Stick The DPI Through The PRISM Of The Contextual Stats > Cricket News, cricinfo, mobilecric, cricbuzz, livescore and more

Cricket news - Batting in IPL through the prism of contextual stats

RCB fall short in the middle overs despite having Kohli and de Villiers

Debates in cricket revolve around context, but cricket statistics hardly do. Scoring 10 runs per over is expected in the slog overs, while scoring at the same rate in the powerplay exceeds expectations. Yet, the archaic numbers automatically borrowed from Test and ODI cricket never include these important conditions-based riders, and continue to be used by broadcasters and fans.

To put T20 batting in context, we need to establish our expectations from each over. It is well-known that teams capitalise on the fielding restrictions, and then slog in the death overs. Why don't we begin with looking at the strike rates each over in an IPL game? This helps us gauge what the "normal" scoring rate for an over is.

This shows that a T20 innings is a story in two parts: in the Powerplay, teams start slowly in the first two overs, and then go hard for the next four. In the 7th over, there is a new beginning. They start slow again, and the scoring rate goes on increasing till the final overs. These, then, are the "expected" scoring rates in each over.

What about the loss of wickets? Similar to the expected strike rates, we can glance at the chances of losing wickets in a given over:

This follows the same story. Teams conserve wickets in the seventh over, and then take more risks progressively. The death overs see caution thrown to the winds, as batsmen swing without much time left.

With these two "baselines" now in place, we can see how each team does compared to the average, in each over. This graph shows team performance in 2018, compared to the "average" scoring rates in IPL games from 2014 to 2018. Let's take Chennai Super Kings as an example.

The meaning of the bars is simple: a value of 1.00 means a team scored as fast as the expected strike rate in that over, just meeting expectations. Higher is better. In the first five overs, CSK score at par with the expected rate, so their values hover around 1. In the final three overs, they score 32, 28 and 21 per cent faster than the average team scores in those overs.

A similar graph can be made for wicket loss rates. A value of 1.00 here means that a team is as likely to lose a wicket in that over as an average team. Lower is better. Take RCB as an example. The graph shows quite a few tall buildings. In the 11th over, for instance, they have a value of 1.63, which means they are 63 per cent more likely to lose a wicket than the expected rate.

To facilitate an easier study where teams scored their runs and lost their wickets in IPL 2018, let's look at batting in the commonly defined three phases of an innings: Powerplay, middle overs, and death overs. We will follow the same strategy for each "phase": look at the scoring rate of an average team in that phase, and compare a team's numbers to that. The numerical nitty-gritty can be forgone, and we can look at a nice graph.

The lower left of this graph corresponds to conservative batting: not losing wickets, not scoring runs. The lower right is ideal, and the top left means a weak batting side. Let's have a look at the first six overs:

This enables us to look at each team's batting strategy during the Powerplay. Remember that a value of 1.00 is par for the course. This shows us how relatively likely a team is to lose a wicket, and how fast they score relative to the expected run rate.

We immediately see that KKR, DD and KXIP were aggressive during the initial phases of play, scoring heavily, while also taking risks and losing wickets.

Chennai were extremely conservative, with a low wicket loss rate and a low scoring rate, starting slow and steady. They scored only 5% faster than the average side, but were 10% less likely to lose a wicket.

RCB were the best combination of scoring freely while also not losing many wickets. On the other hand, Hyderabad struggled to find a settled opening combination, and it shows in their very low scoring rate.

Moving on to the middle overs.

Here we see clear shifts in performance. CSK are now more likely to lose a wicket compared to an average team. Bangalore make a high jump owing to a middle-order muddle after Virat Kohli and AB de Villiers. While SRH, buoyed by Williamson, get partnerships and score 5% faster than the average team.

On the other hand, Punjab's non-existent middle order to follow an in-form KL Rahul shows. Partnering them are the Royals, trudging at scoring and more than 25% more likely to lose wickets compared to the average. KKR decrease their scoring rate in context, but also are less prone to wickets falling.

Lastly, in overs 16-20, we again see changes in strategy.

Five out of the eight teams jump across to the undesirable quadrant where their wicket loss rates are high, and their scoring rate is low. The 2018 IPL saw most sides underperform at the death. CSK, owing to a spate of late heists, rule the slog overs. They score 16% faster than the expected rate, and are a whopping 52% less likely to lose wickets at the end. Delhi too improve on both counts, just behind Chennai.

In conclusion, some insights into strategy these graphs clearly show:

RCB fall short in the middle overs despite the Kohli-AB combine. A lack of support in the batting unit is clearly visible. Teams should attack them in overs 7-15. SRH are the clear "bowling team", consistently under-performing in all three phases of an innings. CSK start conservatively, have a low-scoring middle order, and are almost untouchable at the death. They should be targeted in the Powerplay. DC need to improve and settle their opening combination. With Prithvi Shaw out of form and Shikhar Dhawan an important India player, just ahead of the World Cup, they might find this a tough gap to plug. KXIP have nothing to write about in terms of batting after their opening partnership exits.

Live Home Series
Player Schedule
Points Table Cricket Rankings
Cricket Clothing Cricket Equipment
Cricinfo Apps Download