the spirit of the board game Monopoly, where simply sticking it out is rewarded, you'll also get 20 new Hubdub dollars ever day you log in.

Hmm... can you say millisecond-scale auto-logins leading to hyper-inflation?   :)

"What I realized was it wasn't the fact that I had lost money, but my pride and feeling of self-worth had taken a real knock," Eccles said. "I think that's the case for a lot of people - it's not so much the cash, it's being right."

Or in his case, being wrong. Methinks someone needs to re-write his press-releases. H/T: Bob Weber.

One of the Benefits of Working in "Cube Land"

Your choice of seating at work may affect much more than your sensory pleasure at knowing when someone is eating tuna fish or onions versus citrus fruit or chocolate. Google has discovered that it's the most important factor in your being in the loop: in particular, whether you win or lose based on your close colleagues' prediction market trading.

(See 'heat map' at right: successful traders are depicted in green; losing traders in red).

(H/T: Jeffrey Henning)

'Real' traders take note! (Pick your desk-mates carefully.)

Random thought: this adds a whole new dimension to the elaborate strategies seasoned business travelers use to get primo seating on airplanes. Exit row? No thanks. Just give me the seat next to Warren Buffett.

The Future Will Be... Different From What We Expect

Great insights from science writer Ed Regis [emphasis added]:

...after watching all those forecasts not come true, and in fact become falsified in a crashing, breathtaking manner, I began to question the entire business of making predictions...

...the source of this incredible mismatch between confident forecast and actual result [is that] the universe is a complex system in which countless causal chains are acting and interacting independently and simultaneously... all of them running in parallel, and each of them often affecting the course of the others...

It is therefore

...hopeless to try to specify in advance what's going to happen as they jointly work themselves out.  In the face of that complexity, it becomes difficult if not impossible to know with any assurance the future state of [a] system...

(other than tides, moon phases, etc.)

Further, it's an illusion to think that supercomputer modeling is up to the task of truly reliable crystal-ball gazing.  It isn't.  Witness the epidemiologists who predicted that last year's influenza season would be severe (in fact it was mild); the professional hurricane-forecasters whose models told them that the last two hurricane seasons would be monsters (whereas instead they were wimps).  Certain systems in nature, it seems, are computationally irreducible phenomena, meaning that there is no way of knowing the outcome short of waiting for it to happen.

After accepting this reality (something that's as or more true for systems in which human beings play a role, e.g., social, economic, political, business, etc.) the next best things to knowing and predicting are imagining and recognizing:

• imagining multiple, archetypal ways in which the future might evolve, and
  
• collectively being able to recognize (if only a little bit faster than the other guy with whom you are competing) the key events and sequences of events that signal (if only weakly) that one or more paradigms (and not others) are starting to unfold.

Corporate Prediction Markets Pass Major Milestone

The airtime that Newfutures' North American VP, Norris Clark got on CNBC with Maria Bartiromo last week is a milestone unto itself in the steady rise of prediction markets into mainstream consciousness and respectability. What he had to say put an exclamation point on that milestone [emphasis added]:

"...we've been setting up over 100,000 prediction markets for our corporate clients."

It's unclear how that statement squares with the undated claim, on Newfutures' website, that they have

"...created and operated more than 100,000 markets on a wide variety of topics both public and corporate."

But by any measure, it's a big number, begging the question: Have we reached, or are we already beyond a tipping point in the corporate use of these tools for strategic competitive advantage?