TMGs vs Cuckfield, 11th July, 2021 by Delbert Covillo

A statistical analysis of winning factors and dynamic match prediction for the Mighty Greys vs Cuckfield CC.

Delberto Covillo1

1Universidad de los Poderosos Grises

First Submitted: 30th July, 2021.

Accepted: 15th September, 2021.

Published: 24th October, 2021.

Updated: 26th October to include a confidence interval band in the dynamic prediction tool and results from a pilot study of 3 matches for match predicting metrics.

Citation: Covillo, D (2021). A statistical analysis of winning factors and dynamic match prediction for the Mighty Greys vs Cuckfield CC, Annals of Improbable Research, 4(4).

Abstract

It can be said that cricket and statistics are like spring rolls and soy sauce – they really go well together. In this paper, an analysis of a Sunday cricket match was conducted with particular consideration of the winning factors in a match and a new statistical tool was developed for dynamic match prediction during the second innings of a cricket match. Some winning factors were shown to be accurate predictors of the match outcome, including hitting more fours and scoring more runs in the first 10 overs, and the dynamic prediction tool was shown to represent the overall state of the game quite well. Further work is required to develop the dynamic prediction tool in order to make it more widely available and further heighten the pleasure for stakeholders of cricket matches.

Introduction

Cricket is a complex, nuanced game where detailed statistical analysis and cutting edge technologies are employed to heighten the pleasure levels of all stakeholders including players, commentators and spectators (Leamon and Jones, 2021) although until recent times such fantasticals have been reserved for the professional game. Nowadays however, they permeate into the lower, more cultured levels of the game, including Sunday cricket.

In this paper, an analysis of a Sunday cricket match was conducted with particular consideration of the winning factors that can be used to predict such matches as has been outlined in the literature. The aims of this study were to a) investigate if those winning factors were indeed indicative of a positive outcome for the team batting second (the Mighty Greys) – if so these may be strategically employed in future games, and b) develop a new statistical tool for dynamic match prediction during the second innings of a cricket match in order to further heighten the pleasure for stakeholders.

The match of interest here was played between the Mighty Greys – a Brighton based club who consider themselves ‘more than just a pub team’ (Mighty Greys, 2021), and Cuckfield Cricket Club – a ‘leading club in the Mid Sussex area, providing first-class facilities, a thriving community-based membership, a welcoming environment and opportunities for fulfilling the potential of all’ (Cuckfield CC, 2021). This match was played on the 11th July, 2021 at Cuckfield’s home ground – their second pitch overlooked by the manor house. The weather was sunny and warm with a gentle breeze. In attendance was a curious horse in the nearby paddock.

Methods

Towards achieving the first aim of this study, a number of winning factors were analysed from the match to see if these were indicative of victory for the Mighty Greys. These winning factors included a comparison of the following metrics:

  • Number of 4’s scored in the innings.
  • Number of 6’s scored.
  • Runs scored by top 4 batsmen.
  • Runs scored in first 10 overs. Note that this was scaled from the first 5 overs as used in the T20 format.

These factors have been shown to be significant predictors for winning a match for the team batting second in either a 50-over ODI or a T-20 match (Schaeffer, 2018).

Regarding the development of a new statistical tool for dynamic match prediction, a simple technique was developed to extend the work of Shah (2017) which utilises the ‘par score’ as determined by Duckworth-Lewis (DL) formula to make a prediction for the match outcome during the second innings of a match. The method presented by Shah, involved calculating the probability of victory for the team batting second according to the following conditions:

However, this approach is limited in that it does not relate directly to the run target set in the actual match being played and hence does not provide a sufficiently accurate prediction in the author’s opinion. Yet the premise of utilising the well-established statistical Duckworth-Lewis dataset (or in the case of this study, the revised Duckworth-Lewis-Stern, or DLS method) seems sound and hence a modified version of the Shah approach has been developed in this study.

Results and Discussion

The match itself got off to an interesting start, with an insistence on the home side (Cuckfield CC) batting first. It is hypothesised that this was a tactical decision to provide a favourable advantage to the visiting Mighty Greys – a friendly gesture, since winning the toss in one-day matches has been shown to advantage the opposition (Bandulasiri, 2008).

The Mighty Greys struck early, with Newland removing Seed for 0 in the 2nd over. The wicket had a healthy slope, and so the Mighty Greys set about exploiting this to their advantage. In particular, Hermitage caused the Cuckfield top order considerable discomfort with full, swinging deliveries then gathering further deviation off the sloping pitch and ultimately removing Sammy H (19) and Sheldon (0), both caught by Covill in the gully leaving Cuckfield at 3-20 in the 5th over. These tactical manoeuvres employed by Hermitage and Covill were both nominated for the PSM as separate albeit intrinsically linked moments – with Covill pipping his partner in crime for the PSM with his sharp catch. Soon after Hermitage struck again with two more victims, one of which was Fenton taking a smart diving catch down the leg-side, leaving Cuckfield on the ropes at 5-41 in the 9th over. A solid partnership was then constructed by the Cuckfield middle order, keeping the run-rate steady with singles and boundaries while the fielders entertained one another – Dom with his ‘Fake throw’, Rick with his ‘Over throw’, Biff with his fabulous football stop of Rick’s overthrow. Dear (54) and Josh D (59) took the total to 127 in the 28th over before Hoare struck with a fabulous straightening ball to clip the top of off, then Mendis sending the jigsaw pieces flying with a quicker delivery going straight through Josh D. The runs kept coming though until the end when Line trapped Paul dead in front for 1, then Mendis finished off Crear on the last ball of the innings – Cuckfield all out for 187 in the full 42 overs leaving the Mighty Greys 38 overs to bat. Hermitate (4-30 off 8), Mendis (2-12 off 4) and Newland (2-24 off 8) were the pick of the bowlers, although it was a rather solid all-round effort from all the bowlers.  

Tea was a brief affair since all 42 overs were consumed in the first innings and there was an important football match that both sides wanted to return home for after the match. Line and Flynn strode out to the wicket for the Mighty Greys, and Line started the innings well with a sturdy flick for 2 before being caught at mid-off off Sear at the end of the first over. Flynn had a slow start, taking 12 balls to get off the mark causing a strange reaction from Hermitage (Figure 1), but then Flynn and Covill pushed things on reaching 79 in the 13th over before Flynn tickled one behind for 29.

Figure 1: The strange reaction from Herimtage after a slow start from Flynn: oddly the only photograph taken on the day.

It is worth pausing here to also appreciate one particular cover drive from Flynn: a majestic, authoritative blow that raced away to the longest boundary as shown in Figure 2 below.

Figure 2: Flynn’s majestic, authoritative cover drive that raced away to the longest boundary for 4.

The fall of Flynn brought Mendis to the crease and this is where the game was taken away from the home side. The pair got busy, running well between the wickets and cashing in on anything wayward putting on 79 in 9 overs, until Covill was smartly bowled by the young Seed for 92. This particular partnership, as shown in the yellow/middle section in Figure 3 below, initiated a divergence from the Cuckfield innings, suggesting that it was the significant factor differentiating the two teams in this particular match. This differentiation was achieved through a sustained increase in the run-rate to 9.0 runs per over during this partnership. Even the fall of Ashton (2) did not slow the rate below that of Cuckfield as Mendis and Day carefully massaged, then brutally dispatched the bowling to take the Mighty Greys to victory in the 29th over with a classy straight drive for four off the back-foot by Mendis finishing things off.

Figure 3: A comparison of the run rates for both sides in the key stages of the match.

The retrospective analysis of the winning factors of the match yielded results presented in Table 1 below. In all cases except the # 6’s scored, these were indeed accurate indicators of victory for the Mighty Greys batting second. Of course, this was only the outcome from one match and so any generalisations here are limited without further analysis of larger statistical dataset.

Table 1: Comparison of the significant predictors of winning the match. Brackets indicate mean average of the runs scored for each batsman. Green indicates an accurate predictor, amber indicates parity between the teams.

A pilot study was then conducted on three other matches played this year against Cutter’s Choice, Mayfield and Headliners. The results of this analysis can be seen below in Table 2, indicating a good prediction for the match outcome using these metrics, and so it is recommended to conduct a larger analysis on this approach, and to consider using these as strategies for match play.

Table 2: Results from a small pilot study to carry out a further comparison of significant predictors for winning the match.

The comparison between the runs accumulated by the Mighty Greys (batting second), Cuckfield and the modelled ‘Par Scores’ (determined using the DLS method and the Required Run Rate at the end of each over) can be seen in Figure 4 below. Also shown in this figure is the probability of TMGs victory along with a confidence interval. While this probability function is a rather simple linear model, it seems to capture the transient nature of the match, in particular the significant 3rd wicket partnership by Mendis and Covill which is seen to increase the probability of victory for the Mighty Greys from the 12th over when the runs accumulated exceeds the DLS Par Score for the first time. From there, the probability of victory steadily climbs until the end of the innings.

Figure 4: A comparison of the runs accumulated by the Mighty Greys (teamB), Cuckfield (teamA) and the modelled ‘Par Score’ as determined using the Duckworth-Lewis-Stern (DLS) method and the Required Run Rate Par Score. Also shown is the probability (with confidence interval) of TMGs victory as calculated at the end of each over taking into account the run target set by the team batting first and the DLS Par Score.

Conclusions

The match itself was a pleasant affair. Some good cricket was played and there was good spirit shown by both sides. As always, the setting was lovely in Cuckfield, the wicket was good and the sun was shining.  An analysis of the significant predictors showed that these particular metrics were (mostly) accurate predictors of the final outcome suggesting that there may be some merit in using this as a strategy for future matches although a limitation of this work is that the analysis was conducted retrospectively and so hasn’t been tested in situ. Furthermore, it is recommended that further analysis be conducted on a larger statistical dataset. In particular, there may be some merit in taking a strategic approach (when batting second), to:

  1. Score as many fours as possible throughout the innings.
  2. Prioritise high scoring for the top 4 batsmen.
  3. Ensure that the total runs accumulated in the first 10 overs of the run chase exceeds that of the team batting first.

Similar, but subtly different approaches are also suggested for batting first and for the different forms of one-day matches as outlined in Shaefer (2018).

The method used to calculate the probability of victory for the team batting second seems to capture the balance of the match relatively well. There is a bias towards the team batting first in the early stages of the run chase due to the proportional influence of the difference between the current score and the target score (i.e. the stage of the game). This would seem a fair assumption, since anecdotally most teams have a tendency to choke, choke, choke when chasing and the Mighty Greys are no exception. Also, the model predicted that the probability of success for the Mighty Greys didn’t exceed 90% until the 26th over. At that point, only 17 runs was needed off 12 overs (72 deliveries) with 6 wickets in hand. This may seem a rather conservative estimate, although again the tendency to choke, choke, choke may render this a fair bias, even in these very late stages of the game. In essence, this would suggest that one in ten innings in this situation would result in a catastrophic collapse to lose the game, and it is the author’s opinion that this does not seem unreasonable. That said, the model could be expanded to include a poly-linear approach to make adjustments to the weighting of the stage of the game and the strength of any tendencies to choke, choke, choke. The model could also be tweaked by mapping it to further historical or team-specific historical data or by using artificial intelligence techniques to predict and include the influence of other factors on the outcome of the match as used by Kaluarachchi and Varde (2010) and Rajesh et al (2021). Further work is also recommended to make this tool more widely available. It may be a playful tool at all levels of the game and help to further heighten the pleasure experienced by stakeholders.

References

  1. Bandulasiri, A (2008). Predicting the Winner in One Day International Cricket, Journal of Mathematical Sciences & Mathematics Education, 3(1). Cuckfield CC (2021). Cuckfield Cricket Club website. Available at: https://www.cuckfieldcc.co.uk Last accessed: 25th September, 2021.
  2. Kaluarachchi, A and Varde, A (2010). CricAI: A Classification Based Tool to Predict the Outcome in ODI Cricket, Fifth International Conference on Information and Automation for Sustainability, Colombo, Sri Lanka, 17-19 Dec. 2010.
  3. Leamon, N and Jones, B (2021). Hitting against the spin. Constable: London, UK.
  4. Mighty Greys (2021). The Mighty Greys website. Available at: http://www.mightygreys.com Last accessed: 20th September, 2021.
  5. Rajesh G, et al (2021). Dynamic cricket match outcome prediction. Journal of Sports Analytics, 7(3), pp. 185-196.
  6. Schaefer, M (2018). Match statistics that discriminate between winning and losing teams in ODI and T20I cricket. MA thesis, University of the Free State Bloemfontain.

Appendix 1: Scorebook pages for the Cuckfield (batting first) and Mighty Greys (batting second) innings.

Figure A1-1: Cuckfield Innings (batting first).
Figure A1-2: Mighty Greys innings (batting second).

Appendix 2: Party Seven Moment trophy awarding ceremony.

PSM winner Delberto Covillo- for his sharp catch in the gully completing a well-excecuted plan with Hermitage to exploit the slope with full swing bowling.

8 responses to “TMGs vs Cuckfield, 11th July, 2021 by Delbert Covillo

  1. Somehow Del manages to bring his work to a match report and produces the most scholarly MR ever which could also change the way match prediction is done. Great work Dr. Greysman.

  2. Pure brilliance, Prof Delberto. What’s the Annals Improb Res impact factor though? You deserve higher🤓. Amazing

  3. Pingback: The Pinstickers Guide from the Flat Capster Napster, Ian Sewell November 2021 | mightygreys.com - home of the mighty greys cricket club·

  4. Del this is amazing! I’ve just finished Hitting Against The Spin and honestly this report should have been included!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.