"Getting Your Eye In" - How long does it take?

There has been various discussions over the years as to the importance of wickets in T20 cricket. Way back in 2008, an article on Cricinfo stated that "wickets don't mean much in T20. This fact doesn't seem to be widely recognised, but the whole concept of the 20-over game relies on it. If wickets were important, then batsmen wouldn't blaze away at 8 or 9 an over." Manish Verma wrote in an article in 2014 on the Red Bull website that "it is wickets that are best at slowing down the scoring rate, even in the shortest form of the game." An article by Tim Wigmore last year spoke about Leicestershire approach to bowling early in the game, explaining how "with the ball, Leicestershire attacked early on, even if it meant leaking boundaries, recognising that ultimately the best way to contain a T20 innings was to take regular wickets." Tim also wrote later in the year about the West Indies approach to batting, explaining that "wickets have been systematically overvalued in T20, holding totals back."

Wickets are generally met with great celebrations from the bowling team

It is interesting how the discussion has gone from feeling that wickets are relatively unimportant in T20 to how they are the best way to limit the run-rate, whilst also recognising that their impact on restraining the run-rate is itself only based on flawed strategy from the batting teams. Wickets may only be important in containing T20 innings because teams think that wickets are important in containing their innings. If teams moved toward a mindset, as the West Indies have done for periods, that wickets are less important and having hitters down throughout the order means that you can sacrifice wickets with little slowdown in scoring, then wickets may well become relatively unimportant.

There are obviously caveats to that. It is important to remember that, regardless of how a player coming in to replaced a dismissed batsman approaches his innings, if that player is substantially less able than the batsman that he is replacing, that is a blow. For example, if Andile Phehlukwayo comes in to replace AB de Villiers, that wicket is important.

Regardless of the importance of wickets themselves, the topic that I wanted to look at for this article is how long it takes a new batsman to 'get his eye in' following a wicket. For each ball in a T20 innings, I currently generate an 'expected runs' value that indicates what one would expect the average batsman to score off that ball. This is calculated by looking at the average runs that particularly ball at that stage of the innings, then adjusting that based on the ground and the bowler. For example, in the 2017 T20 Blast match between Durham and Nottinghamshire at Trent Bridge, the final ball of the powerplay with Durham batting was bowled by Jake Ball to Michael Richardson and was given an expected value of 1.50 runs. The final ball of the powerplay with Nottinghamshire batting was bowled by Ryan Pringle to Alex Hales, a far superior batsman, and this was given an expected value of 1.61 runs.


Using these expected runs values, we can look at how batsmen perform compared to expected runs when having faced certain numbers of balls. For example, across my entire database, batsmen facing their first ball would be expected to score 1.25 runs, but in actuality, they only score 0.75 runs. From this, we can see that they score 0.5 runs less than what we would expect. The table shows the values for the first 20 balls of a batsman's innings. We can see that averaged across the entire database, it takes roughly 15 balls faced for batsman to reach the average scoring rate (in this case, given it includes the entire database, this average would be 0.00).

However, one might ask the valid question of whether elite or very good batsmen are likely to get scoring at a quicker rate? If we narrow the focus down to those batsmen with a rating of 1.10 or greater in my rating system, we do indeed find that it takes these batsmen quicker to reach the scoring of an average batsman - just 6 balls for this group of players. Similarly, for poorer batsmen with a rating of 0.9 or lower, it takes them roughly 24 balls to reach this scoring rate.

Now this is not particularly surprising. What is interesting is that if we look at how quickly each group reaches the expected scoring rate for that specific group, rather than for the average player. For example, for the elite group, they have an overall expected scoring rate of 0.10 runs above average per ball faced, for the very good group, this is 0.05 and for the poor group, this is -0.15. In other words, the elite batsmen are worth 0.25 runs per ball more than the poor batsmen. However, whichever group we look at, we find that it takes 7 balls faced to reach the expected scoring rate for that group.

We also find that for both the elite and very good groups of batsmen, the scoring rate levels out at between 0.05 and 0.1 runs above the expected rate for that group from around balls 10 to 30, then we find it tends to take off again around the 30 balls faced mark, which suggests this is the point when the batsmen feel properly settled and confident to launch their attack. For the poor batsmen, we again see the levelling out after around ball 10 at 0.05 to 0.1 runs above the expected rate for the group, but there is no sign of any acceleration again after that stage.

Do batsmen consciously play more defensively during the first few balls faced?

From this, we can see that it is not only the fact that they have a higher scoring rate during the middle parts of their innings that sets the top T20 batsmen apart from the rest. It is also the fact that they have an additional gear to step up into once they are settled that the lesser batsmen do not appear to have. However, it seems that it takes almost the exact same number of balls for all batsmen to 'get their eye' in and reach the rate that we would expect them to score.

Returning to the initial discussion, what does this mean about the importance of wickets in T20s? Well, clearly it shows that dismissing an elite batsman and replacing him with a lesser quality batsman is important, but it did not really require any analysis to realise that. However, it does show that taking wickets is currently important to restrain the scoring due to this 'getting their eye in' period that all batsmen appear to perform. However, we cannot really tell from this whether this is a conscious action on the part of batsmen (i.e. they are actively taking less risks at the start of their innings) or whether it simply does require time for a batsman to up his scoring rate and there is nothing that can be done.

One way of looking deeper into this might be to look at whether we see more wickets than expected in these early balls of an innings. Logically, one might expect settled batsmen to be dismissed less than expected, while new batsmen might be dismissed more quickly. This is something that I will return to in a future article.

Predicting the Future of the ATP Rankings

So much can change in tennis over the course of just a couple of months. Even more can happen over the course of a year. To try to predict what might happen in five years time is arguably almost impossible. In the last few days, an old prediction made by L'Equipe has surfaced again and has been fairly widely mocked for how bad it has turned out.

Their prediction, made in 2013, was for the top 10 ranked players at the end of 2018. As the image above shows, just two of their predicted players ended up in the actual top 10, with 6 of their picks not even ranked inside the top 50 in the world.

Rather than just deriding their efforts, closer inspection suggests that some of the picks are not as ridiculous as they might seem. Grigor Dimitrov has never made it to #1, but if we go back twelve months, he had just reached a career high ranking of #3 and only dropped out of the top 10 in the world at the start of November. Were it not for injury, Andy Murray would almost certainly be ranked inside the top 10 and it would not have been a surprise if he was actually ranked at #3 behind Novak Djokovic and Rafael Nadal. Until his injury-problems, Milos Raonic spent the best part of three years as a top 10 player. Jack Sock has been dreadful in singles this year, but it is worth remember that it was only twelve months ago that he broke into the top 10 for the first time.

There are admittedly some major misses in their list. Benoit Paire has barely even threatened the top 20, let alone reaching #2. If we are being completely honest, it is difficult to look back and even spot anything that might have suggested he would go on to achieve the success that L'Equipe expected. It is easy to suggest that Bernard Tomic's attitude problems were clear, even back in 2013, and maybe his current plight might not be all that surprising.

Ernests Gulbis had just had a season where he raised his ranking from #138 to #24 with titles in Delray Beach and St Petersburg, a victory over Andy Murray and twice pushing Nadal deep into deciding sets in Indian Wells and Rome. He would start 2014 with titles in Marseille and Nice before that run to the French Open semi-final. That was obviously as good as it got and he seemingly lost interest in tennis after that, but again, it was maybe a slightly risky pick, but not completely ridiculous.

Jerzy Janowicz has not even played a match this year, but having burst onto the scene with a final at the 2012 Paris Masters, a QF in the Rome Masters in 2013 and pushing Andy Murray in the 2013 Wimbledon semi-final to close the year just outside the top 20 at the age of 23, a top 10 prediction doesn't seem an absurd suggestion. Injuries have derailed his career, although whether he would have gone on to become a top 10 player, who knows.

Looking at the actual 2018 top 10, there are a number of obvious names missing - Roger Federer, Rafael Nadal and Juan Martin Del Potro being the most glaring omissions. I am not going to mock L'Equipe for excluding Roger Federer, given that back in 2013, I myself wrote:

"It could well be that Roger Federer has played his last ever match as a top two player. His victory at Wimbledon was a glorious reminder of why we should not write him off, but could it be a final hurrah?"

At the time, he had just dropped down to #7 in the world following a year where he had lost to the likes of Sergiy Stakhovsky, Daniel Brands and Federico Delbonis. To suggest that he would not still be in the top 10 five years later was not ridiculous.

One assumes the exclusion of Nadal was based on the assumption that his body would have finally succumbed to the many injuries that he would suffer, although that has clearly not been the case.

The exclusion of Del Potro is a slightly strange one though. At the time, he had been a fixture in the top 10 for almost 18 months. It seemed as though he had overcome his first wrist injury and whilst he might not have been looking as though he could return to the very top, a top-10 ranking seemed quite likely.

However, overlooking the likes of Marin Cilic, John Isner and Kevin Anderson seems entirely plausible. Isner had generally lurked around the #20 spot in the rankings and would be 33 in 2018. Cilic had never reached the top #10 and had just returned from a doping suspension - I suspect few would have predicted that he would become a fixture in the top 10 and a Grand Slam champion. Kevin Anderson had only just peaked at the #20 spot by late 2013 and there was little sign that he would become an established top 10 player at the age of 32 five years later.

So, whilst it is easy to look at L'Equipe's prediction now and mock it, at the time, it may not necessarily have looked the worst prediction.

And I suppose that no article about this would be complete without opening myself up to future ridicule, so it only seems right to make a prediction for my ATP top 10 at the end of 2023. No doubt I will look back in five years time and wonder what on earth I was thinking, but here it is: