Before you start reading this, let’s just check in with WinViz. On the basis of data synthesised from previous pieces, this column currently has a 51% chance of being a success. You don’t need to know the precise workings of the algorithm, but rest assured it takes into account everything from page views and completion rates to the volume of anger registered in below-the-line comments – along with a unique proprietary sneer metric. Intuitive, AI‑powered analysis can even detect irony fatigue. And I see we’ve just dropped to 38%.
Cricket fans will know the drill. WinViz may have been around only a few years but already it is impossible to imagine the sport without it. During the recent Ashes, the most unpredictable in the contest’s history, its pie-face was forever popping up on screen to offer its percentage opinion on how likely England or Australia were to win. Which, under Bazball conditions, is about as useful as being shown where are you on a county map of Bedfordshire while you’re flying over it in an F-15.
Related: Glorious English summer hustle has been an Ashes rush like no other | Geoff Lemon
We’re now galoshes deep in the Hundred, where the same “win predictor” plays a prominent role amid a headache-inducing schema of hyperactive graphics. Given the tournament’s target audience, it’s fitting that its fluctuations tend to mimic the mood chart of a five-year-old on a diet of sweets and cola. And yet this evanescent, momentary insight has become embedded in the game’s contemporary discourse: fans quote it, players reference it and commentators continually ask to see it, often so they can instantly disagree with it.
How did WinViz worm its way so deep into our lives? It feels like the friend of a friend who showed up at a house party a couple of years ago, stayed over in the spare room and hasn’t left since. There it remains, taking up the sofa, getting drunk every night and telling us how really, no, really, Trent Rockets are a banker. All right WinViz, but they’re now nine down and 40 adrift. “No, no, mate, you misheard me. I’ve been saying all along that Phoenix are the sure thing.” Fine, sure, whatever. Just try to clean up some of these bottles before you go to bed, OK?
This isn’t an anti-analysis rant: the data revolution has been a fascinatingly revelatory era in sport, even for those of us who don’t have the technical or mathematical skills to process it all. As for AI, its practical applications within the industry are undeniable. Coaches and scouts across top-flight sport can already save themselves hundreds if not thousands of hours of watching video through its ability to read game footage. Athletes are using it to analyse and improve their training.
Inevitably, it will infiltrate the fan experience too and I, for one, welcome our new robot overlords. AI commentary might sound creepy (it certainly does the way the Masters did it), but the summaries created by cloning Hannah England’s voice at the European Athletics Team Championships weren’t half bad. The fact that a Dutch media group has managed to cover 60,000 local football games through automated reporting proves that the technology can widen access to all levels of sport. Even if the union won’t thank me for saying so and the chances of this piece pleasing people are now at a rock-bottom 5%.
One of the things at which AI is exceptionally good is predictive modelling, which is why WinViz’s creators have incorporated machine learning into their product’s most recent iteration. This makes sense, given that one of the most common fan complaints about WinViz is that it doesn’t know what it’s talking about. The algorithm has been wildly off its game since Ben Stokes and his New Model England went on the attack – not least because the vast archive of historical data on which it was basing its future outcomes is being rendered increasingly obsolete by the topsy‑turvy principles of Bazball.
There is, in fact, an entire subgenre of Twitter thread dedicated to owning WinViz’s fails. On the last day of the Oval Test, with Australia needing 146 runs for victory and Steve Smith and Travis Head well set at the crease, WinViz was giving them only a 33% chance of pulling off the heist. “Is this thing broken?” was the kindest response. (If you’re tempted to give the clever bot credit for getting the outcome right, remember it had England at only 48% – the draw got the rest.)
Obviously this is not a soothsaying device, merely a data-based probability machine. And that is the fundamental issue. The information itself may be accurate to within a gnat’s crochet at the moment of delivery but, on a philosophical level, what it communicates remains meaningless. Sport’s very uncertainty is its draw, its charm and its ultimate raison d’etre. If it cleaved only to the most likely results we’d have bored of it long before they started erecting stadiums in Ancient Greece.
As someone with no understanding of neuroscience, I shall confidently claim that, while watching live sport, three-quarters of my brain function is dedicated to the question of who is going to win. In my flawed human way, I’m drawing on my personal bank of knowledge and experience to parse the action in front of me, then horribly skewing the results with a massive input of irrational emotion. This, I suspect, is why so many of us react to these win-predicting bots with irritation and downright hostility: they’re stealing our turf, and doing it better.
What price our endless, pointless, gloriously stimulating debates about the state of the game, if WinViz is sitting at the table with its smugly statistical answer, like someone who won the pub quiz by Googling all the questions? Where’s the fun in that? If AI truly learns from everything that has ever happened, perhaps it should take a tip from the Romans and leave us our bread and circuses. Congratulations on reading this far, by the way. You’d never have thought it back at 51%.
Article courtesy of
Source link
