Wednesday, August 13, 2008

A tangled web

Image: Screenshot from Web of Fate

A startup site called Web of Fate, launched in July 02008, crowdsources a timeline of users' predictions, goals and other ideas about the future:

Web of Fate is a social experiment that harnesses the collective intelligence of the web to visualize and uncover hidden relationships among future and historical events.

Maybe it's the liberal use of bold type that makes this read like it belongs in the voice bubble of a comic-book's mad scientist.

Web of Fate is ambitiously eclectic. It runs on user submissions, with four options; "make a prediction", "quote a prediction", "create a capsule", or "set a goal", thereby serving as a hybrid of prediction registry and to-do list. There is apparently a "semantic analysis engine" under the hood that produces links between the fragments of submitted content, but exactly what it does, and how it relates to grand questions posed on the site (such as "What can we learn from the past and make the future better?") is unclear.

The visualisation tool looks extremely interesting (see the image above); a cloud of prediction stems hover onscreen in a way that promises a dynamic exploratory interface (a la Visual Thesaurus), but although you can drag the nodes around, this yields no additional insights -- the function appears to have no informational referent. Also, it piles on the disparate elements (users' predictions, reported predictions, and goals) in a way one might hope would be thought-provoking, but which in practice generates more confusion than enlightenment. So a statement like "I will go to the gym everyday" (with no username or date given) sits alongside "$200 dollar oil prediction" (by the end of this year; from Daron in San Jose) and "By 2009, more than one third of IT organizations will have one or more environmental criteria in their top six buying criteria for IT-related goods" (submitted by a user citing IT research outfit Gartner). Those interested in accessing the material primarily through the visualisation may be disappointed when clicking on any prediction takes them to back to a page of static text, and similarly, any given category (users' predictions; quoted predictions; goals) is viewable only in lists, at this point.

A review of the site at Mashable (social networking news) says:

[M]y sense is that registered users will be quite pleased with what they’re offered. Given enough members, this place could become a real hotbed for forecasters.

I find myself wondering whether the conditions are in place, at the moment, for this service to generate that kind of momentum. In their current form, the range of functions seems too diffuse, the quality control too lax, and the interface too unwieldy, to attract and keep a core base of users, who may find more forecasting edification elsewhere at prediction markets like Predictify and NewsFutures, and dedicated "time capsule" functionality at future email services like FutureMe or EmailFuture. There's also the Long Now Foundation's Long Bets project [disclosure: I'm an advisor on it], which has since its inception addressed the quality control issue by requiring users to provide real names and front real money in support of their speculation. The tradeoff, however, is that it doesn't act as a repository where users can freely submit other people's ideas about the future whenever and wherever they appear. (However, Long Bets cofounder Kevin Kelly points out that such a registry has always been part of the plan.) I believe a reputable, well-designed prediction registry has long been needed and is technologically possible, but is yet to be done.∞

According to a Fast Company blog, Web of Fate is the brainchild of Arizona State University student Max Yuan, who, writes Francine Hardaway, "has a head full of projects I hope he has time to complete because they all seem great to me". I'm certainly pleased to see a prediction registry that enables submission of one's own ideas as well as reporting the prognostications of others; the latter an important prerequisite to improving accountability and quality in future-oriented public discourse. Web of Fate makes a brave effort to tap some of the potential for such a forum, but I hope Yuan can find the time to revisit and overhaul some elements of its design. Otherwise, I venture to suggest, this web will (to stretch the metaphor horribly thin) remain too tangled to catch the prey for which it was spun.

∞See futurist David Brin's 02005 article on this topic for more.

[Also posted at Long Views, 14/08/08]


Anonymous said...

Hi Stuart,

I also looked at the Web of Fate. I thought it was a neat visualisation but in the 'not sure exactly how I would use this' bucket. I think it is in its early stages to looking forwards to seeing how it develops.

A prediction registry is actually a key part of what we are developing at Hubdub. We are currently tracking about 50,000 predictions on 2,000 questions. Most users elect to make those predictions public. For example here are all my predictions:

I'd love to hear more about the Long Now. Drop me an email some time to nigel [at] hubdub [dot] com

Stuart Candy said...

Thanks for this, Nigel. I see from that link that you're quite a prolific predictor!

An advantage of tracking short-term events is of course that there are many more cycles, so you can build up a large corpus of predictions, and results, relatively quickly. Long Bets, being part of an institution dedicated to long-term responsibility, has a two-year minimum time horizon, intended to tune attention to slow/significant register as opposed to fast/trivial. What's tricky is that the web tends to thrive on the latter, and the former intrinsically offers less action. So the low liquidity and volume of traffic on long-term predictions in the market would be a challenge.

It'd be interesting to know a bit more on how those 2,000 questions (being tracked at hubdub) are distributed on the dimension of time horizon. What proportion are looking further ahead than, say, six months? Two years? Or a decade?

nigeleccles said...

There is a very strong bias to shorter term markets. I think they average at about 1-2 months, although it varies by category. Sports has lots of short term markets. Science and technology have longer term markets. The key though is to incentivize trading on longer term markets. We are working on a couple ways to improve that.