Cricket is having its Moneyball moment
Michael Lewis’ blockbuster bestseller “Moneyball” not only immortalised baseball coach Billie Bean’s use of stats for player recruitment & player selection (the movie with Brad Pitt in a starring role also helped), it changed the nature of the sport. Now baseball and basketball in the USA and football in Europe are heavily stats oriented sports. Coaches look not just to acquire stars with big price tags but they identify stars with specific skills which the broader market might be undervaluing eg. goalkeepers who excel in penalty shootouts but who might not be very good at defending high balls. This riveting read from Wired talks about the same phenomenon coming to T20 cricket courtesy a bunch of English and Aussie cricket addicts: “In 2013, a British businessman, Aidan Cooney, was preparing to sell his sports-data company, Opta, for a life-changing sum of money. Cooney had bought Opta in 2002 and positioned it at the forefront of an information revolution in football. He realised that there was more going on in a match than was captured by official statistics, so Opta began counting different types of action and analysing them. Then it started selling its findings to clubs and broadcasters who were hungry to understand the game at a deeper level. As he was preparing to hand over the firm, a couple of colleagues persuaded Cooney that there was an opportunity to do in cricket what Opta had achieved in football. Because cricket consists of a series of individual plays, it could be broken down and analysed more easily than football, they argued. Cooney agreed and in 2014 he launched a new company, named CricViz.”
The data that Cooney collected next is critical because in investing jargon we would say it gave him a ‘sustainable competitive advantage’ over his potential rivals: “To offer something new, Cooney realised he needed unprecedented amounts of data. To help him find it, he pinched one of his Opta analysts, Phil Oliver, who is now CricViz’s managing director. Oliver believes CricViz has the most comprehensive and sophisticated cricket database in the world. He explains its design as a four-storey stack, with each layer adding more detail. The base layer is a set of complete scorecards. This gave the firm the number of runs scored, wickets taken and catches held and the overall results of every documented match since Test matches began in the nineteenth century. But this was just the start. Searching for the next layer, Oliver struck gold. He found an Australian cricket statistician, Charles Davis, who had been painstakingly translating physical scorebooks into a ball-by-ball digital archive as a hobby. Oliver flew to Melbourne and pleaded with Davis to share it. Davis acquiesced. Now CricViz knew not only the results of every match, but some of what had happened with each ball.
Next, Oliver licensed a data feed of information that Opta had been recording for years on fielding positions, shots played by batsmen, the types of ball delivered by bowlers. Finally, it added what OIiver describes as “the magic ingredient”: data from two ball-tracking systems, Hawk-Eye and Virtual Eye, that have been logging cricket matches for 20 years. This stack means that for any single ball in international or major club cricket in the past generation CricViz knows who the bowler was; the sort of ball they delivered; where it bounced; how it moved in the air; if the batsman played a shot and what kind; if any runs were scored and where; and if a wicket was taken and how.”
Once all the data was in place, CricViz asked Nathan Leamon – now the data analyst for the England cricket team – to build “to build models to suggest how teams might perform in the future based on previous outcomes from the past…Cricket fans first took note of CricViz because of WinViz, an algorithm built by Leamon that takes account of the players, the venue and the state of the match to run a series of simulations to determine the most likely outcome. It then ascribes each team a percentage chance of winning the game.”
Now, CricViz caters to a clientele which includes TV broadcasters, cricket boards like the ICC itself and the leading T20 franchises across the world. Former England Test captain, Michael Artherton, who is a commentator for Sky Sports is a user and a fan. He explains why tools like CricViz are useful: “I’ll say to myself, ‘Who’s winning this game?’, and then I’ll price it up in my mind and check it against WinViz to see how it compares. The interesting thing is that most of the time I’ll be within two or three percentage points, but when I’m massively out of sync, that’s the starting point for a conversation”.
What is even more interesting is how IPL franchises use (or should use) CricViz and WinViz. Freddie Wilde, who entered CricViz five years ago straight out of college (in this day and age, nobody wastes their twenties getting credentialled by a multinational firm), and who is now one of the world’s top T20 strategy analysts says: ““We understand Twenty20 as much, if not more, than most commentators and broadcasters”, he says. Analysing a game with Wilde is a mind-expanding experience. He talked at length about a recent match in the IPL, in which Rajasthan Royals batted first and built a huge score of 220 runs for the loss of just three wickets. Their opponents, Sunrisers Hyderabad, wilted, and could only muster 165-8 in response. It seemed a straightforward thrashing. But for Wilde, there was lots to talk about.
One of the key features of the CricViz database is the ability to look at how well batsmen play against different types of bowlers. This enables Wilde to say which player match-ups are likely to be particularly good or bad for each team. In this game, there was only one Sunrisers bowler, a spinner, Rashid Khan, who had a favourable record against the Royals’ two most destructive batsmen, Jos Buttler and Sanju Samson. Khan is usually most effective at bowling in the middle overs of an innings, but the Sunrisers captain bowled him early on to try and get Buttler and Samson out quickly. According to Wilde, “This was a huge strategic battle. The game hinged on Rashid against these two players”. But Buttler and Samson were wise to what was going on. They played Khan respectfully. The Royals scored 24 runs from Khan’s 24 balls and then went berserk, scoring 196 runs from the other 96. “These are the sorts of things that are shaping the match”, Wilde went on, “but they are only really apparent if you have done your research, or if you have numbers in front of you”.”