The selection process is not fit for purpose - Analysis of Sam Curran's T20 call-up

25th January, 2018.


Since official sport ever started taking place, the selection process in the majority of sports has been the same - subjective, and biased 'eye tests'. In some sports such as baseball, famously with 'Moneyball', this situation has changed for the better, but cricket is still lagging hugely behind.

This was confirmed as recently as the Ashes series, which ended this month in a comprehensive victory for Australia, 4-0 over an inept England team, particularly from a batting perspective, which I predicted here in advance. However, rather laughably, the Australian selectors were given the credit for this victory, with coach Darren Lehmann suggesting that 'gut feel' selection regained the Ashes. If you are interested in reading more about this, here's a link to a decent cricinfo article, but I'll quote an interesting piece here:-

"We were just going with gut feel to be perfectly honest and really pleased for the players that they turned up and played really well," [Darren] Lehmann said. "I was impressed with each one of those selections and the way they turned up in the series so far.

"There is [satisfaction for getting calls right] I suppose but full credit to Trevor [Hohns] and Mark [Waugh] and Greg [Chappell] for that. Yes I'm on the selection panel but it's a tough, thankless job, it's just pleasing those guys get the credit where it's due to be perfectly honest.

"Those guys we picked were fantastic. You don't know how it's going to go when you first start and you cop a lot from people and media. That's the way it is, so really pleased for the selection panel."

The simple fact of the matter is this. Australia are pretty much always strong in home conditions - you only have to look at their reasonably recent Test batting and bowling data to understand this - and these 'gut selections' were actually put into an environment where they were more likely to thrive than struggle.

Would Mitchell Marsh, Shaun Marsh and Tim Paine do any better than those who were rumoured to be in competition with them for places in the team? It's an impossible question to answer, but it's far from impossible to discount the fact that those who rivalled them for squad places would not have actually done as well, or even better.

Australia's selectors got the credit for Tim Paine's selection, but we have no idea if other options would have been better...

Certainly, I'd be interested to see how these 'gut feel' selections fared when they play away series in the subcontinent. My suspicion is that the selectors will be highlighted for rather different reasons.

By no means are Australia the only team with such issues. I've already highlighted England's selection of James Vince - my pre-Ashes critique made national newspaper coverage, as well as several cricinfo articles - and anyone who missed this can check out my article here, and also the linked article above regarding England's batting pre-Ashes in general as well.

As mentioned at the start of this article, the main issue with any selection process (or indeed, recruitment process) which doesn't rely on data, is that it is subjective and has strong potential - read likelihood - to be biased. Let's give some basic examples, starting with Vince, whose aesthetically pleasing style has won him plenty of admirers. However, a textbook cover drive for four does not score ten. It scores the same amount as a slog over cow corner, and it's important to remember that.

Look at the celebration of some players. We have Tom Curran's wide-arm celebration, Rashid Khan's extravagance, Tabraiz Shamsi's phone impersonation, and Kesrick Williams' notebook. All are very noticeable and stick in people's minds, offering the potential for inherent bias there and then. The human mind frequently makes us think that a notable event happens more often than it actually does.

In the aforementioned Moneyball, Michael Lewis' excellent book - a true story, if you are not aware - he describes the lack of development of the book's hero, Billy Beane's playing career.  Beane was undoubtedly the scouts dream - perhaps even the equivalent of James Vince of the baseball world:-

'They all missed the clues. They didn't notice, for instance, that Billy's batting average collapsed from over .500 in his junior year to just over .300 in his senior year. It was hard to say why. Maybe it was the pressure of the scouts. Maybe it was the other teams found different ways to pitch to him, and Billy failed to adapt. Or maybe it was plain bad luck. The point is: no one even noticed the drop-off. "I never looked at a single statistic of Billy's," admits one of the scouts. "It wouldn't have crossed my mind. Billy was a five-tool guy. He had it all."'

This happens so much in cricket. How many times has a commentator praised James Faulkner, suggesting he knows how to find a boundary? This type of commentary happens a lot - don't you remember, once upon a time, he played the innings of his life against England - but the statistics actually show Faulkner now to be one of the worst boundary hitters in professional T20 cricket, particularly against spin bowling.  Cricket journalist Jarrod Kimber's recent excellent analysis of Faulkner's batting can be found here.

Of players worldwide with a decent batting sample size, James Faulkner's boundary-hitting percentage is one of the worst...

England's Cricket selectors are also seduced by youth and excitement, as well as the already discussed aesthetics. So often, we see a random young player fast-tracked into the national team, without ever justifying why they should have ever been included to start with. Mason Crane is a prime example, with there being are a plethora of county spinners offering much better numbers than the young prospect at the time of writing.  Sam Curran, Surrey's 19 year old bowling all-rounder, is another, and he was called up to the England T20 squad for the first time on Tuesday.

From a recruitment perspective for a club team in a variety of sports, putting faith in young players makes a lot of sense. Giving young players a chance has the obvious benefits of being able to pay less wages to younger players than established players, building team loyalty from players at a young age and in football, for example, being able to sell the player on for a profit.

However, it makes far less sense for a national team. I made the point previously that if James Vince averages 50 in division two and 30 in division one, how on earth could anyone expect him to do better against Mitchell Starc and co, and the point obviously still holds. While I can understand why a veteran might be disregarded by selectors, but in general, why wouldn't a national team pick the best players available to them? 

James Vince was always likely to struggle in Test cricket against bowlers of the calibre of Mitchell Starc...

Unfortunately, at this time, Curran isn't the best player available to them. In the future, he may be, and he may even stand a better than 50-50 chance of doing so (at 19 years of age, he has a ton of upside), but at this time, he's not even close. The squad has lost a top-order batsman (Joe Root) and a batting all-rounder (Ben Stokes), so why would a bowling all-rounder be picked to replace them? It's like a football team losing a striker and attacking midfielder, and picking a defensive midfielder in their place - an obviously illogical move.

Regular readers may also remember that in September after the T20 Blast season, I published statistical analysis of English players in T20, and this can be read here. Analysis included all-rounders, and of the English all-rounders with a big enough sample size, Sam Curran ranked 19th on my list, and with six spin bowlers ranked above him in that list, by definition he was 13th on the list of non-spin all-rounders.

In this article, Curran's expected T20 Blast average for next season was 12.29, and strike rate 106.06. This makes him out to be a tail ender who should be hidden down the order. His bowling numbers weren't bad (expected average of 26.90, expected economy of 8.11) at all but they still ranked him 31st in the bowlers list.

Curran's supporters will no doubt point to his improved batting data in the New Zealand domestic T20 competition, the Super Smash. For Auckland, in six completed innings, Curran scored 157 runs (average 26.16) at a strike rate of 161.85 - seemingly impressive data, and a marked improvement from my expected T20 Blast numbers for him.

However, such enthusiasm needs to be calmed down by the fact that the New Zealand domestic T20 event is one of the easiest leagues around the world to bat in.   My algorithm reduced Curran's exploits in New Zealand to an expected average of 22.34, and strike rate, 155.04 if playing in the T20 Blast, and considerably lower for a number of other franchise leagues around the world.  When these six completed innings are treated in isolation, these are still really good numbers for a bowling all-rounder, but they aren't quite as impressive as the original data suggests.

Treating six innings in one competition in isolation is also fraught with danger, in that it opens up analysis to considerable amounts of variance. On a side note, I anticipate IPL franchises to exhibit a lack of awareness of this when they over-bid for players who have performed better than expectations in the recent Syed Mushtaq Ali T20 event in India.

Who has better data that should have stood a chance, if England were looking for a seam bowling all-rounder? 

One name stands out - a player who has demonstrated over a large sample of data that he is a top bowler, as well as a better batsman than Curran, and that player is Benny Howell. However, Howell plays for an unfashionable county, Gloucestershire, and has unspectacular red-ball data. In effect, Howell is a T20 specialist, but that shouldn't be an insurmountable issue - we are looking at T20 selection, so why would his ability in other formats be an issue?

If England need an all-rounder, Benny Howell offers a much better current option than Sam Curran...

As well as Howell, other better statistical options ahead of the younger Curran brother, who aren't currently in the squad include the veteran, Tim Bresnan, a boundary hitter who can bowl in Ryan Higgins, and Paul Coughlin.

It really is a mystery why the England team is not using data to drive team selection, when it is virtually impossible for three selectors to effectively cover 18 teams, without basing their decisions either on inherent bias or variance. This interesting article from The Independent, written in July, also showed the lack of input from head coach, Trevor Bayliss, on the selection procedure, and I quote:-

"This is a man who earns upwards of £500,000 a year. It’s about time he started working a bit harder for that money.

It is not being churlish to ask exactly what Bayliss does.

We know what he doesn’t do because he is not shy in admitting he often defers to the greater knowledge of England’s other selectors precisely because he hardly watches any first-class county cricket."

Intelligent usage of data would address such shortcomings. Until England's selectors, and other cricket teams, learn to embrace data, instead of fearing it or not considering it relevant, they will be at a significant competitive disadvantage when playing against teams that do - something that will happen more and more frequently in the future.