Morgan following the numbers to brink of World Cup glory

Have pushed their use of data to new levels as their predictive algorithm plays out every match in advance

When England played Australia in the Twenty20 World Cup, a curious thing happened. Moeen Ali, who had been England’s opening bowler and best bowler in their first two games, did not bowl at all, even though he was fully fit.

Yet the strangest thing of all, perhaps, was how predictable it was. Analysts knew that Moeen would bowl much less against Australia, if not at all. While he had taken four for 35 from seven overs in the first two games, on each occasion bowling the first over and three overs in the powerplay, Australia suited his gifts far less well.

The reason was simple. One traditional piece of cricketing logic that data has endorsed is the notion that spinners are more effective when turning the ball away from the bat. Moeen’s off-spin is an asset against left-handers, but is far less effective to right-handers. Australia’s opener and captain, Aaron Finch, eviscerates off-spin. And, so, against Australia, Adil Rashid opened the bowling for the only time in the competition. Rather than Moeen, England used Liam Livingstone in the role of fifth bowler: Livingstone’s off-spin is less effective than Moeen’s, but he bowls very useful leg-spin. With Finch batting until the 19th over, Livingstone duly bowled four overs – each ball he delivered to Finch was a leg-break – and returned one for 15.

These decisions, like all those taken by England, were influenced by Eoin Morgan’s discussions with Nathan Leamon, England’s whiteball analyst.

Before every match, Morgan analyses how opposing batsmen fare against different types of bowlers and where they score their runs.

This has been standard practice for several years, but Morgan now has a new, cutting-edge tool: an algorithm, which predicts the impact of different combinations of bowlers against a certain batsman. This allows him to assess different options for who should bowl when.

Data in cricket is often accused of being overly historical. Focusing on par scores can be too conservative, and not take account of how the game is evolving. Graeme Swann recalls that against Sri Lanka in the 2011 50-over World Cup: “We’d batted to our batting plan perfectly, got 229, everyone said ‘brilliant’.” England lost by 10 wickets.

England’s use of the predictive algorithm represents an attempt to harness data to illuminate not what has already happened, but what could happen. It uses information about players’ records and, crucially, takes into account the ground and conditions.

Each game is simulated several times in advance, to assess all the likely scenarios. Yet, for all the work that Morgan does with Leamon before matches, he is renowned for keeping his team talks brief and to the point – generally, no more than

15 minutes with the batting and bowling groups alike.

While such tools help inform Morgan’s planning, the art of cricket captaincy is to react to the game. To this challenge, Morgan brings his cricketing intuition, acumen and calm, unemotional demeanour.

Yet, in the cauldron of matches, he enhances these qualities by regularly looking towards the dugout. There, he will see Leamon, with two clipboards in front of him. These show a number and a letter – for instance, 4E. These are signals, believed to refer to who should bowl the next over and what the field should be. England have used the system since the start of 2020.

What the signals advocate is not predetermined. England constantly run live computer model simulations during the game. The aim is to ensure that, rather than playing the game as they may have intended to if all went right – the victory over West Indies, who were bowled out for 55, was perhaps the closest to this ideal – England can be tactically flexible. There is, for instance, no pre-determined bowling order.

“It’s trying to integrate all the information with exactly what’s going on in my brain and throughout the game,” Morgan explained. Having access to “what the blackand-white decision might have been”, is “an instrument to try and make me be a better captain.

“It’s helped myself and Jos [Buttler, the vice-captain] come to better decisions at different times and, equally, sometimes we’ve gone against it and it’s worked as well.”

Morgan is free to ignore the signals if they seem out of kilter with what he picks up out in the middle, yet his grasp of T20 is such that his decision generally aligns with the data. Across the three-game T20 series in South Africa last year, Morgan said that he made only seven bowling changes that differed to what Leamon’s signals suggested.

One hallmark of England’s World Cup campaign has been the degree to which it is possible to predict their decisions. Tymal Mills, normally a death-bowling specialist, bowled in the sixth over against the West Indies to exploit Chris Gayle’s vulnerability to pace. Rashid did not bowl until the left-handers in their top six were all dismissed.

Similar logic has permeated the approach to batting. Against both West Indies and South Africa, England elevated Moeen to use him as a floating hitter against spin, replicating the role that he performs in franchise cricket. They reflect how England are embracing the logic that what matters is not a player’s batting position, but in which over they arrive at the crease.

No decision is the product of England’s predictive algorithms or inmatch simulations alone. Morgan has a brilliant, nimble cricketing mind; Buttler advocated Moeen’s promotion against Sri Lanka to target spin and the shorter boundary.

Perhaps more than any other international team, it is possible to deduce what England are about to do from crunching the numbers.