15th October, 2017.(Cross posted from my other site, sportsanalyticsadvantage.com) For my next article, I wanted to look at the world's foremost T20 batsmen, with a view to establishing the relationship between boundary hitting and strike rate, and also looking towards highlighting poor recruitment from numerous worldwide franchises - not to mention questionable national team selection decisions. This is as much the case in the upcoming Bangladesh Premier League as ever - there were a number of bizarre recruitment decisions from franchises, particularly regarding overseas players, with a number of players left unsold who, statistically, look to be able to do a much better job than those signed. Having said this, I would commend the recruitment of some Afghanistan players from BPL franchises - batsman Najibullah Zadran and all-rounder Mohammad Nabi are excellent boundary hitters who can maintain a high strike rate while Zahir Khan - who looks a superb prospect - and of course, Rashid Khan, offer quality with the ball. It's refreshing to see some more unheralded names getting chances as opposed to some mediocre names on the T20 merry-go-round. In the research for this piece, I looked at players who faced at least 500 deliveries in major domestic T20 leagues (Big Bash, IPL, T20 Blast, BPL, CPL, PSL, South Africa and the New Zealand Super Smash) and T20 internationals from the 2015 Big Bash onwards - effectively the last three years - and in total, 148 players had reached this landmark. 19 of these 148 players achieved strike rates in excess of 150.00, and they are listed below. These players really should be mandatorily discussed by franchises around the world as viable recruitment options.
Some of these players don't have the best averages, but with good recruitment, that's not necessarily critical - with the number of players available to choose from in T20 drafts and auctions, there really isn't any excuse for a team not to be able to bat down to 9 or 10 in the batting order, and it would be relatively straightforward to organise a squad who had the potential to bat at a high strike rate down to these positions, without particularly weakening bowling options. In the future, teams will realise that striking at a rate below 130 (and possibly higher benchmarks) is totally unacceptable in T20 cricket, and bowlers who can hit regular boundaries and strike at a high rate in cameo innings will become increasingly valuable. With each ball being 0.83% of a team's batting resources in an innings, even an innings similar to 6 (10) costs a team almost seven runs, given that the mean strike rate across the major domestic T20 leagues and T20 internationals combined was 127.18. Therefore, having a squad who can maintain consistently high strike rates without playing too many dot balls is critical. We can see from the table above that not only do these players boast strike rates in excess of 150, they also have high boundary percentages - only four of the 19 players have boundary hitting percentages below 20% - and even the lowest of these 19, Mohammad Nabi, at 18.20%, is still 2.57% above the mean boundary hitting percentage across the major domestic T20 leagues and T20 internationals combined of 15.63%. Given this, and the obvious logic behind such a statement, we can clearly see that there's decent evidence of a relationship between a batsman's strike rate and their boundary hitting percentage, and this is evident from the graph below also:- Here we can see that generally, as the strike rate (horizontal axis) increases, so does the boundary hitting percentage (vertical axis). However, there were also a number of anomalies, and it's worth breaking down these anomalies into some sub-categories, in an attempt to label batsmen, and also to work out their strengths and weaknesses. This was done by working out the deviation from the mean for both strike rate and boundary hitting - as I mentioned earlier, the mean strike rate was 127.18, and boundary hitting 15.63%, so using Andre Russell as an example, his strike rate mean deviation would be (166.08/127.18) = 1.31, while his boundary hitting mean deviation would be (23.96/15.63) = 1.53. This would then allow us to work out the difference between the strike rate and boundary hitting deviation for all players, with a negative figure meaning that the player is more boundary hitting orientated than strike rate orientated, while a positive figure indicates that a player is more dominated by their strike rate than their boundary hitting. The most boundary orientated players are listed below:-
These players all had mean deviation differences of over -0.20, meaning that they were extremely weighted towards hitting boundaries as opposed to maintaining a strike rate. Of course, all with the exception of Dwayne Smith also have strike rates above the mean, but the fact that these players have considerably bigger boundary mean deviation than strike rate also indicates that a very high proportion of their balls faced either end in dot balls or boundaries. Improving the majority of these players will be quite tough. It's obvious that their tolerance of risk is pretty high, given their propensity to hit boundaries, but perhaps these players could focus more on nicking singles as opposed to playing dot balls. Conversely, there were a number of players who had high positive figures for mean deviation, meaning that their strike rate mean deviation was better than their boundary hitting mean deviation. In short, this should mean that they are more strike-rotators, as opposed to boundary hitters. Those with a mean deviation difference of 0.15 or higher are listed below:-
What is immediately apparent from looking at this table is that none of these 17 players have a strike rate in excess of 130, and only Moises Henriques, Paul Collingwood, JP Duminy and Manish Pandey have strike rates better than average. Therefore, while it's reasonable to suggest that despite these players seemingly being quite adept at turning dots into singles, this is an overall worse strategy than dots and boundaries, as evidenced by the significantly lower strike rates. All of these players have boundary hitting percentages over 2% worse than the mean. "In a country where one-day cricket is king, big-hitting batsmen are seen as risky flashes in the pan and without their body of work properly evaluated. After all, this is where a six followed by a dead defensive block earns serious praise for sensible batting. So when a Pappu, Ezaz or Hasan hammers a six, a club official shifts in his seat. If he plays and misses the next ball, he swears under his breath. And if the big hitter gets out caught in the deep, at nearly any point in an innings, he will earn the official's wrath. The coach will then explain to him how he should have played it straight, with his head down." In fact, in the course of research, I noticed that it is fairly obvious that this isn't just a Bangladesh problem, but also faces a number of players from Pakistan, and despite their success in T20s, also some West Indian players - in fact, an analysis of West Indian players would probably conclude that their approach is either extreme risk or extreme caution, with there being little middle ground. The following players had boundary hitting percentages below 13.13 (2.5% below the mean):-
Three mainstays of the Bangladesh national team, Shakib Al Hasan, Mushfiqur Rahim and Mahmudullah, are included, while there are a number of mediocre English players in the list (probably due to the bigger number of players in the 18-team T20 Blast, as opposed to any other inherent reason). The players above, unless they offer considerable value with the ball, such as Benny Howell, should be avoided by franchises. Given that 54 players of the 158 who have faced at least 500 deliveries in this time period have boundary hitting below the average, and 60 of the 158 have strike rates below the mean, it is clear that many regular T20 batsmen either do not possess the required skills to succeed in the format, or need a considerable mentality shift. It truly is bizarre how franchises and national teams around the world select many of these players despite their obvious limitations, and the numerous better options left unsigned. Please feel free to enquire for further details on player analysis, and bespoke draft or auction strategies at sportsanalyticsadvantage@gmail.com. |