NEW - Weekly Cricket Analytics & Discussion Blog

Email: Sportsanalyticsadvantage@gmail.com

If these articles has given you insight into the data that Sports Analytics Advantage can offer cricket teams around the world in formulating team strategies, selection, draft or auction plans, or any other work, please feel free to get in touch at sportsanalyticsadvantage@gmail.com.


As we move towards the new year, I have started a new weekly cricket analytics and discussion blog, with regular posts throughout 2021.  I wanted to start this for a number of reasons - firstly, it's sometimes very difficult to concisely give my thoughts on matches and situations in matches on Twitter, which obviously has a character limit, and also because I get a lot of contact from aspiring analysts around the world both via email and also on Linkedin, so I thought it would be interesting and useful to give some thoughts on my personal experiences of trying to get work in cricket and the industry in general as the blog progresses.  Also, I'm happy to answer any questions or give my thoughts on specific match situations, so please feel free get in touch - if there are enough questions, perhaps I'll do a Q&A on here at some point in the future.  As the posts continue throughout 2021, they will be added below and all posts can be viewed via this page with the most recent at the top.

If you enjoy my content, please feel free to buy me a coffee - it would be much appreciated!
1st April, 2021.

Cricket teams need to be more evidence-driven

In the previous post (below), we discussed how perceptions of player often vary from their data, and I want to expand a little further on that discussion with this piece, but with a focus on benefits of success, and implications of failure.

As I've mentioned in my writing and on Twitter a number of times before, decision-makers in a team - be it in cricket, football, whatever sport where recruitment is involved - need to be accountable for their recruitment to their bosses (General Manager, CEO, Owner, etc).  If they can't explain why they recruited a certain player, and give evidence-based rationale as to why that recruitment took place, then in my view they've failed at their job, it's that simple.

Cricket has a lot to learn in this respect.  Even many teams who are mentioned as being 'data-driven' often appear to use data when it suits them and hunches when it suits them.  Being data-driven isn't a sometimes thing.  It should be the bedrock, the foundation of recruitment strategy.  Once you determine a shortlist of players from data who are suitable to fill a pre-identified gap in a squad, then you can do the more subjective due diligence - such as finding out about their personality, video analysis, and so on.

Some teams understanding age curve dynamics better than others

While I don't think football is even close to running at maximum recruitment efficiency, I do perceive it to be far more advanced than cricket.  One area where they are more advanced is understanding the economics of legend players getting towards of past peak age, as I tweeted earlier this week:-


CSK with the most reliance on older players by some distance

However, just as with football, some cricket teams are better off at dealing with this issue than others.  A snapshot of the current IPL squads illustrates this (ages of players correct at 1/4/21), with the chart below illustrating the number of players at each franchise aged 34+:-


There is a clear 'leader' in this area - Chennai Super Kings with seven players aged 34+.  They also lead the way in the financial figures, with the chart below showing the Crore value spent on players aged 34+:-

Teams should follow the MI model

As you may have noticed, Mumbai Indians have no players aged 34+.  Many people perceive MI to have the best recruitment strategy and processes of the eight IPL teams, and I agree with this.  By a long distance.  A pretty good rule of thumb for rival IPL teams would be to assume that if MI are doing something (or not doing something) then it's probably a good idea.  All teams should try and replicate MI's strategy, but as is the case with England in ODIs, teams appear unwilling to take a similar approach and breaking 'conventional wisdom'.  MI and England perform their strategy in plain sight, yet teams are extremely reticent to try and adopt these strategies themselves.

Greater rewards for success and implications for failure needed

One reason why I think teams don't look to perform strong strategies with high positive expectation is that in franchise T20 leagues, there really isn't a great deal of meritocracy.  There aren't many incentives for success, or implications for failure.  Coaches routinely get new jobs despite poor records because they have 50+ caps for their country.  Lower-profile coaches with good records and good processes can't get marquee jobs.  From personal experience, I can tell you that most teams aren't interested in specific recruitment analysts.  

Why would an owner want someone to tell them that they can find a better player for half the price, but it removes the opportunity to have a selfie or dinner with a superstar?  
Why would a coach or captain want someone interfering with their ability to pick their mates?

I can guarantee you that owners mentalities would change overnight if these T20 franchise leagues brought in huge prize money for them if they qualified or won tournaments.  Extending this also, a better reintroduction of the Champions League would be fantastic too.   It happens in football, and it needs to happen in franchise T20.   Immediately, there would be rewards for success.

Furthermore, there would also be an intolerance of failure.  If an owner saw a rival owner earn, say, $20m for winning the IPL, and another rival owner got $10m for coming in the top four, they would want to know why their coach didn't deliver, and would almost certainly question their processes.  I would also like to see relegation - the ultimate implication of failure.  For example, a two-tier IPL with 12-16 teams split into two divisions, with the ability to play 5-6 overseas players per team, would be incredible.  Perhaps it would take relegation for some teams to rip up the script and start adopting better, evidence-based processes.

T20 franchise leagues need to develop themselves by having better rewards for success and greater implications for failure.  This must happen for these leagues to fully develop to the maximum extent.


16th March, 2021.

Imagine a world where, before an auction or draft for a T20 league, each player's name on the draft list was simply replaced by a random number.  Next to the random number was detailed data on each player, spanning various time-frames.  How different do you think the squads would look compared to how they currently appear?  

My opinion is that for most teams, they would look markedly different - probably unrecognisable to most squads which are picked.  Now think of international teams - would they also look similar to the squads picked on a regular basis?  Again, I'm sceptical that they would be.  I've often thought - with data-driven evidence - that reputation (among other factors) frequently skews recruitment and selection, so I thought I'd try a little experiment on Twitter...

Over the last few days, I posted the data for three hypothetical players on Twitter and set up polls which asked voters to value the player for next year's IPL major auction.  As many may have suspected, these players were actually real players, and the data used were their ballpark numbers from 2019 onwards (at the point that the polls were tweeted) across T20 leagues and T20 internationals.

The three players were Virat Kohli, Ben Stokes and Rahmanullah Gurbaz.  You may be wondering why I picked those three players, and I'll explain here now.  The first two players are players with extremely high all-format reputations, and have commanded big salaries in the IPL in recent years.  However, there are arguments as to whether their performance output in T20 justifies these reputations and salaries, so I thought it would be interesting to see whether people felt their data was positive or negative without taking into account their name (and therefore, reputation).  The reason why I picked Gurbaz was because he went unsold at a low base price at the IPL auction recently despite having pretty strong abilities in recent years as an ultra-aggressive opener and wicket-keeper.

I want to make something very clear at this stage.  The entire point of this piece is not to assess whether the players are good or bad, but simply to try and create a discussion as to the extent which a player's reputation (which can also be generated from other formats to T20) affects their market value in auctions and drafts and whether this information can improve team's processes.

Virat Kohli:-


This was Kohli's ballpark data from the start of 2019 to the end of the 1st T20 against England last Friday.  The data illustrates a player who has elite-level death-hitting strike rates, but struggles to generate a high strike rate in the two phases in advance of this.  Against pace, he strikes very well, albeit without a stellar balls per dismissal figure, while against spin he's ultra stable with a magnificent balls per dismissal figure, but has some issues with his strike rate.  

Kohli is an elite rotator and runner between the wickets, as evidenced by his non-boundary strike-rate of 83, but his boundary percentage around 14% will see him struggle to generate many innings with high strike rates - it's virtually impossible to hit a strike rate of 140+ without being able to hit 16% boundaries at a bare minimum.  His four to six ratio of 2.04 (he hits 2.04 fours per six hit) isn't bad, but it's not top level either.  

This data suggests Kohli is a clear stability-driven batsman, and an extremely reliable anchor.  However, given the strike-rate bias towards pace bowling, there would also be an argument to suggest that he'd be a pretty decent opener.  Coming in at a stage when the spinners start bowling a high percentage of overs could be a major strike rate issue and potentially a contributory factor as to why Royal Challengers Bangalore had a mediocre middle overs team strike rate in the 2020 IPL.

Is he worth 10+ Crore at auction?  Not many people, without knowing who Kohli was, thought so in the poll:-


Surprisingly, more people suggested a player with Kohli's data (without knowing who he was) should go unsold as opposed to going for 10+ Crore.  This insight is fascinating, given that it potentially shows how much reputation still influences market dynamics at the IPL auction.

Interestingly, a similar scenario presented itself for Ben Stokes, as we will see below.

Ben Stokes:-


The above is Stokes' ballpark data from the start of 2019 to the end of the second T20i between England and India on Sunday (the time scale is very marginally different to Kohli simply due to the time that the data and poll was tweeted).  Some may argue that the data suggests a batting all-rounder, but one who doesn't necessarily have a top-level strike rate upside, and I think that would be a fair assessment.

With the bat, Stokes' boundary percentage of 17% is solid but fairly unspectacular - certainly not in the Andre Russell or Hardik Pandya bracket, or even the Lewis Gregory (a positional rival for England) bracket.  As with Kohli, he's a very strong rotator of strike (a non-boundary strike rate of 78 is excellent) but the data above suggests he's being misused by England as a finisher given his strike rate in the last four overs of just 157.  Arguably, he'd be better used as an opener, given his strong strike rate in overs 1-6, a better strike rate versus pace compared to spin and also a fairly mediocre six-hitting percentage of 4.65%.  Stokes has a reasonable bias towards hitting fours as opposed to sixes, given a four to six ratio of 2.74 (he hits 2.74 fours for every six).  In the IPL, Rajasthan Royals used Stokes as an opener and I think that was a smart decision.

Bowling-wise, Stokes fits a similar dynamic to Ben Cutting and Marcus Stoinis with the best role being utilised in the middle-overs.  It's difficult to make a case for him being used in other phases given the economy rates detailed above, and it's also tricky to suggest he'd be a regular four-over bowler.  

Overall, in my view, Stokes' numbers suggest a dynamic of a player who would be best utilised as an opener with the bat and bowling several overs in the middle overs of opposition innings.  I would strongly consider opening Stokes and Buttler, if I was involved with England selection.  Rajasthan Royals can do this in the IPL as well.

Without knowing who he is, did people think this data would make him a high-value player at next year's IPL major auction?


As with Kohli, less than 10% of voters thought Stokes' data (without knowing who he was) represented a 10+ Crore player.  28% of respondents thought he would go unsold.   Again, this potentially shows how much reputation still influences market dynamics at the IPL auction although in Stokes' case it is very prudent to make the point that overseas pace all-rounders are a very scarce resource and therefore cost premium prices at auction as we saw with the likes of Chris Morris, and to a lesser extent as all-rounders (but either batters who bowl or bowlers who bat) such as Glenn Maxwell, Kyle Jamieson and Jhye Richardson recently.

Finally, the flip side - Rahmanullah Gurbaz.

Rahmanullah Gurbaz:-


While Gurbaz's data comes from leagues which are arguably lower standard than the IPL, as well as T20 internationals for Afghanistan, it is also very fair to point out that against high level bowlers, his numbers don't drop-off much at all from these.

Based on the above, Gurbaz's average innings is 28(18), which is pretty decent for an opening batsman and he also has the extra skillset of keeping wicket.  His boundary percentage around 25% is elite level - there are very few batsmen in T20 cricket who can beat this - and this is also the case for his six percentage of 10.44%.  His four to six ratio is very low at 1.37, which indicates that he hits a six for every 1.37 fours, and you'll note that this ratio is much lower than Kohli or Stokes, suggesting Gurbaz is a real six-hitter and a player who can contribute very strongly to his team winning the boundary percentage count, which is a huge driver of eventually winning the match.

Gurbaz's non-boundary strike rate of just 52 is pretty mediocre though, but this doesn't really matter with such a high boundary percentage and six percentage - it's a very similar dynamic to the likes of Sunil Narine, Chris Gayle, Andre Russell and Evin Lewis.  Gurbaz also strikes almost identically well against both pace and spin and is a real threat in the Powerplay, striking at 165, and while this drops off in the middle overs to 138, a middle overs strike rate in this region is still well above average.

This is where the experiment gets even more interesting:-


Again, without knowing who the player was, the same percentage of people voting in the poll thought that Gurbaz deserved a 10+ Crore price at the next IPL auction as Virat Kohli did.  More respondents thought that a player of Ben Stokes' data would go unsold than a player of Gurbaz's data.  Gurbaz had the highest priced winning percentage, with 49% of people thinking he would be sold for 5-10 Crore.  At this point, it seems like a good point to remind you that this was a player who went unsold at the recent IPL auction.

The factors concerning why a player is overvalued (or undervalued) are plentiful, and I detailed numerous reasons in an article for The Cricketer magazine previously.   You can read that here.  Despite having written the article almost two years ago to the day, in my view, not much has changed.

In football, things aren't the same.  A big name wouldn't be able to get away with a year or two of below-elite performance output without being under intense scrutiny.  Almost certainly, their market value would fall.  In cricket, matters are different in my view.  Big names are sometimes viewed as untouchable, undroppable and unquestionable.  

Perhaps a big part of this is the media, and potentially in particular, commentators.  It may well be that when analysing a player's level, commentators simply know no better than to simply go on reputation.  This could be a negative side-effect of having numerous commentators who simply are big name ex-players who seemingly do little research on players, instead relying on their own big name and experience from playing in the past.  The problem is, those playing experiences were often in formats which have markedly changed to how the format looks now, or these ex-players didn't play T20 at all.  It should be pointed out this certainly isn't a problem for all ex-players in commentary - some, particularly on Sky Sports, are superb.

The other main complicating issue for commentators in cricket compared to football is the three-format style of cricket.  It's very easy to judge a player on reputation (or indeed, performance level) in Test cricket, for example, and let that influence your opinion of that player's ability level in T20.  If a group of commentators were asked what the benchmark levels were for various phases of a T20 match, or asked what boundary percentage or six percentage was good or bad, I'd be extremely surprised if I didn't get a large variety of answers, some of which are unlikely to be particularly accurate.

Commentators are simply a side issue here though.  The key point that I'm trying to reinforce here is simply that teams need to ignore the noise and do their own due diligence on a player before recruiting and selecting a player.  Basing those decisions on reputation and the opinion (or reaction) of others must be avoided.  Many teams can make considerable positive changes to their processes in this area - be they franchises or national teams - and a more data-driven approach while ignoring the noise will frequently be the foundation to this.


8th March, 2021.

England fail the test of spin in India

Following the Test series between India and England, where 67 England wickets across the four matches fell to spin bowling and just 13 to pace, it seems like a pretty good time to write an article on spin bowling in Test cricket.  The series was always likely to be a test of spin, and this was obviously the case - and a test which England resoundingly failed.

I've discussed England's batting selection policies - as well as their spin policies - already following the Fourth Test.  Anyone interested my thoughts on that can check out The Cricket Podcast here.  It remains my point of view that England need to be more evidence-based with their decisions, and I wouldn't like to be in a position of having to explain to my boss the reasoning for certain recent selections, because I'm not sure I could.  However, that's not the point of this latest blog post - I want to discuss something related, but not altogether linked, and it's a topic which I've never seen covered anywhere else - what is the benefit of playing an off-spinner in Test cricket as opposed to a different spin type?

Spin match-ups must be considered with selection

It's important to me to do make a small point of order here.  Unlike some which others have written, this isn't a post at all aimed towards targeting the inclusion of Dom Bess in the England team - it's much broader than that.  Bess' data will be included in the analysis, and a little later in this article I will be making one data-driven specific point about how there could have been dramatic improvement in the situations in which he was chosen to bowl in during the matches he's played so far in his career.

When a seam-bowling orientated Test team picks a spinner at home - let's say, England, Australia or New Zealand - there doesn't seem to be any particular discussion on the type of spinner picked.  It's just 'a spinner'.  If the type of spinner isn't considered in a selection process, it is my view that the selection process is extremely flawed.  As this piece will demonstrate, there's an extreme difference between picking a right-arm off-spinner and an orthodox slow-left arm spinner.  Both are finger spinners, but in terms of match-ups and the likelihood of success in Test cricket, the two spin types are completely different.

The first piece of analysis I want to illustrate is the difference in bowling averages between right-arm off-spinners (highlighted in red) and slow left-arm spinners (highlighted in black) in Test cricket from 2016 onwards (minimum 25 wickets):-

Right-arm off-spinners with much better numbers against left-handers

Here we can see a clear grouping of the right-arm off-spinners in red towards the left hand side of the chart - stronger averages against left-handed batsmen than right-handed batsmen.  Conversely, we see more players highlighted in black - slow left-armers - towards the right-hand side of the chart, being relatively weak against left-hand batsmen, but stronger against right-handed batsmen.  It's the classic match-up scenario that we very frequently see in T20 cricket as well.

There are also a few slow left-armers with elite data against both types of batsmen, with Shakib Al Hasan and Ravindra Jadeja boasting excellent numbers from a bigger sample than Axar Patel, who is also in the most ideal bottom-left corner.  Likewise, Ravichandran Ashwin - England's tormentor in chief - and perhaps surprisingly, Akila Dananjaya, featuring as right-arm off-spinners with good numbers against both types of batsmen.  From these players I think it's fair to suggest that a spinner - regardless of type - needs to be elite level to average sub-30 against both left and right handed batsmen, which is a feat only a few can manage in this sample. 

As you can see from the chart above, most off-spinners - shown in red - have struggled against right-handed batsmen.  Even Nathan Lyon, who is one of the greatest spinners in his era in Test cricket.  This provides a massive problem when it comes to selection, which brings me back neatly to the original question - what is the benefit of playing an off-spinner in Test cricket as opposed to a different spin type?

Around 70% of batsmen in cricket are right-handed, so there really is little benefit towards picking an off-spinner unless they are elite (Ashwin) or add a decent amount of expected value with the bat over the average bowler (Ashwin again, Moeen Ali, Joe Root).  Utilisation of some strong match-up planning can increase the percentage of balls bowled for right-arm off-spinners against left-handers, as is evidenced by the chart below which shows the balls bowled to left-handers for right-arm off-spinners taking 25+ wickets from 2016 onwards:-

Bess utilised less against left-handers than other regular right-arm off-spinners

Every single one of these right-arm off-spinners bowls less than 45% of their balls bowled against their favourable match-up versus left-handers, showing how difficult it is for right-arm off-spinners in Test cricket.  However, all bowled more than 30% of balls to left-handed batsmen, with one exception - Dom Bess.  Here is a young player who is making his way in Test cricket, who has been under pressure for his place and the subject of intense media scrutiny yet he's been bowled less in his favoured match-up than every single right-arm off-spinner in this entire sample.  I genuinely feel sorry for Bess, because his numbers against left-handers are pretty decent - averaging 29 runs per wicket with an economy rate just below three runs per over.   It's also worth noting that Moeen Ali's usage against left-handers is also pretty low, while Joe Root's percentage of balls bowled against left-handers is markedly higher.  Is the captain bowling himself in more favourable match-up scenarios, for some reason?

This brings me to my next point.  If there's not much need to pick an off-spinner unless they are elite or if they add much more with the bat than the average bowler, why on earth have England been playing two right-arm off-spinners in Root and Bess during the last couple of years?  There's an obvious point that some will say here, that Root is an occasional bowler, but actually, their numbers against left-handed batsmen - their primary match-up target - aren't hugely dissimilar:-

Root a capable right-arm off-spin option for England

The chart above - again showing the right-arm off-spinners taking 25+ Test wickets since 2016 - shows that while Root's average versus left-handers is a little worse than Bess, it's not markedly so and Root offers a better economy rate in this targeted match-up.  Root certainly looks a pretty competent bowler for an 'occasional option' and looks good enough to hold an end up when required, at the very least.

If England need to play an additional spinner to Root, it should be a spinner who matches up well against right-handed batsmen.   As mentioned previously, these are much more plentiful in the player pool - India's top five in the last Test were all right-handers, for example - and therefore Jack Leach should be at the front of the queue.  If there are situations which dictate playing multiple specialist spinners, England should strongly consider playing Leach plus either another slow left-armer or a leg-spinner.  There is little benefit in England playing an off-spinner when they have Root unless the opposition's top order is extremely left-hander heavy.

Positive match-ups more plentiful for slow left-armers

Finally, to illustrate the more plentiful better match-ups for slow left-arm spinners, the chart below shows the percentage of balls bowled for these bowlers against right-handed batsmen by slow left-armers taking 25+ wickets in Test cricket from 2016 onwards:-

Here we can see that that all slow left-armers bowl a minimum of 65% of balls against right-handed batsmen, and many bowl much more.  This is simply because right-handers are more plentiful in the player pool as discussed above, and this provides the ultimate explanation as to why a specialist right-arm off-spinner isn't needed in a team unless they are elite or add much more with the bat than the average bowler.  Why would a team want to pick an off-spinner, who bowls around 30-35% of balls against their ideal match-up, when they can pick a slow left-armer who will usually bowl around 75% of balls or more against their ideal match-up?

For all their issues with the batting group, if I was able to ask the England selectors one question, it would be simple - why on earth do you play a right-arm off-spinner in your team when you already have Root? 


27th February, 2021.

Every day is a learning day

As the old phrase goes, every day is a learning day.  In a work environment, it's a mantra I take pretty seriously, because in my view, it's the only way that someone can get to the top of their profession.  I'm always reading articles, listening to podcasts, and not just about cricket either - I also look at a lot of other sports and look at the parallels between them and cricket.  I try to understand how they might approach similar problems, and both baseball and in particular, football, are two sports that I find very useful in this respect.

One person who I try and listen to and read about as much as possible is Stuart Webber, who is Sporting Director at Norwich City Football Club, and he's one of a number of Sporting Directors/Directors of Football who I think have very strong processes, vision and an ethos geared towards achieving long-term success at their respective clubs.  Given this, it was fantastic to have the opportunity to chat with Stuart yesterday with our Leicestershire management group to try and learn more about best processes in football and how we can then take those learnings back to cricket.  

Does Ashwin's tweet hint at the recency bias which affects drafts and auctions?

At this point, if you've clicked on the Twitter link, you're probably wondering what this all has to do with Devon Conway, and that's fair enough.  However, if you listen to this interview given by Stuart several years ago at https://www.youtube.com/watch?v=jw5WL5AMDq0 then you might begin to understand where I'm going with this.  The entire interview is great - particularly the part about why high-profile ex-players shouldn't walk into management jobs - but the particular segment applicable to this discussion starts at around 32 minutes onwards.  In this, he talks about League Two teams not willing to take Todd Cantwell on loan several years ago because they weren't willing to take the risk on a 'soft' and 'inexperienced' young player.  I think the vast majority of football observers would now agree that Todd Cantwell, as one of the best young players currently in the Championship, probably has a £20m+ value at the bare minimum.  Stuart then discussed Phillip Billing, who also has commanded a sizeable transfer fee since the interview was released.  He couldn't get him a loan to a League Two team, yet three months later he scored the winner in a Championship match and all these teams who rejected him previously were straight on the phone asking to take him.  Stuart's point was that Billing, as a player, 'hasn't changed', yet because of this greater exposure, these teams who didn't want him on loan three months before were now desperate to take him.

Before the draft for The Hundred on Monday morning, I had the pleasure of watching Devon Conway's superb match-winning 99*(59) for New Zealand against Australia, but this isn't going to be a debate about the merits of Devon Conway as a specific player - that's not the point of this piece.   After Conway's innings, Ashwin tweeted the following:-


Ashwin is referring to the IPL auction, which took place the preceding Thursday, and perhaps he was being a little humorous, or perhaps he was being 100% serious - we don't know.  However, it provokes some thought as to the recency bias that clearly takes place at auctions and drafts worldwide.  I actually have little doubt that if Conway had played this innings a few days before the auction, he'd have been sold, and potentially for big money as well.  However, the timing of his innings, coming four days after the auction, arguably contributed to him being unsold.

Conway hasn't changed!

Why is this?  As a player, like Stuart Webber said about Phillip Billing, he hasn't changed!  Conway was, to all intents and purposes, exactly the same player before or after the auction.  My model, which takes into account performances around the world and weights them to historical assessments of the standard of various competitions, suggested that after his innings, Conway's IPL expected strike rate rose by 0.57 runs per 100 balls - an extremely negligible amount.  His IPL expected average improved by 3.56 runs, but again, comparative to the previous figures, this was a fairly small percentage rise.  Before his innings, a probability distribution model would assign him an x% chance of playing such an innings, and after his innings, the same model would give him an x+y% chance of playing such an innings in the next match, but 'y' would be an extremely small figure.

In my view, this shows the distance that many T20 franchises - not just in the IPL, but worldwide - have to travel to get towards some sort of recruitment efficiency.  There were players signed at last week's IPL auction who struggle to justify either their recruitment full-stop, or their price point because there are no statistical metrics which could possibly advocate that particular player or their price.  I'm not even talking about the 10+ Crore players (perhaps Kyle Jamieson apart given his relatively small sample size of T20 data).  If you want to read my thoughts on the auction in detail, some links are below:-

Pre-Auction Preview with The Cricket Podcast: https://www.youtube.com/watch?v=f9i9kZVmtF0&t=185s
Post-Auction Quick Review with The Cricket Podcast: https://www.youtube.com/watch?v=kSMT_go1wrE&feature=youtu.be

Until teams do further layers of due diligence on players, reduce weight in their recency bias and stop being seduced by reputations and personal relationships, they aren't going to come close to recruitment efficiency.  As I've said many times before, a strong recruitment-focused analyst can be a considerable positive disruptor to the current status quo, and fortunately, several teams who I am working with understand this - however, they are currently in the minority.



12th February, 2021.

T10 Format with so many positives if done well 

The T10 competition drew to a conclusion last weekend and following it, there's been quite a bit of speculation towards the development of the format moving forward.  

From a personal perspective, I'm a big fan.  The format is short, which is beneficial in terms of attracting interest from an often time-poor society, and rewards attacking intent.  Plus, there are little in the way of overseas player restrictions.  There has been talk of this being the best format to bring cricket into the Olympics, and I think it would be really fun to do so - we also have the potential for more shock results in the shorter format, as on occasion, all it takes is one batsman to play an absurd strike rate innings to win the match for their team.  

Standard currently mixed due to questionable recruitment

It was also interesting to note that a number of people in the media, and some media articles, focused on the mixed standard of the tournament.  Calling it mixed would probably be being kind - there were some huge mismatches, in my view.  Ironically, there isn't really a logical reason as to why there were mismatches or mixed standards - teams were able to pick players from the same brackets as their rival teams.

The mixed standards were simply due to some bizarre recruitment, with some teams in particular seemingly having done little in the way of research or preparation - one team at the draft didn't even have a laptop on their table!  Three teams, and two teams in particular, appeared to have little idea on how to value players or what type of players were required to achieve success in the format.  Some players, who don't even represent their domestic teams as a local player on a regular basis, plus several pretty much retired players, were even drafted - yet numerous high quality T20 players who have skillsets likely to translate well to T10 cricket went unsold.

Short-format franchises need to have more accountability for poor processes

My opinion is simple - until the attitude of franchise cricket (be it T10 or T20) turns from 'oh well, we'll go again next year' to greater rewards for success and increased accountability for failure, the growth of shorter formats will be constrained.  The difference between football and cricket in this area is clear.  Coaches should be under pressure for poor recruitment, selection and strategies.  Coaches should be rewarded, and have merit-based opportunities when they are successful via good processes.  I'm not sure that is the prevalent mindset among owners currently, but it should be. 

There is huge scope for a data-driven approach to T10 (and T20, still, because many teams either don't do it, or don't do it well) and in my view, there's a massive expected edge likely when adopting such an approach compared to the recruitment of many teams.  While the draft was taking place, I tweeted exactly this:-


Big gap between best and worst performing T10 teams this year

The struggles of some teams were accurately illustrated by basic batting and bowling scoring outputs, as can be seen via the chart below (qualifying teams highlighted in red):-


Despite there being a relatively short group stage season, the league table pretty accurately represented performance outputs.  The four teams with the best bowling economy rates qualified, and the three teams with the best batting strike rate qualified.  The mixed standard of the tournament was also illustrated here, with two teams - Pune Devils and Maratha Arabians - having team performance metrics far worse than any of the other teams, and this was largely predictable following the draft.

Boundary-hitting yet again a huge driver for short-format success

At this point, some readers might be thinking 'well that's obvious, the best batting and bowling teams qualified, hardly groundbreaking news and he's given no insight as to how to pick a strong squad for the format'.  However, I did after last years tournament!  If you haven't read that article, you can check it out here.  In addition, either directly to head coaches that I had contact details for or via an agent, recruitment services were marketed to teams.

Moving on from that, I want to specifically focus on batting recruitment in this current article, because in my view many teams are leaving so much expected value behind with their squad selections.  Anyone who has read a lot of my previous work will know how highly I value boundary hitting as a key metric in short-format success, and this is arguably exacerbated even further in T10 cricket.  Unsurprisingly, given the close relationship between boundary-hitting percentage and batting strike rate (again, something I've written about previously), the top boundary-hitting teams qualified.  The worst teams, again, were Pune Devils and Maratha Arabians:-

It is clear that boundary-hitting is a key driver of success in T10 cricket, just as it is in T20 cricket.  In major T20 leagues, around 85% of teams with the higher boundary percentage win the match, and this figure drops around 20% to a ballpark 65% for teams with the lowest dot ball percentage.  Avoiding dots is nice to have, but the ability to hit more boundaries than your opponent is virtually mandatory.  The 15% of teams who win with a lower boundary percentage almost always fall into one of two categories - they either scored a lot more sixes (so they hit fewer boundaries, but the ones they hit were often worth two runs more) or the match was extremely close, frequently coming down to the last few balls.

This year's T10 tournament had 29 matches take place, and 23 were won by the team with the highest boundary count with two further matches having tied boundary counts.  Of the 27 matches with non-tied boundary percentage counts, 23 were won by the team with the highest (85.19%) - a figure which is extremely similar to the ballpark T20 figure for major leagues.  

In short, T10 and T20 teams need to work out realistic processes with which to win the boundary percentage count on a regular basis - through both boundary-hitting (good batting recruitment and in-play strategy) and boundary prevention (good bowling recruitment plus taking advantage of match-ups).

Player boundary-hitting output also quite mixed

Looking at the scoring output of individual players, output towards the left-hand side of the chart below should be avoided (lower than average boundary percentage), with the players in the bottom-left corner facing particular issues with scoring output in the recent T10 competition:-

The requirement for wicket preservation is pretty much non-existent in T10 cricket, and having a group of batsmen who are capable of teeing off from ball one is critical.  Not only this, but the teams with the highest six percentages - Northern Warriors (15.28%) and Delhi Bulls (13.68%) featured in the final - finding batsmen with strong boundary-hitting ability in conjunction with strong 6/4 ratios would be really obvious filters with which to run through a player database.

Given this, it's useful to ascertain how players managed to translate their T20 boundary percentage to T10 this year.  The chart below shows the boundary percentage for batsmen facing 50+ balls in T10 in the recent competition compared to their T20 boundary percentage (time period from the 2017-2018 Big Bash onwards):-


Basic T20 batting database filters could positively impact T10 franchise recruitment

A number of players, such as Azam Khan, Rahmanullah Gurbaz, Paul Stirling, Chris Gayle, Evin Lewis and Nicholas Pooran are strong boundary-hitters in both formats, and look really logical picks for T10 franchises.  However, there were also a number of players recruited by T10 teams with a boundary percentage of below 18% in T20 cricket, and only two - Ravi Bopara and Rovman Powell - managed a T10 boundary percentage above 30%.  This looks a realistic point for T10 recruiters to consider whether a batsman with a T20 boundary percentage below 18% can really add value over the average batsman in T10 cricket.

In fact, batsmen with an above 18% boundary percentage in T20 during this time period who faced 50+ balls in this year's T10 competition hit 30.16% of balls faced for boundaries - a figure which would have been top-level for the competition, and nicely above the overall 27.67% figure.  Particularly if a player can hit 20%+ boundaries in T20, and has a 6/4 ratio of 0.5 at the very minimum, or higher (they hit at least one six for every two fours) then they should be strongly considered by T10 recruiters.

Not only this, but of the 24 bowlers bowling 10+ overs in the competition, 18 of them were pace bowlers (75%), so prioritising recruiting extremely strong pace hitters (a generally abundant skillset in T20 cricket) would be an extremely sensible, and logical further layer of due diligence to take place.

This is just an example of how usage of really basic database filters can really help teams recruit better than they currently do.  There's a lot more to go into with player recruitment, and I definitely don't want to share a lot of the work which I'm currently doing with teams, but I do think this is an example of where a really simplistic data-driven approach can dramatically improve player recruitment, which in this format still appears to lack strong processes and evidence-based decision making.

Noted football manager Pep Guardiola was once quoted as saying that '80% of success is down to scouting', but I think this figure is close to 100% for T10 cricket right now.


28th January, 2021.

IPL franchises should be careful about overvaluing Big Bash performance levels

The group stages of the Big Bash concluded on Tuesday with a dramatic triple-header which saw Stars and Hurricanes eliminated, joining bottom of the table Melbourne Renegades in failing to qualify for the knockout stages.  With five of the eight teams qualifying for the latter stages, a team doesn't even have to be an average team to qualify - although as we will see later on in this post, one team in particular did look quite unlucky when coming in the bottom three.

With the IPL auction confirmed for February 18th, again the Big Bash gives players opportunities to put themselves in the shop window for the auction.  However, as I wrote about in my book, it would be unreasonable for IPL decision-makers to put too much weight into Big Bash performances, unless they have been top level and backed up by strong longer-term data.  Australian players are, generally speaking, extremely popular in the IPL - most likely due to the traditionally high number of Australian coaches, and the recency bias of the Big Bash which often takes place either partly or fully just before the auction - yet from 2017-2019, there was a considerable general drop-off in BBL performances in the IPL when looking at the overall numbers of domestic players who played in both the Big Bash and the IPL.

For domestic batsmen playing in both tournaments in this time period, there was a noteworthy drop in batting average from Big Bash to IPL, as can be seen by the chart below:-


Having said this, there was very little difference in average strike rate levels for players playing in both competitions, which shows a further interesting dynamic - ground dimensions in Australia (often big) aren't necessarily conducive to high strike rates, and this enables Australian batsmen to at least strike to a similar level in the IPL (with matches often played at smaller venues) despite the varying standards of the two competitions.  

Moving on to Australian bowlers playing both competitions, there was a similar drop off in performance output, as the charts below indicate:-


While the numbers above cover both pace and spin bowlers, interestingly, when delving deeper into the sample, Australian domestic spinners performed relatively better in the IPL than their pace bowling countrymen.  Splitting the sample into bowler type isn't ideal, given sample size constraints, but this is certainly interesting to note - do the more spin-friendly conditions in India have an influence here?

Essentially, the reasonable conclusion to draw from the above is that Australian batsmen generally have a lower batting average (due to negative impact on balls per dismissal) in the IPL than when they play in the Big Bash, and Australian pace bowlers in particular tend to relatively struggle in the IPL compared to the Big Bash.  From this conclusion, I would recommend that IPL franchises ensure that they bear this in mind when looking at Big Bash performance levels.  While some players have produced stunning performance levels this year in the Big Bash (more on this later) it is these players who should largely be prioritised by IPL franchises, as opposed to those Big Bash players with consistently good, but not top-level, performance levels.  These 'good' Big Bash players have been picked up by numerous IPL franchises in the past, and on the whole, haven't impressed in the IPL.

Scorchers, Sixers and Thunder the best batting teams so far

Having got the IPL discussion out of the way, I want to move on further to look at Big Bash performances in more detail.  First of all, some basic batting analysis which looks at boundary percentage and non-boundary strike-rate for each team:-


Looking at these two metrics allow us to look at the scoring output for each team.  Sydney Sixers and Sydney Thunder were closest to the ideal top-right corner, although Perth Scorchers were also in this positive quadrant.  These teams clearly are closer to maximising their scoring output compared to other teams.  Melbourne Stars, who failed to qualify, also had strong boundary-hitting but weaker non-boundary strike rates.  On the left-hand side of the chart are four teams with below average boundary percentages, and three of the four teams (Strikers, Heat and Hurricanes) also had below average non-boundary strike rates. 

We can also look at the approach which each team has taken with regards to attacking versus stability.  The chart below looks at strike rate versus balls per dismissal for each team in the group stages:-

 
Again, the four teams with the above-average boundary percentages all had above average strike rates, which isn't a surprising conclusion to draw - there is a clear correlation between boundary-hitting and strike rates, which I've demonstrated previously.  Renegades' poor boundary-hitting is reflected in them having the worst batters strike rate in the competition, which in conjunction with the lowest balls per dismissal figure shows the problems they have had with their batting group.  Adelaide Strikers were the only team situated in the top-left corner (weak striking, strong stability) which perhaps indicates that they could benefit from dialling up the aggression slightly - although they don't possess numerous above-average boundary-hitters in their playing group, so perhaps this problem can only be solved via recruitment for them.

Scorchers and Strikers with the best bowling output

Continuing on to the bowling performance levels for each team, a simple look at bowlers economy rates versus balls per wicket provides some useful insight into which teams performed well or poorly with the ball:-

Here we can see that Perth Scorchers by some distance have had the best bowling attack in the competition so far, with Adelaide Strikers the next best.  Melbourne Renegades had the worst bowling performance levels overall, again by some distance - given their batting performance levels as well, it really wasn't surprising that they came bottom of the table.  

Melbourne Stars arguably unlucky not to qualify

This is reflected in looking at the mean deviation for each team.  Mean deviation looks at the batting and bowling performances (averages and economy/strike rates) in conjunction with each other.  For this I used a 50:50 split in weight between averages and either economy/strike rate, although you can play around with different ratios depending on the importance you assign to each metric.  If you changed this to 40:60 for example, valuing economy/strike rates more, it wouldn't really change much though.  Here's the chart:-

Based on this, there's some evidence that Perth Scorchers are the best team in the competition so far, placed in the ideal top-right hand corner (strong bowling, strong batting) while unsurprisingly, bottom-placed Melbourne Renegades were the opposite, in the least ideal bottom-left corner.  Adelaide Strikers had a notable bias towards bowling strength over batting, while the opposite was in evident for both Sydney teams.  Given Melbourne Stars had above 1.00 mean deviation for both batting and bowling, it could certainly be argued that they were very unlucky not to qualify, particularly considering the five qualification spots among the eight teams.

Philippe and Hales among batsmen to impress

Having looked at overall team performance so far, I thought it would be useful to conclude by looking at individual player performances, for players either batting or bowling a minimum of 150 balls in the competition so far.  The first chart I want to discuss is the scoring output for batters in the tournament group stages (overseas players in red):-


Here we can see that Josh Philippe was the closest to the ideal top-right hand corner, with strong boundary percentages in conjunction with strong non-boundary strike rates - it's not a surprise that Royal Challengers Bangalore retained him in the recent round of IPL retentions.  Alex Hales had stunning boundary percentages but below-average non-boundary strike rates, while the reverse was true of Jimmy Peirson.  However, a player with the dynamic of Hales has considerably more strike rate upside than the reverse type of player, such as Peirson.  Consider it like this - around a 4% increase in boundary percentage is broadly similar to a player increasing their non-boundary strike rate by 20.  The group of players slightly inside the top-left corner, or towards the far left hand side, or in the bottom-left corner, have struggled in this year's Big Bash to maximise their scoring output.

Bringing stability into the picture, I want to look at boundary percentage versus balls per dismissal for the same group of batsmen:-


Ideally, a batsman would be situated in the top-right hand corner (attacking, high stability), and teams should be very careful not having too many players expected to be in the top-left corner in their lists if they want to be capable of hitting high scores - there really is no need to have more than one anchor in a team at the most, if you recruit well in other areas (e.g. some bowlers who are capable of lower-order high strike rate batting cameos to add depth).  Players in the bottom-left hand corner had below-average boundary-hitting and balls per dismissal numbers - far from ideal.

Jhye Richardson and Mujeeb among the bowlers with strong run prevention metrics

Finally, bowlers.  I want to look at a comparison between dot balls earned and boundaries conceded.  Generally speaking, lower boundary concession percentages are more beneficial than high dot ball percentages, although some players were capable of producing both as the chart below illustrates (pace bowlers highlighted in black, spin bowlers highlighted in red):-


Closest to this ideal bottom-right corner are the likes of pace bowlers Jhye Richardson, who was highly impressive in the tournament, as well as Jason Behrendorff, Adam Milne and Riley Meredith.  For spinners, the Afghanistan duo of Mujeeb and Rashid Khan clearly impressed.  We can see that most spinners were grouped closely around the 30-36% dot ball area with below 17% boundaries conceded, and this is a dynamic which occurs in most T20 competitions worldwide - spinners concede lower boundaries in general than pace bowlers, but have mostly lower dot ball percentages as well, and the phase usage of the two bowling types is key in this dynamic as well.

With the knockout stages starting tomorrow, I hope this analysis and discussion has been useful, and as always, I'm happy to respond to questions via email or Twitter (links at the top of this page).



21st January, 2021.

IPL retention list announced yesterday

Apologies for no post last week, I've had a lot of recruitment work on so had to put that post onto the back burner for a few days time - I'm planning to get a few posts done between now and the end of January.  However, the delay does allow a natural starting point for this week's post - the IPL retention lists, which were announced yesterday.  If you didn't see the lists of retained/released players, you can check them out here.  According to the IPL website, 139 players were retained and 57 were released.  Of the released players, 28 were overseas - around half of the released players - and actually I think there's a decent case for this number to actually be pretty low.  

What criteria should teams use for overseas player retention?

Strictly speaking as treating players as a commodity point of view, there isn't really much need to retain overseas players unless they tick at least one of the following criteria:-

1) World-class, genuine marquee player (current ability, not reputation).
2) Possesses rare skillsets which are likely attractive to numerous rival franchises (such as strong death bowlers, all-rounders who are both genuine hitters and regular four-over bowlers, or 20%+ percentage boundary-hitters with 20+ balls per dismissal).
3) Offer strong financial value based on their current salary (if they went back into auction, it would be likely that they'd cost more than their current salary).

That's it.  No other overseas player really needs to be retained - the overseas market is completely different to the domestic market.  It has plentiful supply in non-rare skillset areas and there are numerous potential overseas players who have never played IPL before who are of equivalent or better quality to some of those retained players.

Going through the squad lists and current salaries, I only really make a case for the following overseas retentions:-

CSK: Sam Curran - improving and cheaper than when signed for KXIP.  Offers rare all-rounder skillset, CSK need left-handers & teams will pay a premium for left-arm bowlers too.
DC: Kagiso Rabada - World class pace bowler and death bowler at just 4.2 Crore.  Anrich Nortje - Would almost certainly be more expensive than 50 lakh if bought back at auction.
KXIP: Nicholas Pooran - Incredible spin-hitter, offers a useful wicket-keeper option and would almost certainly be more expensive than 4.2 Crore at auction.
KKR: Andre Russell - Has had injury issues but no batsman offers strike rate and boundary-hitting upside like Russell - a very rare skill-set.  Perhaps more of a batter who bowls, as opposed to being a genuine all-rounder, but again, a very rare skillset.  Lockie Ferguson - Offers strong value based on 1.6 Crore price - would be likely to cost more to buy back at auction.  While he only bowls around 15% of his overs at the death, his numbers in Powerplay and middle overs are impressive.
MI: Quinton de Kock at 2.8 Crore.  He's been well used by them at the top of the order and offers a useful wicket-keeper option as well.  Would probably cost more to be bought back at auction.
RR: Jos Buttler - excellent keeper/batter who would be in demand at auction.  Jofra Archer - World class pace bowler.   The English duo offer great value at around 12 Crore combined.
RCB: AB De Villiers - world class, marquee batsman.  In my view their main challenge is to try and get Kohli cheaper than 17 Crore.  Josh Philippe - High potential, plus his strong spin-hitting is a rare skill-set.  Would be likely to cost more than 20 lakh to be bought back at auction.  Adam Zampa - excellent leg-spinner who has done very well in the current Big Bash.  Will be likely to cost more than 1.5 Crore to buy back at auction.  Moeen Ali - boundary percentage and strike rates compare very favourably to most batsmen and offers useful off-spin option.  Chris Morris - was fairly expensive in the previous auction but offers a rare all-rounder skillset as a genuine hitter and four-over bowler who bowls at the death.  Likely to be in demand at the auction.
SRH: Kane Williamson - Mr reliable - would be likely to cost more than 3 Crore to buy back at auction.  Jonny Bairstow - again would be likely to command more than 2.2 Crore at auction.  Rashid Khan - world-class leg spinner and dangerous lower-order hitter, would cost 15+ Crore potentially in an open market.

Mumbai Indians look like they have adopted this strategy

So this makes 17 overseas players that, in my view, tick at least one of those three boxes, which works out at an average of just over two overseas players to retain per team.  Due to a quirk of the rules, it appears that teams have options to retain players both originally purchased at auction and replacement players, so there are around 65-70 overseas players associated with the competition - from a value perspective, it's difficult to make a case for a number of overseas players currently retained by IPL franchises.  This isn't to say that many of these aren't good players, but simply that there isn't much logical sense in retaining them at this point in proceedings.  IPL teams are generally quite risk-averse when it comes to retentions and trading - I already have made a data-driven point in my book that many players (particularly domestic players) are retained for one expensive retention too many when they are past peak age - so this general approach from franchises isn't particularly surprising.  However, I'm far from convinced that their approach in this area is close to optimal.

Interestingly, current champions (and overwhelmingly the team with the best strategy in the competition in recent years in my view), Mumbai Indians, were one team who adopted quite close to this strategy.  Of the seven players they released, five were overseas and none were genuine first-team options.  Given that they've released four overseas pace bowlers, it strikes me that they might be clearing squad space/budget for a high quality overseas pace bowler capable of bowling well in both the Powerplay and death overs - perhaps something their overseas pace bowling options didn't consistently offer last year - and while I'm not willing to share the name of the particular bowler that I think they will have in mind (he has these skills plus can hit lower down the order) , if they do sign him they will be incredibly difficult to beat.

A number of questions about overseas recruitment in advance of the auction

The auction dynamics this time around will be fascinating.  Even with just 28 overseas players released, I'm looking forward to seeing overseas recruitment strategies from teams, and a few of the questions I have are as follows...

Will teams go for 'reputation' players or look to pick up younger players with huge upside?   There are plenty of these high upside players in their early/mid 20s around currently, and in the next couple of years I expect a huge turnover from the current players on the T20 'merry-go-round' who somehow get regular franchise league contracts, to these potential superstars coming through.  Will there be recency bias towards the Big Bash, as has potentially occurred in recent years?  Will teams overvalue more red-ball orientated Australian players who have performed against India recently in the Test series?  Will franchises realise (some still don't seem to) that giving big contracts to batsmen with strong rotation abilities but below-average boundary percentages is a poor usage of resources given the plentiful supply of these type of players in the domestic market?  

It's worth pointing out here that the domestic market has a completely different dynamic which cannot be compared to the overseas market.  Domestic resources are scarcer in some areas (in my view, aggressive batting and death bowling in particular) and sometimes this leads teams to overpay for certain skillsets, or take risks signing players for big money who have performed well at lower levels of domestic T20 cricket.  Even so, there still looks like a bias towards reputation among a number of domestic players, with some players who, statistically, look past their best still retained on high value contracts.  It will be interesting to see if this dynamic changes in the future as more teams start to improve their recruitment efficiency.

Should players have a say in their retention?

According to reports, this is the last year before a major auction in advance of the 2022 season, with the potential for several new teams to be added.  It's interesting to see that, unlike some other leagues, there doesn't appear to be a limit on the number of retentions and players don't appear to have the ability to renegotiate their salaries.  It appears that all the power currently lies with the franchises, as opposed to the players, and some players on low contracts could well have been in a better financial position if they had been released and were available at auction.  Take Ishan Porel as an example.  In my book, in advance of last year's tournament, I discussed Porel as a young bowler with high potential, yet he didn't play a single match for Kings XI Punjab last season - this was a real surprise to me.  Having performed very well in the Syed Mushtaq Ali Trophy recently (13 wickets in 5 matches at 10.84 average) he would almost certainly be in demand at the auction, and would likely command a salary well in excess of the lowest 20 lakh base price, which is the price he was signed at by Kings XI previously.  In some other leagues, to retain a player, both player and team must agree on the retention and price point of that retention, and I think this would add an interesting dynamic to the IPL process moving forward.


8th January, 2021.

Conveying ideas to decision-makers 

For the first blog post of 2021, I want to discuss some of my own experiences which I hope will benefit a lot of the aspiring analysts who get in touch with me asking for advice.  Some aspiring analysts have even gone as far as send me samples of their work, which has been fascinating for me to look at and also illustrates the considerable drive and work ethic which many aspiring analysts without current work with teams are willing to demonstrate in order to try and catch the eye of people in the industry.  The one general theme of a lot of this work I've read - both in the way of work which people have privately sent me, but also that I read online - is that it is too 'technical' - often reading like a University dissertation.  I can understand why this approach has been taken, given a lot of analysts have a very academic background, but if my eyes are glazing over reading formulas, references and footnotes then it's pretty likely that most people in the industry also will lose interest pretty quickly as well.  Working on an approach to convey your ideas and findings concisely, perhaps via charts as opposed to formulas or numbers, is likely to get more 'buy-in' than the reverse approach.

One particular analyst has impressed me considerably with the independent work he's been sending me from his own choice, and while I'm not in a position to recruit anyone currently, should I need any assistance with future projects he will be at the top of the queue for that.  In this post as well, I'll also talk about the difficulties of being a 'nobody' trying to obtain work in the cricket world and some of my own personal experiences, which I hope will give some fascinating insight into some of the battles which I have personally faced, and to some extent, continue to do so.  

Difficulties bringing a product or service to market

Imagine a hypothetical world where you, as an individual, living in a random place and knowing no-one in professional cricket, has 'solved' T20 cricket completely independently through your own work - I'm talking about having complete game theory optimal solutions for both recruitment and in-play strategy.  As far as I can see from my own personal experience, no-one (including myself!) is remotely close to achieving this, but just take a quick moment to think about how you might proceed from that point onwards in order to put your strategies into action.  How do you get a team, or multiple teams interested in what you can offer them?  More importantly, how can you market this without giving your work away for free?

In an ideal world, teams would be falling over themselves to get access to this solution.  But the problem is twofold - firstly, at this point, you are a nobody.  No-one knows who you are.  Secondly, even if teams understand the value of what you can offer them, it's still far from a given that they want to invest in it - sport still isn't necessarily a meritocracy.  Some teams are beginning to appreciate that some analysts (often people who have never played cricket professionally) can offer alternative points of view which challenge the conventional wisdom of those involved within the sport for decades, but these teams still aren't close to being a majority.  Answering the question in the previous paragraph isn't particularly easy, particularly if you don't know a single person in professional cricket! 

From personal experience, mass messaging on email/social media/Linkedin has a very low chance of even getting a reply, let alone a positive response, so that option is unlikely to yield much in terms of bringing your product to market.  Personally, I've even written post-season reports on teams, being careful to try and point out problems without giving specific solutions and sent these to any contacts at overseas franchises, but my experience was negative in terms of getting work from this marketing method.  The long-term option that I took when I first started was to try and create a variety of free content on this website and on Twitter in an attempt to showcase what I could offer teams.  This has benefits, in that eventually there is a decent chance of attracting some interest, but also negatives - you probably give away more for free than would be ideal.  Certainly, I have been told by several people in the cricket industry that some teams have used the content that I have given in free articles, as opposed to actually recruiting me to work for them.  

Could teams be more proactive recruiting bloggers?

The balance between showing your worth and not giving too much away for free is, without a doubt, a tricky one.   The entire process is extremely difficult.  I would imagine the majority of people with the skillset to be a sports analyst would probably earn more money and have far less self-doubt (which is easy to acquire given the puzzling lack of demand from many teams to be more efficient) working in a different industry where that skillset will still thrive.  On the subject of self-doubt, as said, this can be easy to come by.  To give you some insight into a situation where I certainly felt this, the first time I ever met with a professional cricket team to market my work I did a presentation to the decision-makers, who then asked me to email them with a formal proposal to work for them.  Subsequently, it took them three months to reply back to me!  In those three months, it's natural to wonder why, and to go over the entire presentation in your head and second-guess yourself on how you somehow messed up without realising.   I guess what I'm trying to say is that any aspiring analyst needs to be prepared to deal with this, and it hasn't been an isolated incident.  As an individual, I value clear and concise communication extremely strongly, and if I had a decision to make between working for two teams, the team with the better communication would be a huge positive for me.

As time developed, I managed to start getting better relationships with people in the cricket world, and overall, these have been extremely positive.  Without these contacts and relationships, I don't think I'd have earned the roles which I have done - word of mouth still seems pretty important, so I'd recommend any aspiring analyst to try and build up their network as well as they possibly can.  In football and some American sports, this word of mouth doesn't appear necessarily as important, with some bloggers being proactively recruited by teams purely on the strength of their personal portfolio of work.  I was interested to read Tim Wigmore's excellent recent piece about Ashwin Raman, a 17 year-old football analyst from Bangalore, who was spotted by Dundee United on the strength of his work online.  There might be a few instances of this happening in cricket, but I'd suggest that it is not commonplace.  Some readers may think that I got roles from the content I've provided online, but as far as I'm aware, it's not been via a direct consequence of that - the work might have caught the eye of someone, who mentioned it to someone else, who recommended it to another person, etc.

Do teams still generally undervalue analysts financially?

Moving on to finances, when I do give advice to people contacting me about how to try and break into the cricket analytics world, I'm also very clear to point out that I would recommend them not to do work for anyone for free in order to gain 'exposure'.  Firstly, their time has value - both in their time invested to get to the point where people are interested in what they can offer, and also their subsequent time doing the actual work - and also because having done this myself, there are some people who will try and get you to do work for free by promising much and delivering nothing.  I probably was taken advantage of in this respect by some people, so my advice would be to any aspiring analyst not to fall into this trap.

Financially, being an analyst probably still doesn't offer the rewards that people with the skills required could earn in other industries, and consideration should be given to how much financial discrepancy you value in terms of achieving what is, to some people, their dream career.  It always puzzles me why analysts aren't paid more, particularly those with niche skills such as recruitment and list management.  Think of it like this from a county cricket perspective - how much does the worst player or the two worst players in a county team earn?  How much do they contribute to the overall success of their team?  The chances are that they'd only play a handful of matches with mediocre returns.  Could that money be better diverted to someone who can help teams recruit better and help their planning for matches, for example?  Of course, I'm biased, but I'd suggest this is a much better usage of finances than giving a contract extension to a fringe player, or recruiting a mediocre squad player.  Will the general financial situation for analysts change?  I think it will - over the next decade I can definitely envisage the better analysts becoming GMs or Director of Cricket at teams, just as it wouldn't surprise me to see more people from this background become Directors of Football in Football as well - and of course the financial rewards from such positions is greater.  Not only this, when more teams start becoming smart with data, there will be a premium placed on the best analysts and this will put them in a better negotiation position financially.  Plus, in that situation, as more smart teams become data-driven, the teams who aren't data-driven are going to fall so far behind there will be pressure on them to take a similar approach, and therefore there will be more demand for analysts full stop.

This is a pretty different post than a lot of my previous content - there's no data or commentary on data here, and much more in the way of my own personal experiences.  However, I do hope that it was interesting, and that any aspiring analyst reading this finds it useful in terms of preparing themselves for marketing their abilities to the industry.  Of course, I'd be happy to answer any specific questions, if you have them, via email.


29th December, 2020.

This first post covers several subjects, including the Big Bash so far, my thoughts on the X-Factor, recruitment and list management in the competition plus also some thoughts on last week's T10 draft...

BBL Teams failing to maximise overseas output

After today's double-header, every team in the Big Bash has now played at least four matches, with most having played five.  It's probably fair to suggest that you can't win a tournament in the first few matches but you can certainly go a long way towards losing it.  Melbourne Renegades (four points from a possible 20) and Perth Scorchers (two points from a possible 16) are already in danger of being cut adrift at the bottom of the table and need a stunning comeback in order to make the playoffs.  

One interesting area to discuss from these opening matches is the recruitment of overseas players, with the quota for overseas players being increased from two in previous years to three this year.  In theory, this should increase the quality of the competition with there being a general relationship between the quality of a competition and the number of overseas players permitted per team, and this is logical because there is only a certain amount of domestic talent to go round before there is a quality drop-off, so having fewer domestic players and more overseas players tends to increase the standard of a tournament in general.  This is something which could affect the IPL in 2022, with there being the potential for two extra teams to be added - if there was no increase in overseas players permitted, there would need to be 70 domestic players picked for each round of matches, as opposed to the current 56, and the 14 extra players will either need to have been previous fringe players or players who were unsold at auction.  However, if five overseas players were allowed in each of the 10 teams, there would only need to be 60 domestic players picked, which would be much closer to the current 56.

You'd think that with three overseas players now allowed per team, the Big Bash teams would be keen to try and maximise output from their extra overseas addition in an attempt to take advantage of the greater expected returns from the average overseas player compared to the average domestic player.  However, it hasn't really happened so far - below is the combined matches played for overseas players for each team:-

Team

Maximum Possible Overseas Appearances

Actual Overseas Appearances

% of Quota Used





Adelaide Strikers

15

15

100.00

Melbourne Renegades

15

13

86.67

Sydney Sixers

15

13

86.67

Melbourne Stars

15

12

80.00

Perth Scorchers

12

9

75.00

Hobart Hurricanes

15

11

73.33

Brisbane Heat

12

8

66.67

Sydney Thunder

15

9

60.00



Only third-placed Adelaide Strikers (Danny Briggs, Phil Salt, Rashid Khan) have used their full overseas quota so far in the competition, suggesting that the other teams have struggled to maximise their expectation from their overseas players.  At this point it's probably fair to point out that overseas list management is a very tricky with regards to players potentially pulling out for various reasons, and even more so in a global pandemic, but one easy solution would have been for BBL teams to focus on players who aren't likely to be required for their international teams - as Strikers did with both Briggs and Salt.  Furthermore, teams needed to have drawn up plans to deal with overseas player withdrawals but with the exception of Benny Howell being signed by Melbourne Renegades (having already been in Australia), teams have arguably left expected value on the table.  I know of one proven-quality T20 Blast batsman who is currently in Australia and who would have added value to a number of these Big Bash teams who haven't maximised their overseas quota.

My view is that teams should have looked towards a specific 'Moneyball' strategy, by picking up lower-profile players who didn't have any commitments which would clash with the competition.  Some of these lower-profile players are easily good enough to improve Big Bash teams, and I want to make a clear point that I'm very strong on - just because international selectors deem a player not good enough for their national team, it doesn't mean that they are not good enough - it's just the (often incorrect) opinion of these international selectors.  In the T20 format, there are numerous players worldwide who are high quality but don't feature in their international teams plans, and at this point in time, many players prospects of being picked up in overseas leagues are influenced by their national team selectors.  To me, international recognition seems a flawed recruitment consideration for T20 teams, given that I'd be surprised if many international selectors had performed detailed research into the drivers of success and skillsets required to be a successful T20 player.

Potential lack of strong boundary-hitting depth in the domestic player pool

Maximising overseas players in a team's squad is imperative, given that already in the competition, Australia and Australia 'A' have taken players out of BBL teams.  It's had a particular effect on the batting pool in particular - of the 23 domestic players who have currently faced 60+ balls in the tournament, just 10 have a boundary percentage in excess of 16.12% (the tournament mean so far).  Eight of the 23 domestic players facing 60+ balls have a boundary percentage below 13%, and all of these have a strike rate of 121 or below (and some quite a bit lower at around or below the 100 mark).  

Last season, 22 domestic batsmen faced 200+ balls throughout the full competition, and just seven of these had a boundary percentage in excess of 16% - Chris Lynn, Matthew Wade, Josh Inglis, Jake Weatherald, Glenn Maxwell, Marcus Stoinis and Nick Larkin.  Based on these numbers, it would appear reasonable to consider that Australia have some potential batting depth issues for above-average boundary hitters, and there were plenty of available players worldwide who could have potentially solved this problem for the various Big Bash teams.  In fact, it would have taken me just several hours to run some database filters and provide Big Bash teams with a shortlist of potential overseas players who could have solved this issue.

This, to some extent, is a problem which has also manifested itself at international level.  Australia frequently pick multiple players with below-average boundary percentages, and it's not surprising to see their team boundary percentage is unspectacular in T20i matches played from 1/1/18 onwards:-


However, Australia do have very strong running/rotating abilities - again a skillset apparent in a lot of their domestic T20 batters - with only India (very marginally) having a higher non-boundary strike-rate.  Unfortunately, as I wrote about in the linked articles here, this is the wrong type of intent and a skillset which has less correlation towards success in T20 cricket compared to boundary-hitting.  I have little idea as to whether this is actually the case, but it's interesting to wonder if maximising running between the wickets but with less priority on boundary-hitting is coached into Australian batsmen from a young age, a dynamic which is rather the opposite to West Indies who have many batsmen with high boundary percentages but low non-boundary strike rates.

Teams still struggling with the X-Factor

As well as the increase to three overseas players, I was particularly interested in seeing some of the other new innovations and rules implemented in the Big Bash this year, and none more so than the 'X-Factor', which enabled teams to make a substitute after 10 overs of the first innings.  When hearing about this new rule, I immediately got thinking about how to try and utilise this strategy with positive expectation, and there were clearly some options, some of which being more obvious than others.  I will say now that if I worked with a BBL team, their head coach and captain would be able to access a report which I would have constructed in advance of the competition which detailed numerous potential strategies which could be implemented if various scenarios developed in matches.

One obvious strategy would be for the team batting first subbing on a batter if they are struggling by the 10 over point, but this is something that Perth Scorchers failed to take advantage of in their innings against the Melbourne Renegades.  In fact, captain Ashton Turner was quoted as saying 'I think we're still getting our head around it and trying to figure out a way we can best utilise the rule changes'.  Given that the rules were announced a reasonable amount of time before the tournament started, there is little excuse for teams to still be 'getting their head around it' - this should have been done in advance.

It's probably fair to suggest that the 'X-factor' hasn't caught the imagination yet, although my opinion is that this isn't necessarily due to the rule being poor - in fact I really like it - but more so due to the lack of creative thought and innovation from teams, plus perhaps some fear on their part about getting it 'wrong'.  Unfortunately, this gives credence to the argument by some traditionalists that these new rules were simply a gimmick.  Teams being scared about something new isn't necessarily unique to the 'X-Factor' in the BBL, with IPL teams extremely reluctant to trade in the mid-season window as well - in my experience, most cricket teams generally have a very risk-averse approach, and are rarely willing to try and challenge 'conventional wisdom'.  The media and social media has a lot to answer for here - if a team tries to be innovative in T20 and loses badly, they will get criticised, but if they play conventionally and lose by 10-20 runs then 'that's the game'.  Teams such as England in ODIs and Mumbai Indians in the IPL are seen by many as anomalies, as opposed to being a template which offers potential positive expectation.  The incredible thing is that England and Mumbai have achieved success in plain sight - it is utterly obvious what their strategies are - yet teams are so reticent to copy it, probably due to this fear.

When I was considering how smart teams would use the X-Factor, my thoughts mainly focused on specific match situations, and these weren't necessarily reactive to these match situations.  For example, let's discuss a potential proactive situation regarding a random opposition team that has one good left-hand top order batter (who is weak against off-spin, like many left-hand batters are) and the rest of their top six are right-handers.  A smart team might be reluctant to play a right arm off-spinner against this opposition batting line-up because there is only one left-hander in the opposition team, but the X-Factor would bring an opportunity to start a right-arm off-spinner in the starting XI to get one over of a great match-up against the good left-handed batter and then sub that off-spinner off no matter what the outcome.  Can you imagine the uproar from commentators if the off-spinner took a wicket and then was subbed off - something like 'he's just taken a wicket and they're subbing him off!' would be pretty predictable - but it's a great specific match-up play, given that the right-arm off-spinner is unlikely to be nearly as much of a threat versus the remaining right-hand batters in that opposition team.

Watching the tournament so far, in my view there have been many strategic errors committed by teams, but here I want to focus on two specific ones that I really didn't like, for various reasons.  The first one was Adelaide Strikers removing Danny Briggs after one over against Sydney Sixers.  At the time, James Vince and Daniel Hughes were batting, with Dan Christian and Jordan Silk carded to come in next for Sixers.   According to my data, from the start of 2018 onwards, the combined quartet of Vince, Hughes, Christian and Silk strike at 121 against slow left-arm, with 21.4 balls per dismissal and a boundary percentage just shy of 11% - a pretty positive outlook for a slow left-armer such as Briggs.  Instead, Briggs was subbed off and Christian promptly hit a 15-ball 50 against better match-ups for him.

The second instance that I really didn't like was Hobart Hurricanes using overseas player Keemo Paul as an X-Factor sub.  My opinion is much more straightforward here - if a team uses one of their overseas players as a sub (whoever the player is), then questions need to be asked about either recruitment or selection.

Moving on ever so slightly from the Big Bash, I was interested to see the Australian quick Aaron Summers get an overseas deal at Southern Punjab in Pakistan.  Such a deal is very rare, but Summers has trodden a rather unconventional path in his career so far.  I mentioned earlier that often a player's career can be somewhat dictated by the opinion of just several people, and it's interesting to see Summers get this opportunity (having also previously played PSL as an overseas player) despite barely featuring as a domestic player in the Big Bash.  It's difficult to assess Summers from a statistical basis due to a very small sample size of matches, but teams in Pakistan clearly have seen something in Summers which Australian teams perhaps haven't done.  I'm not sure who is right or wrong but it will be fascinating to find out.

T10 draft shows the potential edge a data-driven team could have

Finally, on a personal note, I was disappointed not to be involved in the T10 draft which took place last week - I think it is a format which has a lot of potential upside in future years. Following last year's tournament, I wrote a detailed strategy guide to the format (which can be seen here).  I watched the majority of the draft on YouTube, and because of some recruitment in previous years not particularly adhering to drivers of success in the format as detailed in this article, I had fairly low expectations of all teams performing strong recruitment generally.  Based on the recruitment performed this year, some teams have assembled much better squads than some others, and some of the picks from some teams struggle to adhere to any statistical rationale.  During the draft, I tweeted that 'it's not difficult to imagine the edge that a smart data-driven team would have considering many of the picks so far', and, concluding this week's blog post, that statement still looks relevant in both T10 and T20 cricket at the current time.




Comments