10th May, 2018. The first Test against Pakistan, due to take place at Lord's between 24th-28th May, 2018, heralds a new dawn in the selection of players for the national team, with the new chairman of selectors, Ed Smith, leading the selectors in picking his first squad. The reason why Smith's first squad heralds a new dawn is clear - he is an advocate of using analytics in sport, and it from media reports that there will be at least some usage of data in his selection methods, which from the outset is something I applaud. With this in mind, and given my obvious background in sports analytics, I wanted to assess the contenders for the England squad for the first Test - an exercise that many media outlets are performing, although perhaps with less, or no, data used. Another area which many media outlets are missing out on discussing are the anticipated conditions at Lord's in a fortnight, and we can use historical data to give us an excellent idea of the likely conditions at the venue. It seems rather obvious that a wicket which will suit England's pace attack will be prepared - a spinner's wicket will suit Pakistan much more - and this has traditionally been the case in Test matches in May. The table below illustrates the bowling data for Test matches played in May in England, from 2010 onwards:-
Here we can see that pace bowlers have been rather more successful, taking wickets at a rate of around 9 runs per wicket fewer than spinners, and in these matches, 79.6% of bowler-attributed wickets went to the pacemen. The figures for Test matches at Lord's in May this decade were even more pronounced:-
The spin average for matches at Lord's in May was almost double that of the pace average, and an absurdly high 86.6% of bowler-attributed wickets went the way of the quicks. Evidently, looking at these numbers, a quality pace bowling attack is imperative, and brings into question the necessity of playing a specialist spinner, who would be extremely likely to have almost a watching brief during the first-innings, at least. With this in mind, if I was on the England selection panel, I'd be advocating the selection of four pace bowlers and any spin coming from the batsmen, such as Moeen Ali (data suggests that he's a batsman that bowls, as opposed to an all-rounder), Joe Root, Dawid Malan or Liam Livingstone, depending on who is selected. This may be harsh on the likes of Jack Leach, who has taken wickets with regularity in county cricket in recent years, but I advocate a 'horses for courses' approach. If England are playing Pakistan - a team that has traditionally found it difficult to score at much of a rate against pace in Test cricket - in English conditions, then they must play to their strengths. Some may argue that teams should be picked with the future in mind - be it young prospects or future away tours - but I'd vehemently disagree. Particularly given England's shocking performances in away Tests this winter, a confidence-boosting win needs to be found, and found quickly. In any case, picking young players for a view to the future seems counter-productive, because frequently they are picked before they are ready, which then leads to them failing, and jettisoned from the team to rebuild their shattered confidence in county cricket. I'd say to the likes of Tom Curran and Mason Crane - come back to me when you average sub-25 with the ball in Division One. The extra benefit of this approach is that it will demonstrate the County Championship performances are a big driver in Test squad selection, motivating players to perform well for their countries, and would by definition illustrate selection is a meritocracy - a word which perhaps hasn't surrounded cricket selection on many occasions. In addition, picking teams now with a view to away Tests in the winter also seems counter-productive. England will be playing Tests in favourable, home conditions this summer, which will suit seam bowling, and these are completely different conditions to those that they will experience in Sri Lanka and the West Indies, where slow, spinning surfaces, are likely to be the order of the day. It may well be that some players picked this summer don't have the skill-set to succeed on these overseas tours, and vice-versa, and that's where the 'horses for courses' approach comes in. So then - who to pick? The table below illustrates the results of the main batting contenders for the England squad, sorted by their Test expected data, based on my algorithm, which accounts for a number of factors, such as recency of performances and difficulty of opposition:-
It's probably of little surprise to most readers that Root, Bairstow and Cook are streets ahead of the rest of the competition. They have consistently been the best three batsmen for England in Test cricket in the last few years and are complete certainties for the team, fitness permitting. However, from Cook (45.12) to the next contender, Rory Burns (39.04), there is a big gap, and there are 11 players (including Burns) with expected averages between 35 and 40 - very little indeed between the other contenders for the remaining batting positions. This indicates that England have a lot of good, but not great batsmen at this level, and expectations - from the fans and the media alike - need to be managed as such. With numbers at this level, it is likely that many batsmen will experience a bad run of 'form' - also known as negative variance - and will face calls for them to be dropped. Looking at these other contenders, Stoneman is the man in possession and on that basis, probably deserves to retain his place, while Burns is a strong contender - he's consistently accumulated Division One runs in recent years, and has already done so this season too, in often bowler-friendly conditions, and ticks the box of being a stable presence in the middle-order - someone who perhaps won't play as many flashy, risky shots, and this is an attribute that England, based on their performances this winter, should crave. Stoneman and Burns' Surrey team-mate, Ben Foakes, has also started the season well - and incidentally, so has another Surrey player Ollie Pope, who is rocketing up my expected average list, and warrants serious discussion among the selectors in the future. Whether Foakes will keep, or Bairstow retains the gloves, is another matter, but perhaps him being selected as a keeper would allow Bairstow the freedom to stablise the middle-order, with Foakes coming in at seven. Sam Northeast is injured - unfortunately coming directly after scoring a century for his new county, Hampshire - but deserves to be in consideration for future squads when fit, while it seems fair to assume Ben Stokes will take the all-rounder berth following his return from the IPL, where he hasn't performed. Some discussion on Stokes is worthwhile here. Ironically, despite his attacking batting style, Stokes has proven his effectiveness much more in the Test arena than in T20s, and my data doesn't make him even an above-average T20 player. His performances last year in the IPL were something of an outlier, and it wasn't a shock at all that he has mean-reverted this season. Assuming he can bowl, he'd be in my Test XI, however. So we have our top seven - Cook, Stoneman, Burns, Root, Bairstow, Stokes, Foakes. This means dropping the likes of Dawid Malan, who is 16th on my list, and also James Vince, who came out bottom of my list of contenders. However, this batting line-up has strength in depth and while there are only three world-class batsmen in it, should be capable of performing reasonably well in home conditions, at least. The second all-rounder spot to bat at eight - Stokes in the top six offers us this luxury - is a difficult decision to make, with few contenders in any sort of touch at all. Both Moeen Ali and Chris Woakes have had little game time in the IPL - Woakes has had more, but was unfairly dropped - and they both had very mediocre winters in the England team as well. Woakes at least does have a superb record in home conditions - see below - and that should be considered, as well as offering another pace option in conditions which are likely to suit. Here are some of the contenders for the all-rounder spots, with their expected batting and bowling averages, based on my algorithm,:-
Interestingly, of these, Moeen has the worst expected bowling average, and there is little in the way of evidence over the last three years that he can be a successful spinner at Test level. Fellow spinner Samit Patel has better expected bowling data than Moeen, but slightly worse batting data. If Moeen had match practice in the last couple of months, picking him to add batting depth while providing a mediocre spin option would be a reasonable consideration, but it's surely impossible to pick him based on recent performances and inactivity. Looking at the data, the decision has to be between Chris Woakes, and the incredible 42 year old, Darren Stevens, an absolute master of these early-season English conditions. While earlier on in this piece, I advocated that age considerations were not particularly important, it's tough to see the England selectors agreeing with me and picking Stevens over Woakes. Picking Woakes would be fine - at least he's played five matches in the IPL - and has taken 36 wickets at 19.63 in England in Tests since the start of 2016. If he can replicate close to those numbers, he'd have a successful summer, so he's our number eight. Current team: Cook, Stoneman, Burns, Root, Bairstow, Stokes, Foakes, Woakes. Who will we pick for the other three bowling positions? The table below shows the expected bowling average for the main contenders for bowling positions in the squad:-
Quite incredibly, the Warwickshire bowler, Ryan Sidebottom, tops this list. I'm not actually sure if he's England qualified (he has a UK passport, but I'm not quite sure how far that gets him) and while his data at county championship level isn't of the biggest sample size, it's incredibly impressive. It's also worth noting that he took 27 second XI wickets at around 13 runs per wicket for Warwickshire last season, data which translates extremely well to future success, and boosts his sample size. However, he picked up an injury in his last match, but Sidebottom is certainly one to keep a very close eye on indeed. Jimmy Anderson comes next, and is an absolute certainly to be picked, and then it comes to two uncapped bowlers, Coad and Porter. Coad has been extremely impressive in his outings for Yorkshire, but has only played 18 first-class matches at the age of 24. If he carries on this progress, he'd be capable of making a big impression in Test cricket, particularly in English conditions. I prefer the bigger sample size of Porter, who has taken wickets consistently in English conditions (50 in 2015 and 55 in 2016, both in Division Two) as well as a stellar 75 in Division One last season, and 14 at under 20 runs per wicket so far this season. For these consistently excellent seasons, he deserves his crack at the team, and I'd stick him in. As with the batsmen, the rest of the bowlers on the list probably fit the 'good but not great' bracket, and at the present time, I'd probably stick with the experienced Stuart Broad, given that he can bat a bit better than the rest, and has had an excellent start to the season, taking 10 wickets at 14.70 so far. The likes of Sam Cook and Jake Ball are also very close to selection, and ensure that Broad will need to perform well to avoid looking over his shoulder, while the spinners, Amar Virdi and Jack Leach, should be considered more for the away tours this winter. Virdi looks an extremely exciting prospect, and is another whose second XI data made it virtually inevitable that he'd be a success in the first-team. So that's our team. It's not quite the 'best' team on data alone, but it's certainly very strong and pretty realistic - potentially not hugely dissimilar to the likely team. It has excellent batting depth - critical for England, who have been prone to collapsing in recent years, and has bowlers capable of exploiting the likely conditions. In addition, it will be interesting to see how much weight the selectors give to the final round of matches before Tuesday's squad announcement, and these start tomorrow. I would urge caution against doing so, unless a standout performer was already on the verge of selection. All told, I expect the selectors to stick with Malan - rightly or wrongly - and perhaps they won't give Foakes or Porter a go, but if performances in county cricket have any bearing on Test team selection, they certainly should. My final team: Cook, Stoneman, Burns, Root, Bairstow, Stokes, Foakes, Woakes, Broad, Anderson, Porter. If this article has given you insight into the data that Sports Analytics Advantage can offer cricket teams around the world in formulating draft or auction plans, selection strategies or tactics, please feel free to chat to us at sportsanalyticsadvantage@gmail.com. |