With the CPL reaching the end of the group stages, I thought it would be interesting to take a look at some player and team data for the tournament so far. Teams would probably benefit from knowing some of this, so if any readers know any people at CPL teams, please feel free to link them this article.
The first thing I want to discuss is the situation with the pitches in the tournament. Every match has been played at St. Kitts, with matches seemingly being split between just two pitches. The benefit which this should have for spinners, in theory, should be huge, and I’ll go on to talk about that a little later on.
It won’t have been easy in general for batters to have a good time on these pitches. We’ve seen some low scores defended with ease, and teams regularly collapsing. Smart teams (and recruiters) in other leagues will be able to see past the basic data and apply value added/subtracted adjustments to it, although I don’t expect many teams will do this. IPL teams, for example, could easily overvalue performances by spinners in this competition when it comes to the mega auction later this year.
In my last Substack, I promised that this CPL would be data-heavy, so here goes. Most analysis will be in charts, just because it looks a bit nicer than spreadsheets full of data. First up, let’s look at scoring data for batters (minimum 100 balls faced):-
Here we can see that only a few players were close to maximising their scoring options, being in the top-right hand corner. Faf du Plessis, after a slow start, got there, as did Roston Chase, who has undergone something of a transformation in his T20 career of late - whether that is sustainable is another debate entirely. Tim David, who probably still just about qualifies as a ‘hipster’s choice’ in recruitment debates but is rapidly becoming a player who people have realised adds plenty of value, was also in this area.
The likes of Nicholas Pooran, Kieron Pollard, and the two Lewis’ - Kennar and Evin - also had very high boundary percentages with pretty low non-boundary strike rates, but those non-boundary strike rates are more acceptable when players are so good at boundary-hitting.
Unfortunately, that wasn’t the case for players like Dinesh Ramdin - who was a really bizarre pick-up by TKR based on historical data - and Shoaib Malik, who has one of the lowest boundary percentages of any batter I can remember in any tournament facing 100+ balls. A few other batters who have arguably disappointed based on pre-tournament expectations include Rovman Powell, Tim Seifert, Colin Munro, Mohammad Hafeez and Shimron Hetmyer. Munro apart, I didn’t have absurdly high expected data for these players - Hetmyer is competent but expected data suggests he’s not an expensive IPL player, which actually he currently is - but plenty of casual watches will have had higher expectations for these players.
With scoring options out of the way, I want to have a look at another way of looking at batter output - balls per dismissal versus boundary percentage:-
In the vast majority of cases, batters trade off one of these metrics against the other - they either hit well or they have high stability. Having both, over a large sample of data, suggests world-class capability (e.g. AB De Villiers) and of course, having neither suggests a real problem for that player!
The likes of Chase, Evin Lewis, Pollard, Pooran, David and Du Plessis again rate well here - these six players look like the best six batters in the tournament - and makes the decision from Guyana Amazon Warriors to bat Pooran at seven on occasion look even more bizarre. Here is a guy who is an excellent player of spin who you want to face enough balls to make a match-winning contribution. My advice for Guyana would be to bat Pooran first wicket down after the Powerplay and don’t even think twice about it.
The likes of Malik, Ramdin, Hetmyer and Powell again struggled here - it’s probably fair to say that this CPL shouldn’t have enhanced their reputation - and this starts to give us some insight into how T20 teams should look at constructing their roster. All this stuff can be modelled in advance of tournaments, and I’ve done so for various leagues, so there really is no need for mediocre, subjective recruitment decisions - a lot of the players who struggled were very predictable, and likewise, those who succeeded, with the exception of Chase.
Moving on to the bowlers, I want to look at one area in particular - run prevention. As I’ve written about a lot of times before & discussed on social media, boundary prevention is a critical metric in terms of being able to succeed as a T20 bowler, so the chart below compares dot ball percentage and boundary conceded percentage for all bowlers bowling 100+ balls:-
Ideally, as a bowler you want to be in the bottom-right corner and three players featured predominantly here - all spinners (who are highlighted in red print). Sunil Narine, Akeal Hosein and Mohammad Hafeez (who made up for his mediocre batting output with his bowling) leading the way. Interestingly, this is a pretty unusual dynamic for spinners, who tend to be grouped towards the bottom-left corner more (low dot ball %, low boundary percentage % conceded). This is primarily because spinners tend to operate in the lower scoring middle overs where batters have less boundary intent and try to rotate the strike, but it seems here judging by the high dot % for a number of spinners (also featuring Jeavor Royal, Imad Wasim, Roston Chase and Imran Tahir) that many of the CPL batters were very poor at rotating spinners - this should be something which CPL recruiters should look towards with their future recruitment, finding good players of spin where possible.
Moving on to team data, understanding whether teams deserved what they got is always useful in a 10-match group stage, where variance can play a part to some extent. Net boundary percentage plays a huge part here - we know that around 85% of teams winning the net boundary percentage in a match win the actual match, and there’s a similar correlation with tournament net boundary percentage - for example, the top 7 net boundary % teams in the T20 Blast this year all qualified. Here’s how it looked for the CPL:-
The table was largely replicated by this net boundary % table - only Tallawahs and Kings were the other way around in the standings - and there’s probably some argument to suggest that Tallawahs were slightly unfortunate not to qualify, and conversely, Kings were perhaps a little fortunate to do so. Knight Riders and Warriors look best-placed to make the final, as the league standings also illustrate.
An overview of the batting versus bowling dynamic for teams is also useful to look at, and this is illustrated in the chart below, which looks at batting boundary percentage and bowling boundary percentage for each team in the CPL group stages this season:-
No team were particularly strong in both areas, with all teams having strengths and weaknesses in one way or another, except for Barbados Royals who were below-average in both areas - as evidenced by having the worst net boundary percentage and coming bottom of the group stage.
Following this, I want to run through comparison of a number of batting metrics for CPL teams, to show where teams performed well or badly. The chart below shows team data for this season’s CPL group stages for teams facing pace and spin:-
If you ever want evidence of the effect of playing on two used pitches throughout the tournament, here it is. Only Kings managed to strike at 106+ versus spin in the entire tournament, with all of the other five CPL franchises really struggling against spinners. If there is one take-away from this tournament for recruiters in other T20 leagues it is this - be extremely careful not to over-value spinners based on their performance in this year’s CPL in isolation. This is about as easy bowling conditions for a spinner that you’ll ever see (possibly with the exception of Bangladesh T20 internationals in recent months!).
Understanding the scoring dynamics of each team is also pretty useful to quantify, in order to assess their batting groups. The chart below looks at scoring ability (strike rate) versus balls per dismissal:-
This chart is pretty interesting as it illustrates quite how bad CPL teams were with non-boundary scoring (rotating the strike, and turning ones into twos). In many leagues, around 70 is ballpark average for non-boundary strike rate, yet no CPL team came remotely close to this. Kings were the only non-disastrous team in this area, and this league dynamic may also be a consequence of the used pitches and dominance from spinners (as spinners tend to concede a higher proportion of non-boundary runs than pacers). However, it also gives some pretty decent insight into how teams scored their runs and both Knight Riders and Royals in particular need to consider numerous changes (read: an overhaul) to their batting group moving forward in subsequent editions of the CPL.
As I said previously, most players face a trade off between attacking (boundary-hitting, reflected via the direct relationship it has with strike rate) and stability (balls per dismissal), and it’s no different for teams. A team who can manage both aspects will almost always do extremely well. The chart below looks at team scoring ability (strike rate) versus balls per dismissal:-
All teams struggled from a strike rate perspective, although relatively speaking, Kings were the best here - closest to the top-right corner. However it must be said that even their batting metrics would rate extremely poorly in many other T20 leagues, again illustrating how the CPL was an easy bowling/difficult batting league this year. Knight Riders, Royals and Warriors really struggled in particular and need to ask themselves serious questions about their batting recruitment.
Because this is getting pretty long, I’ll skip a few phase performance charts and just publish the raw data instead, looking at the Powerplay/middle/death over performance for each team’s batting group via strike rate and balls per wicket lost:-
St. Lucia Kings had far more intent than other teams in the Powerplay, with a few quick starts given to them by Faf du Plessis and to a lesser extent, Andre Fletcher. Trinbago Knight Riders had the opposite dynamic (similar to a lot of IPL teams) where they try and preserve wickets until the death overs and then look to hit out - this is a very ‘old school’ strategy and despite the high balls per wicket lost in their innings, something of a high variance strategy. It’s not a methodology I particularly advocate. Barbados Royals’ issues with the bat were adequately illustrated with mediocre phase performance throughout.
Getting on to bowling, we’ve already discussed the benefits of bowling a lot of spin on the used pitches of St. Kitts, so let’s see which teams worked that out better than others:-
Guyana Amazon Warriors have built up a reputation in previous editions of the CPL of being a very spin-orientated team, and even at a non-home venue, that was again their game plan, with Imran Tahir in particular excelling. Given the spin bias at the St. Kitts venue, it’s actually pretty surprising to see that Guyana were the only team to bowl in excess of 50% of their balls via spin - did other teams (particularly St. Kitts & Nevis Patriots and Barbados Royals) miss a trick by going for pace-heavy bowling strategies? It looks like it.
That spin bowling bias is again shown when looking at the pace economy versus spin economy for each team in the CPL group stages:-
Warriors, Tallawahs and Knight Riders had spin groups a class above the other three teams, while Knight Riders spin economy - led by Sunil Narine and Akeal Hosein, remember those two bowlers who had a high dot ball % and low boundary % conceded earlier in this article - was the best in the entire group stages by some distance. There’s plenty of work to do for Patriots and Royals, and Royals in particular look like they just need a complete overhaul, both via their batting and bowling.
Finally, as with the batting data, here’s a look at the phase numbers for each team’s bowling group during the CPL group stages:-
The bowling group of Trinbago Knight Riders excelled in all phases, and was a big part of them winning the group stages and also the net boundary percentage count, as discussed previously. Essentially, they got away with a questionable, and dated batting strategy because their bowlers were so good. Tallawahs, Kings and Royals again, were the worst two bowling groups, with that being part of the reason for Tallawahs and Royals finishing in the bottom two league positions and having poor net boundary percentages.
I hope you enjoyed this deep dive into the group stages of the CPL, and I’ll be returning over the coming days in advance of the IPL with a detailed look at the replacement players added to each squad - it’s fair to say that some IPL teams look to be far better at recruitment than others!
Will you be posting Expected IPL batting averages and strike rates for this year?