A few weeks ago, I was invited to the Newsroom Fantasy Premier League. As an avid new American staff member of a British newspaper, I of course accepted the opportunity to assemble a fictional team of real players and earn points based on their actual statistical performance. I don’t want to do anything.
That’s when the problems started. Especially since I didn’t know who to pick for the team. I enjoy football as an aesthetic experience, but less as an analytical one. I’m not a great student of formations, forms, the transfer market, injury reporting, or management politics. T: Machine learning.
On GitHub, I found code developed by a real programmer for a system called AIrsenal, stole it, and implemented it on my machine. AIrsenal uses a “Bayesian approach”, estimates “conditional distributions”, leverages “Monte Carlo integration”, and uses two or more Greek letters in the underlying mathematics. Select, copy, paste, type. A curtain of blocky green numbers geysered over my black terminal window, simulating a match yet to be played, a goal yet to be scored, a victory yet to be celebrated.
This is how artificial intelligence assembled my fantasy team. All I did was follow the instructions. It was very modern, a little strange, not just fantasy football, but artificial fantasy football. Nevertheless, I was completely proud of “my” creation. After all, it was a geyser on my screen. In honor of an influential newcomer to the (real) Premier League, my team is called the American Billionaires (AB). I have no plans to violate computer recommendations in the next few months.
Under FPL rules, you’ll soon get £100 million to build a fantasy team of 15 real players. When they play in a real match, you can earn points from the menu based on their performance (e.g. Midfielder goals = 5, Keeper clean sheets = 4). Each week, you can buy and sell players whose prices fluctuate depending on their popularity. About 9 million people are currently playing.
“Computers are useless,” said Picasso. “They can only give you the answer.” Pablo was right enough, but the answer — my fantasy roster, the result of cold mathematical optimization — begged the question. Who are these computer-predicted human soccer players in Computer Fury? Instead of separating me from the humanity of sports, AI has brought me closer.
When Bruno Guimarains of Newcastle Appearing on my team, the output of an invisible algorithm, I learned all I could about the real man. attached. Chelsea’s Reece James, American billionaire right-back, dreamed of playing for a club (Chelsea, not AB) as a little kid kicking balls all night in his local park. Manchester United’s Anthony Elanga is AB’s reserve midfielder, fluent in French and Swedish, and looking to master Spanish and Portuguese.
Technology has the power to democratize. It has the ability to improve the quality of human knowledge and accelerate its dissemination. This power allows us to explore faster and deeper, digging up the rich veins at its core. But all the while, I could imagine the tweets and counterarguments rippled through the newsroom. He is cheating!
We spoke with AIrsenal creators Nick Barlow and Jack Roberts, elementary particle physicists and researchers at the Alan Turing Institute in London. Like most good ideas, AIRsenal started as a train conversation about four years ago.
“The problem isn’t the game’s complexity, it’s the information capture,” Barlow said. “If you read the newspaper, you can see that this player is either arguing with the manager or about to be sold to Barcelona.
Roberts added, “It’s much more difficult to represent the state of the world on the FPL than on a chess or Go board.”
Nonetheless, AIRsenal (opinions differ on pronunciation) had its eloquent moments. Referring to an unimaginably beautiful Go move by his AlphaGo AI in 2016, Barlow said, “We have our own move number 37. When he moved to Burnley, it was Joe’s I chose Heart,” he said.they also avoided Liverpool Mohamed Salah When he returned from last year’s Africa Cup of Nations, he was unpopular but proved to be a smart move. “Perhaps it was Airsenal’s lack of emotion,” Roberts said.
I also asked about the ethics of my decisions and was pleasantly reassured. “The more, the more fun,” said Barlow. “I welcome our robot overlords, for example.”
Airsenal finished last season at 976,423, firmly in the top 10%. If they win top prize, the Ethics Committee of the Alan Turing Institute has decreed that their creators will donate it to charity. I plan to do the same.
As I write AB sits third from the bottom in the relegation zone.
Oliver Roeder is a Senior US Data Journalist at the FT and7 Games: Human History(WW Norton)
follow @FT mug Be the first to discover the latest stories on Twitter
Americans looking to join UK newsroom fantasy premier league
Source link Americans looking to join UK newsroom fantasy premier league
The post Americans looking to join UK newsroom fantasy premier league appeared first on Eminetra.