The impact of data mining in T20 cricket
Adelaide Strikers won the toss and elected to field. This shouldn't come as a surprise because, prior to Friday's match, teams have preferred to chase 14 out of 17 times in the BBL and the three who batted first lost. I know all this because of ESPNcricinfo's newest recruit, Gaurav Sundararaman. He is a senior stats analyst, and before this, he worked with Barbados Tridents and the West Indies national team at the 2016 World Cup.
Before the Hurricanes-Strikers match, I asked him to prepare the kind of dossier he would for a team. Although we have some data, it's nothing like what he had access to when he was a team analyst.
There were things I already knew or at least thought I did. Like how Strikers have a weakness in their middle order because they have three batsmen - Travis Head, Kieron Pollard and Brad Hodge - who take a while to start scoring quickly. I also heard from one of the analysts who works with a Big Bash team that D'Arcy Short - from a pretty small sample size - smashes quicks and only goes at run-a-ball against spinners.
With Hobart batting first, Gaurav said this of their team strategies: "The average run-rate for the Hurricanes in the Powerplay was 6.94 in 2014-15 and 6.25 last season. So far this season, their run-rate has been impressive at 8.66. The welcome addition of Short has been vital towards this change, and the Hurricanes have won both the games in which he scored a fifty."
On the Hurricanes new singing, the replacement for Ben Dunk, Gaurav said, "Short has two fifties and has 140 runs from 82 balls. His strike-rate in the Powerplay is 197. He scored 60 off 39 in the previous game against Strikers. However, off the 21 balls he has faced against spin in the competition this year, he has scored only 24 runs and hence starting off with a genuine spinner could be an option for Strikers. Liam O'Connor and Head could open to reduce the run flow."
Head had only opened the bowling twice according to ESPNcricinfo's data, which is not entirely complete for the Big Bash. Tonight was the third instance, and as Gaurav suggested he was brought on for the second over. The first ball Short played a shocking shot that just got past mid-on at a catchable height. The next ball he was out. Short improved his strike-rate against spinners, but he's going to get a lot more chances in the future to face them.
Gaurav's research showed that it wasn't just Short, but almost the entire top order of the Hurricanes that were weak against spin. "Tim Paine has been dismissed three times in the Powerplay versus spin and his strike-rate is 96, while against pacers it is 131. In the 11 knocks he has played in the Big Bash in the last three years, Kumar Sangakkara has played more than 20 balls only once, and has fallen to spin once every eight balls. George Bailey has been dismissed by legspinners 11 times in the last three years in all T20s. Dan Christian gets out once every six balls, and has a strike-rate of 87 against spin. Jonathan Wells' strike-rate is 109 against spinners and 121 against fast bowling."
You might ask yourself how a T20 team pack itself with so many players with question marks against spin, but I myself didn't get the time as they lost 5 for 37 in the Powerplay to quick bowling, two each for Billy Stanlake and Chris Jordan.
The Hurricanes also may have made a tactical error in asking Christian to bat so early, considering that if they were to consolidate, they wouldn't have their best slog-overs batsman to build on it. In the end he faced one ball, was out lbw - and probably caught - and Bailey was gone next over. So even if they had held Christian back, and promoted Wells, he would have been in at roughly the same time.
Darren Berry tweeted, "Take 3 wickets in Powerplay, win 75% BBL games. Powerful KPI." This stat, that many other coaches have told me, stands up through all T20. Although they are yet to invent a form of cricket where losing three wickets in the first six overs is a good thing.
Something else Gaurav predicted was that the Strikers would "play an extra spinner to leverage on Hobart's weakness against spin". Michael Neser - seasoned seam bowler - was dropped for Tom Andrews, a 22-year-old left-arm spinner playing his second match.
After the initial carnage of the five wickets, Wells and Beau Webster put together a partnership. It was the second time that Wells has had to come in and bail out the Hurricanes against the Strikers. Two years earlier he'd come in early and made an unbeaten 58. This time he wasn't on his own, as Webster, in his first T20 innings, stuck around with him. Webster had no data to mine from T20, but he's played 21 matches in first-class cricket, has three hundreds, and has already played for Australia A. Their partnership was 89 runs - a strike-rate of 158 against pace, and 106 against spin.
Webster hit out at the end off O'Connor, but before that over, other than Jordan, the three lowest economy rates were those of the spinners. The total of 161 was better than what they deserved, but still below-par on a small ground.
It - the ground and the total - looked smaller with Dunk opening up. The Hurricanes know all about Dunk - he smashed them around Bellerive for a few balls last game - and, more dramatically, it was them who traded him for Hamish Kingston at the start of this season. Since that trade, Kingston has been handy, although was left out for the last game. Dunk, meanwhile, has been brutal. Before this innings of 79 off 49, he'd already scored more from four games than he had all of last season. He has even outscored Short, who took his place in the Hurricanes line-up.
Dunk's form should have meant more wins. Strikers lost the first game, after he had already broken the chase with a rapid 85, because they have an inefficient middle order.
"Hodge, Head and Pollard are all slow starters. The strike-rates of Head and Pollard indicate that they can catch up and Hodge can finish well. However, in the process of getting in, they do take up a lot of mid-innings resources. Can a team afford to have three slow starters in the middle order? Based on the situation, Strikers need to relook their batting positions. Just picking players with a high strike-rate is misleading and may not result in the desired outcome."
It hasn't helped that Head and Pollard have looked in pretty ordinary form, either. But, Dunk. Good old Dunk. He is in such good form that when Jake Weatherald was dismissed early, Head could come in at No. 3, with Dunk on 14 off 6. That meant there was no reason for Head to smash the ball around, he could play himself in, and then attack once set. On top of being a slow starter, and dreadfully out of form, Head also has issues against spin.
"One big weakness that the Hurricanes could look to exploit is his inability to play spin. He has a strike-rate of 97 against offspin and has been dismissed six times against it." So considering all that, it was a bit of a shock when the second ball he faced, off Webster's offspin, he smashed a straight six.
It looked like Head was trying to attack the spinners more and trying to start quicker - the one time he didn't need to do either. A few balls later he went a bit hard at a Kingston slower ball and was caught for 9. So far in this tournament, he has scored 23 (25), 13 (13), 7 (8), 24 (26) and 9 (8). Strikers might not have worked out their middle order to perfection, but they would never have expected Head to be this poor. That he was 7 off 2, and ended up 9 off 8, suggests that even the six didn't hurry up his start.
Then Hodge came in. Dunk was now 27 from 14, and Hodge started slowly as Gaurav thought he would. For the next four overs Dunk hit a four, or six, off every one, except the over he was dropped off the bowling of Short. And Hodge scored six runs from his first 11 balls. Then, perhaps as Dunk was tiring, or at least, slowing down, Hodge hit six balls for 24 runs, going after Kingston in particular. Had Hurricanes brought Stuart Broad on when Hodge was batting slowly and taken him out then, it would've meant another Strikers batsman had eaten up balls in the middle before being dismissed. Instead, Kingston bowled and Hodge destroyed him.
Both batsmen took boundaries off a poor Short over next. Broad came on after that, and took the wicket of Hodge straight away.
Pollard and Jono Dean both came and went, but Dunk's batting and Hodge's cameo had broken the chase, so Jordan, who has a modest 118 strike-rate and hits a boundary every nine balls, only needed to chip the ball around and stay in with Dunk. He did this, and in his 14 balls, he hit one boundary, scored 15 runs, and hit a six to finish the game. Dunk sat at the other end nursing a healthy not out.
Think about this, you only get eight games a year, and Hurricanes have just lost one to the man they traded away for a handy player who's made no impact this season.
On Dunk, Gaurav had written, "He has been explosive at the top of the order, scoring at a strike-rate of 200 in the Powerplay and will look to continue his good form. He has nine sixes so far in the tournament. He has been strong against spin in the Powerplay. In the last three years, Ben Dunk has a strike rate of 171 in the Powerplay against spin and a strike-rate of just 111 against the pacers. Hurricanes should be looking to bowl the pacers in the Powerplay and hope to dismiss or contain him."
The tried Gaurav's plan, but couldn't contain or dismiss him. Ultimately, that meant they lost to him.
Jarrod Kimber is a writer and Gaurav Sundararaman a senior stats analyst at ESPNcricinfo @ajarrodkimber