Posts

AUSTRALIA’S CAUTIOUS SPIN APPROACH EXTENDS THEIR LOSING RUN IN ASIA

As Australia slumped to a series defeat against Sri Lanka in Galle, their shortcomings against spin bowling were exposed in ruthless fashion. Now 2-0 down in the three match series after losing the first Test in Pallekele as well, Steve Smith’s side are set to lose their spot at the top of the world rankings as their wretched record in Asia continues.

Going back to the start of the India series in 2013 – the tour that became infamous for coach Micky Arthur’s ‘homework-gate’ fiasco – Australia have now lost eight consecutive Test matches in the subcontinent and the UAE. The common denominator across all eight matches has been the dominance of the home sides’ spin bowlers, as compared to the limited impact seen by the various spinners Australia have tried in the traditionally spin-friendly conditions.

The eight defeats in India, UAE and Sri Lanka have seen Australia lose a total of 159 wickets with 131 – or just over 82% – taken by spin bowlers. Conversely, of the 111 wickets that Australia have taken in those matches, only 51 of them (46%) have gone to their spin bowlers. As one would expect, Nathan Lyon has been Australia’s most prolific spinner in that period. He has taken 27 of the 51 wickets having played in seven of the eight defeats but the dismissals have come at a worryingly high average of 46.7. His strike rate of 67.5 is adequate and only a few balls higher than his career strike rate of 61.9 so the high average illustrates that batsmen are going after him, meaning that each wicket is far more expensive than his career average of 32.8.

The approach that Asian batsmen have taken to thwarting Lyon is something that Australia would do well to take on board. During the first innings of the Galle Test, Sri Lanka’s batsmen attacked 55% of all deliveries bowled by Lyon and debutant Jon Holland. When Australia batted in their first innings, that figure was just 38% and seven of the nine batsmen who fell to spin did so playing a defensive shot – or, in the case of Usman Khawaja, no shot at all. Sri Lanka were even more aggressive against Lyon and Holland in the second innings, attacking 63% of the deliveries they bowled and while Australia’s figure for their second innings was up to 47%, it nonetheless illustrated the tentative approach they had to the turning ball throughout the match.

There is one mitigating factor to consider and that is the state of the match when Sri Lanka started their second innings. Having bowled out Australia for just 106 and holding a lead of 175, the home side could bat freely without scoreboard pressure and targeting Holland, a left-arm spinner playing in his first Test, would have been a logical step. However, the data shows that Sri Lanka did not target Holland specifically as his percentage of balls attacked was almost identical to Lyon’s.

The difference in mind-set was evident to anyone who watched the extraordinary events unfold on Day 2 in Galle. Australia were tentative and prodding during a collapse that saw them lose 47-8 on their way to being bowled out for 106; Sri Lanka were aggressive and carefree as they put on 237 to bat Australia out of the game. For a side that prides itself on positive, attacking cricket and scored at an average rate of 4.3 runs per over during their most recent home summer, the contrast in Australia’s differing game plans could not be more pronounced.

After Australia lost the first Test in Pallekele, the Chairman of Selectors Rod Marsh lamented that there was little more that they could do in terms of preparation, highlighting the success Australia ‘A’ had had during their tour to India last year and questioning if the extra pressure of a Test match environment was playing on the players’ minds. It is as good an explanation as any for Australia’s failings against spin; batsmen famed for their bravado and aggression retreating into their shell when confronted with high quality spinners on a turning track.

So with the series lost and the number one ranking surely gone, Australia have the third Test in Colombo to salvage some pride. However, if their batsmen continue with the tentative approach to playing Sri Lanka’s spinners, their long wait for a win on Asian soil might yet continue.

ANALYSING ADIL RASHID

England’s stand-out bowler in the first two matches of the five match ODI series against Sri Lanka has arguably been Adil Rashid. In both games, bowling ten consecutive overs, he has recorded his best economy rate in ODI cricket, first 3.60 and then 3.40. He did not take any wickets at Trent Bridge, although he did build the pressure for Moeen Ali to dismiss Dinesh Chandimal, but took two at Edgbaston in a pivotal spell which ensured Sri Lanka set an under-par target that was subsequently chased with ease. While Jason Roy and Alex Hales’ record-breaking partnership stole the headlines, Rashid’s 2 for 34 was possibly the more important performance.

Length

The defining feature of Rashid’s bowling has been his accuracy. Leg-spin is a delicate art in which the slightest error in action or release can cause dramatic misdirection but in the first two ODIs Rashid has displayed ability and skills that suggests he may have taken his game to the next level.

Over the 120 deliveries Rashid has bowled his average pitching length has been 4.88 metres from the stumps and he has bowled 100 of his deliveries in a three metre range between three and six metres. From this length-range he has conceded 52 runs and taken two wickets. He has over-pitched from this range on five occasions and has dropped it shorter on 16, although 12 of those are in the six to seven metre range from which only eight runs have been scored.

Screen Shot 2016-06-25 at 15.42.42

Interestingly, analysing his lengths by match clearly shows how he bowled fuller in the second ODI. Having gone wicketless in the first match perhaps Rashid, striving for wickets, pitched the ball further up in an effort to tempt the batsmen to drive. Having bowled ten deliveries in the three to four metre range in the first match he bowled 17 in the second and bowled a more consistently full length delivering 51 of his 60 deliveries in the three to six metre range, compared to 48 of his 60 in the first. The charts below illustrate this.

Screen Shot 2016-06-25 at 15.42.53

Line

Rashid has also displayed excellent control of his line in the series bowling 85 of his 120 deliveries on or outside off stump from which he has conceded just 51 runs and taken one of his two wickets. As well as bowling fuller in the second ODI Rashid also bowled straighter, as the charts below show.

Screen Shot 2016-06-25 at 17.01.53

After delivering 34 of his 60 balls outside off stump at Trent Bridge he delivered none there at Edgbaston; the result of this was that it forced the batsmen to play at more deliveries and to hit more against the spin through the leg-side. This proved to be the case as Sri Lanka scored 52% of their runs off Rashid through the leg-side in the second match compared to 27% in the first.

Speed

Rashid’s average speed across the two matches has been 49.11 mph which, although slightly faster than his average speed in his last ODI series against South Africa of 48.55, is still relatively slow for a leg-spin bowler. By sake of comparison Piyush Chawla, Amit Mishra and Imran Tahir all normally average in the mid-50s.

Terry Jenner, the leg-spin coach who mentored Shane Warne during his career, advised that “the right pace to bowl at is the pace where you gain your maximum spin.” Our data shows that if Rashid is seeking more turn he should in fact bowl slightly slower still. This is illustrated by the chart below which shows degrees of deviation on the Y axis in relation to the speed of delivery on the X axis. The yellow line represents Rashid’s average speed.

Sacrificing pace for turn is a risky strategy however. The slower the pace the longer a batsman has to adjust his footwork, shot selection and shape. Anyhow, Rashid’s bowling figures suggests there is no need for anything to change. It is useful to know at least, that a slower speed may produce some additional turn.

Screen Shot 2016-06-25 at 19.41.20

Variation and Strategy

The large majority of Rashid’s deliveries are leg-breaks. However, he is not afraid to use his googly and bowled seven in the first ODI and five in the second (deviation to the leg-side displayed in the chart below does not necessarily represent a googly due to deviation caused by natural variation off the pitch and angle of delivery).

The value of the googly to a leg-spinner is pervasive: once a batsman knows a bowler has one and is willing to use it new dimensions are added to the contest. Firstly a batsman is worried about reading the spin and secondly he has to adjust how he plays each delivery. Even if the ball is a leg-spinner the batsman has to be wary of the one that turns back in. The danger of a googly does not end with the delivery itself.

The chart below maps out Rashid’s two spells ball-by-ball and gives clues as to the pattern of his strategy.

Screen Shot 2016-06-25 at 17.33.18

It is noticeable how Rashid does not employ the googly early. This could be because he wants to find his rhythm with his leg-break before moving onto variations. Trying and failing to land an accurate googly releases pressure and for Rashid, bowling in the middle overs, maintaining pressure is imperative. In both matches Rashid spins his early leg-breaks hard and far.

Examining patterns of deliveries is interesting. Googlies are generally followed by big-turning leg-breaks, except in the case of his 45th delivery in the first match, which seems to be a top-spinner or slider. In the second ODI, his first wicket came from the 41st delivery which was immediately preceded by a delivery which went straight on. His second wicket came from the 43rd delivery, which this time was preceded by a googly. Only once has he bowled consecutive googlies but in both matches he ended his spell with one.

***

The limits of analysis such as this is that while our data allows us to look inside the mind of a bowler, only by actually talking to Rashid can we truly understand what his thought processes and strategies are.

This analysis has been based on a small sample size but it has demonstrated Rashid’s growing mastery of his art. If he can maintain this form deeper into the series then we may be able to say with confidence that his game has indeed reached a new level.

THE ANATOMY OF A THRILLER

Freddie Wilde analyses the key moments in the first ODI of the five match series between England and Sri Lanka that ended in a tie.  

Read more

ENGLAND’S PAINFUL ASHES MEMORIES

Sri Lanka are expected to face friendlier conditions in the third and final Test of the series at Lord’s. Having struggled to cope with the seaming and swinging ball at Headingley and Durham, their batsmen should find batting at the home of cricket an easier task.

This season’s County Championship scores at the venue support this. The first innings scores in Middlesex’s home games at headquarters this year have been 376, 423, 354, 203/3, 452 and 468. Pace bowlers have found it hard to trouble batsmen on pitches that have lacked pace and movement.

However, the Lord’s wicket does not prevent decisive Test results as it once did. There were six straight draws between 2006 and 2008, but just two of 15 matches since have been stalemates. Sri Lanka secured one of those draws, hanging on nine wickets down after scoring 453 in their first innings.

That result contributed to England’s run of one win in their last four at Lord’s, with last year’s Ashes hammering a particularly chastening experience in placid conditions. The home side’s bowlers struggled to find the lateral movement to trouble Australia, whose pace men were far more menacing.

A Hawk-Eye comparison of England’s seamers in that match and this series reveals how they were blunted in that contest and lethal at Headingley. It is no coincidence that as conditions improved at the Riverside, so did Sri Lanka’s batting.

England seamers averagev Australia, Lord's 2015v Sri Lanka, Headingley 2016v Sri Lanka, Riverside 2016
Bounce height at stumps (metres)0.790.930.82
Speed (mph)82.0082.4082.40
Seam (degrees)0.550.810.55
Swing (degrees)1.061.460.76

England’s seamers have averaged around the 82mph mark this series, as they did against Australia at Lord’s last year. However, on the lively pitch in the series opener at Leeds, they found far more average seam, swing and bounce, the key factors in their crushing innings win inside three days.

At Durham they carried on where they left off. They fired out the tourists for 101, but Sri Lanka were far more resilient second time around, posting 475 in 128.2 overs.

England’s pace attack produced the same amount of average seam in that Test as they did in the 2015 Ashes Lord’s clash and only marginally more bounce. They swung the ball less overall, a reflection of how conditions eased as the game wore on.

England know they will need to work harder to take 20 Sri Lankan wickets this week. Their seamers took just five between them as Australia piled up 566/8 last year and whilst England can be expected to bat better than they did in that contest, a major challenge faces the bowling unit.

RATING ANDERSON’S MASTERCLASS

Just how good was England’s bowling at Headingley? Sri Lanka’s batsmen struggled in tricky conditions against a skilled attack and CricViz can measure how much more dangerous the hosts’ seamers were than their counterparts.

The BatViz model analyses ball tracking data to produce wicket and run ratings for every ball. We conduct a nearest neighbour analysis of the six Hawk-Eye categories that comprise each ball: speed, line, length, seam, swing and bounce.

This process, counting the runs and wickets associated with the 1,000 most similar deliveries in our database based on those categories, allows the measurement of wicket threat and ease of scoring.

England’s bowlers had an average wicket probability of 1.87% per ball, Sri Lanka’s 1.38%. The top five bowlers in this ranking were members of the home attack, led unsurprisingly by James Anderson (2.13%).

Average wicket probability per ball bowled 
Bowler%
Anderson2.13
Stokes1.90
Vince1.83
Finn1.74
Broad1.71
Eranga1.60
Pradeep1.55
Chameera1.50
Herath1.45
Moeen1.18
Mathews1.13
Shanaka1.12

The Hawk-Eye data from the first Test testifies to Anderson’s mastery of seam and swing. Of the frontline seamers, only Shaminda Eranga had a lower average speed, but the Lancastrian’s 81mph is plenty when combined with lateral movement that no other paceman in the world can match.

Eranga actually swung the ball more on average, but Anderson’s ability to move the ball both ways is crucial. 16 of the 25 biggest inswingers (as faced by a right-hander) were delivered by England’s talisman.

Dangerous swing bowling is partly about controlling the movement in favourable conditions and Anderson is adept at finding just the right amount. Eranga bowled 13 of the 20 biggest outswingers (to right-handers) in the match, but these were not of the right line or length to trouble the batsmen.

Anderson can famously switch between inswing and outswing with little discernible change in action, a skill that is especially useful in the context of expert seam bowling. He possessed the highest average seam movement in the match.

Average wicket probability per ball faced 
Batsman%
Herath2.38
Mathews2.08
Karunaratne2.07
Mendis1.98
Finn1.94

Applying the wicket probability ratings to each batsman, the struggles faced by the visiting batsmen become clear. Of frontline batsmen the highest average wicket probability per ball was faced by Angelo Mathews (2.08%) and Dimuth Karunaratne (2.07%).

That the best was kept for the two most experienced opposing batsmen says much about the efficiency of England’s bowling. Anderson’s unique combination of seam, swing and accuracy, a combination that has brought him 443 Test wickets, was too good for the tourists.

4TH ASHES TEST ANALYSIS

On the first morning of the Trent Bridge Test match, Australia batted first and at the first drinks break were 38 for 7, their top seven all back in the pavilion. England started batting 50 minutes later and an hour into their innings were 30 for 0. The Ashes were, barring a freak turnaround, already on their way back to England.

What happened? Why did Australia collapse so dramatically? Great bowling? Poor batting? A green-topped, bowlers’ dream that simply handed the match to the captain lucky enough to win the toss?

Why was the first hour of England’s innings so different to that of Australia’s an hour and a half earlier?

Did the conditions get easier?

A little. The ball kept swinging; the average deviation in the air when the Australians bowled was 2.1 degrees, slightly more than the 1.9 degrees when England bowled. Both teams swung roughly 60% of the balls they bowled by more than 1.5 degrees, the amount of swing that starts to have a significant impact on a batsman’s performance.

There was more seam movement when Australia batted. 31% of the balls in the first hour deviated by more than one degree off the pitch, whereas the figure when England batted was 18%. The average seam movement faced was 0.7 degrees for Australia and 0.5 degrees for England.

However, this was part of a pattern in the series. England’s seamers got more lateral movement off the wicket and were more accurate throughout; the Australian pacemen consistently bowled a little quicker on average and got more movement in the air.

Conditions had got a little easier by the time England batted, but not drastically so.

Did England out-bowl Australia?

England, and Stuart Broad in particular, bowled very well. A traditional good length in Test cricket is usually defined as balls pitching six to eight metres from the stumps. These are the balls that have the lowest average (runs per wicket), regardless of pitch, conditions and opposition. When the ball is moving around in the air and off the wicket, the metre or so fuller than that (5-6m from the stumps) becomes equally, if not even more, dangerous. England landed just over 60% of their deliveries in these areas, and these balls accounted for all but one of the wickets in that innings. Australia though, bowled even more balls on these lengths, 67% in their first 11 overs.

The England bowlers also bowled unusually straight. Their average line was middle and off, very straight for Test cricket; 49% of the balls they bowled were within the line of leg stump and six inches outside off stump. It was the balls on these lines that did the bulk of the damage to the Australian top order.

Australia bowled significantly wider. Their average line was six inches outside off stump – they put 52% of their deliveries wide of this mark, compared to 35% of England’s. This allowed England’s batsmen more easy leaves than the Australians got, nearly half as many again.

So, as was the case all summer, better areas and more movement off the pitch from England, albeit at a slightly slower pace. When the pitch offered assistance, England were the more dangerous attack. When it didn’t, Australia’s pace and swing posed the greater threat. Trent Bridge was no minefield, but nor was it the pitch where you wanted your great strength to be taking the pitch out of the equation.

Did Australia go too hard at the ball? Play too many shots? Not leave well enough?

Using the BatViz system we can compare how Australia played the deliveries they faced with how an average Test side would have played them.

Given the balls they faced, we would have expected 25 attacking shots in the first hour. Australia played 22. BatViz projected 14.5 balls to be left; they played no shot on 19 occasions.

For comparison, we would have expected England to play 24.5 attacking shots and they played 21. They got more balls to leave, as Australia bowled wider than England. 17.5 leaves were forecast – they actually left the ball 25 times.

First hour BatViz shot analysis   
AustraliaEngland
Attacking shotsExpected2524.5
Actual2221
LeavesExpected14.517.5
Actual1925

There therefore seems to have been little difference in the overall intent of the two sides and it is worth noting that only three of the seven Australian wickets fell to attacking shots. That might be three too many given the situation and conditions, but it is easy to criticise attacking shots when they don’t come off and applaud them when they do: England showed a very similar level of attacking intent and left the ball marginally better.

Was it therefore poor shot selection and execution?

Given the balls received, BatViz projected 11.9 false shots – edges and misses – from the Australians. There were 19. For comparison, we would have expected eight false shots from England and there were just six (five misses and one edge). On average in 11 overs of Test cricket there would be 4.5 false shots.

England had to play fewer balls and the balls they played at moved a little less. They also played them better than par, whereas the Australians underperformed against the balls they faced.

First hour BatViz false shot analysis  
AustraliaEngland
False shots - predicted11.98
False shots - actual196
Missed105
Edged91
Wickets from edge60

Even so, 19 false shots to six can’t be the difference between seven down and no wickets very often.

So were Australia just unlucky?

They certainly were to an extent. Of their 19 false shots, nine were edges (47%). Generally only about 37% of false shots are edges, so they were unlucky to nick almost as many as they missed. England played and missed five times for their solitary edge.

About 15% of edges result in a wicket. Australia’s nine edges produced six wickets, so the picture of a perfect storm is forming. The pitch had good carry, so there was little chance of edges with the new ball falling short of the slip cordon. The England bowlers’ areas were good, so the edges produced were more likely to find catchers than fly to safety. Two wickets in the first over meant that for the remainder of the innings Alastair Cook employed five or six catchers, so any edge was likely to find a catcher rather than a gap.

And what about the catching?

The first hour brought five slip catches, the innings as a whole comprised eight. Every single one of the chances offered were held, including Ben Stokes’ stunning one-handed grab.

On average in Test cricket roughly 70% of slip catches are caught. PlayViz goes deeper by rating chances according to where they come and the reaction time the fielder has. In doing so we can estimate that the five chances presented in the first hour would normally have resulted in two or three wickets (2.65 to be exact): the English cordon hugely over-performed.

A bit of everything?

The Australians were hit by a perfect storm of several factors, each multiplying the effect of the others that together created a manic 11 overs that devastated their Ashes dreams.

The ball swung and seamed enough to trouble the batsmen. The bowlers – Broad in particular – used the conditions very skilfully, and allowed the batsmen little respite. The Aussies didn’t cope with the moving ball particularly well and didn’t have a lot of luck when it came to playing and missing. A pitch with good pace and bounce ensured the edges carried and early wickets meant a packed slip cordon. The chances went to hand and the fielders caught exceptionally well. 38 for 7. Ashes gone.

3RD ASHES TEST ANALYSIS

After ending the Lord’s Test with a batting collapse, England inflicted one of their own on the opening day of the third Test. Conditions were rather different at a damp Edgbaston to those which Australia prospered in at Lord’s and James Anderson duly delivered a seam bowling masterclass.

The Lancastrian’s 6/47 helped bowl out Australia for 136 inside 37 overs, a collapse which saw the tourists’ starting win probability of 31.9% reduce to 11.7% at the change of innings. Anderson’s lateral movement proved too difficult for Australia to deal with, but their shot selection played a major part in their slump.

The analysis of Hawkeye data for each delivery reveals how the Australian top order got caught in two minds when dealing with Anderson’s movement. As well as producing the average wicket probability and run total for each ball based on similar deliveries in the CricViz database, BatViz can analyse the type of shots played (see below table).

Anderson delivered 88 balls in Australia’s first innings. Of these, 10 had an attacking shot percentage between 40% and 49% – based on the similar delivery evaluation, these type of balls are typically attacked somewhere between 40% to 49% of the time.

All four of Anderson’s top order wickets – David Warner, Adam Voges, Mitchell Marsh and Peter Nevill – fell in this range. Warner (playing defensively too late), Voges (withdrawing his bat to leave too late) and Nevill (no shot) paid the price for tentativeness; Marsh (flat-footed drive away from body) ill-advisedly took the attacking option.

BatsmanShot typeDismissalLeave %Attacking %
WarnerDefensiveLBW3.745.3
VogesNo shotCaught31.344.9
MarshDriveCaught38.748.9
NevillNo shotBowled3.441.1

The type of deliveries Warner and Nevill received were clearly ones to play at. BatViz takes into account the speed, line, length and deviation when picking out similar deliveries and these two balls that were on the stumps were left alone just 3.7% and 3.4% of the time respectively.

This highlights the seeds of doubts that Anderson can plant in a batsman’s mind, despite the lack of extreme pace. Warner’s wicket was 82mph, Nevill’s 83.6mph. His day one wicket burst against a confused batting line-up was the crucial factor in England’s victory, a template that was followed emphatically at Trent Bridge by Stuart Broad’s opening salvo.