1st March, 2018. Email: Sportsanalyticsadvantage@gmail.com In recent weeks, T20 player rankings have been becoming more high-profile, with Rashid Khan becoming the first Afghanistan player (as well as the first to achieve this feat largely from their performances in associate matches) to top the official ICC T20 bowling rankings, which can be viewed here. In addition, the T20 Player Performance Index - created by FICA and The Cricketer magazine - was also released recently, and you can check this out here. Firstly, I'd like to give some observations about the two ranking systems. ICC T20 rankings: Split between batsmen, bowlers and all-rounders - A definite positive, as it's very fair to split players by their playing role. Appears to overvalue performances against associate teams - Can only be the case when Hamilton Masakadza is ranked 11th in the batting rankings, despite having only played six completed innings against traditional Test-playing nations since January 2016. Shaiman Anwar is also in the top 20, from the United Arab Emirates, while Hamza Hotak and Mohammad Naveed are further associate players in the top 20 bowler list. Appears to overvalue batting average compared to strike rate - Babar Azam was recently ranked number one, and is the current number three - despite striking at a mediocre 122.50 strike rate in T20 internationals since January 2016. Only takes into account T20 internationals - Relies on a player to be selected by their national team to be included in the rankings, a selection process that I have discussed as being flawed in numerous previous articles. T20 Player Performance Index: Ranks players by their overall performance - As opposed to splitting players by their playing role. Appears to overvalue all-rounders - Eight of their top ten players have taken at least five wickets in T20 in the last three years, and seven of their top ten can be considered genuine all-rounders in the format. Unclear as to which tournaments they include - Do they include English Second XI data, or Hong Kong Blitz data, for example? It is my firm belief that a player's ability can be best graded from the biggest possible sample of data, ideally as recently as possible. This means that whatever data I can possibly obtain on a player - no matter how weak the competition or opponent - I'll include it. The critical facet is being able to weight that data correctly. As an example, I'll include data from the English Second XI, but I'll be able to accurately translate that to higher standard competitions, even to the extent of the T20 international level. The same can be said of say, the Tamil Nadu Premier League in India, enabling me to assess whether players are likely to be successful in the IPL. Certainly, anyone who analysed Washington Sundar's performances in this competition, prior to last year's IPL, would have been less surprised at his impact in IPL 2017. As I discussed in my recent Strike Rate Superstars article, including as much data as possible has a clear advantage in that it allows me to find extremely low profile players who haven't previously been able to get much of a chance in higher-profile events, such as Sean Dickson, Zak Crawley and Tom Banton (in the T20 Blast, in England) and also add depth to the samples of players who haven't got a large sample of data in the higher-level of competition, such as Calvin Dickinson and Ed Pollock. Some other low-profile English players will be mentioned in this article, as well, including two very highly-rated brothers. Top T20 Batsmen, 1/3/18 At the end of February 2018, the unique Sports Analytics Advantage T20 international algorithm assessed 69 current batsmen (based on their data over the last several years) to be greater than 10% better than the average Test-playing nation T20 international player (average of 23.24, strike rate 132.30), and they are listed below, sorted by a 45-55% split between their average and strike rate mean deviation:-
While there are a number of big names in the top 10, it may be a surprise to see Farhaan Behardien and Heinrich Klaasen in particular included in this bracket, but both have superb domestic numbers which translate well to T20 international success. Certainly, the rest of the top 10 would fit the 'eye test' - they really are the elite of either cricket, or T20 in isolation. Interestingly, and adding further weight to criticism of the selection process of numerous countries, there are a number of players listed who don't even feature regularly in squads of their respective countries , including Stiaan van Zyl (Kolpak deal), Rishabh Pant, Daniel Bell-Drummond, Rassie van der Dussen, Krunal Pandya, Chris Nash, Joe Denly, Mayank Agarwal, Paul Coughlin, Riki Wessels, Richard Levi, Sohail Akhtar and Patrick Kruger - who was actually dropped by his Ramslam team after several matches of the recent competition (again, illustrating why teams should not take such decisions based on the short-term variance associated with several performances). It may come as little surprise to regular readers that England's selectors don't particularly come out of this with much credit - of the nine English players in the list, only Jos Buttler is a regular member of the T20 set-up. Several players who many would expect to be included in the list were not, and here are some examples and reasons why:- Brendon McCullum - Mean Deviation 1.07. Suffered from a poor 2017 BPL and PSL. Alex Carey - Would have comfortably made the list, but was ruled out on sample size grounds. Baba Aparajith, Mohammad Taka - Have not played a major domestic league in recent years. Would have been included otherwise, on their records in lower-profile Indian T20 events. Mukhtar Ahmed - Have not played a major domestic league in recent years. Would have been included otherwise, on their records in lower-profile Pakistan T20 events. Babar Azam - 112 expected T20 international strike rate pushed him below the required level for entry. Kane Williamson - 114 expected T20 international strike rate pushed him below the required level for entry. Top T20 Bowlers, 1/3/18 At the end of February 2018, the unique Sports Analytics Advantage T20 international algorithm assessed 59 current bowlers (based on their data over the last several years) to be greater than 10% better than the average Test-playing nation T20 international player (average of 26.04, economy rate 8.23), and they are listed below, sorted by a 55-45% split between their average and strike rate mean deviation:-
While there were a few surprise names in the batsmen's list, there were really no completely left-field inclusions. However, the bowler's list was somewhat different, with Matt Parkinson, Tom Helm, Callum Parkinson, Michael Hogan and Oliver Hannon-Dalby all lower-profile T20 Blast players whose data was good enough to include them in the top 20. Benny Howell's talents need no introduction to regular readers. Other lower profile players in the top bowlers list include Ankit Rajpoot, from India, Nazmul Islam 'Apu', from Bangladesh, and Mohammad Asghar, from Pakistan. With many English bowlers in the list, it really does appear that not just the national selectors, but also T20 Blast coaches, are failing to understand the merits of T20 specialists - instead being content to pick players who are regulars in other formats, leaving players best suited to the shortest form to be languishing in the second XI for large parts of the season. Of the Parkinson brothers, and Tom Helm, all three are huge prospects and deserve much more game time in T20 events, with small Blast sample sizes backed up by completely stellar Second XI numbers - one more quality Blast competition and they will surely be on the radar of astute overseas franchises. It's also fair to suggest that the T20 Blast - in its current guise - lacks the potential to give players exposure. With a huge number of matches not televised, many players are frequently shown on TV just several times across the whole season, and as anyone involved with statistics will tell you, that puts players at the mercy of variance. It's easy to envisage a scenario where a world-class player fails several times in a row when on TV, and therefore viewers around the world - including decision-makers at franchises, and their own national team selectors - are less aware of their talents than they should be. Finally, several players did not make this list due to small sample sizes, and both would have been in the top five if sample size was bigger - Mitchell Starc (1.58 mean deviation) and Mark Wood (1.45). Regarding all-rounders, given the skill-set that is required to be a top batsman or bowler in isolation, few qualify solely as a batsman or bowler in these rankings. However, we will be featuring analysis of the top all-rounders in the world, using several different metrics, in another article in the near future. It is obvious from some of the players featured being rather low-profile, that there are a number of players who would be something of a 'Moneyball' selection for franchises. If this article has given you insight into the data that Sports Analytics Advantage can offer cricket franchises around the world in formulating draft or auction plans, please feel free to enquire for bespoke draft and auction strategies via sportsanalyticsadvantage@gmail.com. |