Category Archives: statistics

India Batting Depth

Posted on by

While writing my post on India chases, I wondered how much depth India does have in its batting ranks. India is known for its strong batting lineup. So I thought I would take a graphical look at the most recent matches starting from the first World Cup match against Bangladesh and continuing to the latest match against West Indies. The big caveat being that many team members are missing from the squad after the World Cup.

What this shows is that the top order has mostly done its work. If you look at the World Cup batting aggregates, Tendulkar (482), Gambhir  (393) and Sehwag (380) lead. The better averages belong to Yuvraj and Raina, but primarily because of the not-outs. We don’t have many data points beyond the 6th wicket partnership, but among those, very few have been significant. In fact 2 matches (the one against the Pakistan and the 3rd ODI against WI) have all but one of the 20+ partnerships beyond the 6th wicket.

During this period, there have been a few collapses of the lower order too. India made 33 for the last 6 wickets against England, 29 runs for 8 wickets against South Africa, and 36 for 5 against West Indies in the World Cup, not to mention the disappointing performances by the lower order in the last two matches against the West Indies. India needs to do a lot more work to have a deep batting lineup that extends beyond the middle order.

On another note, you can find examples where the middle order has rescued India from a top order failure. But it is important to realize that the middle order (No. 5 and No. 6) are supposed to be people who can bat. And they are there to contribute to the team. If a team never fails to make a score when its top 3 fail, then that is a strange team indeed. It doesn’t mean that a team bats deep.

Sachin Tendulkar and Ricky Ponting

Posted on by

Sometime back, I tried to analyze whether Ricky Ponting could overtake Sachin Tendulkar in Tests. At that time, the stats stood at

Player Tests Runs Avg 100s Age Comparison
Tendulkar 157 12499 54.58 41 - 4 tons ahead
Ponting 131 10948 56.43 37 -1y 8m 1551 runs behind

Today, it stands at

Player Tests Runs Avg 100s Age Comparison
Tendulkar 166 13447 55.56 47 - 8 tons ahead
Ponting 142 11859 55.67 39 -1y 8m 1588 runs behind

Ponting still looks in striking distance of Tendulkar’s total run aggregate. Assuming that he plays for two years after Tendulkar retires, he only has to get some 800-odd runs a year. Since Australia plays more Tests than the average nation, this is somewhat easy. Consider that in 2009, Ponting only scored at a poor average of 38.77, but because Australia played 13 Tests, he ended up scoring 853 runs for the year. You will also notice that Ponting has played two more Tests than Tendulkar in the same period. But if Ponting overtakes Tendulkar, it is also likely that Kallis will overtake Ponting very quickly. So it would probably be a very short reign at the top.

Overtaking the number of centuries Tendulkar scored looks more iffy. Since the beginning of 2007, Ponting has only scored 6 centuries. On the other hand, no one expected Tendulkar to go on a ton-hitting spree at age 36. So Ponting may be able to tap something similar. Also, India’s Test calendar looks barren at this point while Australia have Test matches lined up against New Zealand, Pakistan and England for this year, so that 8-ton deficit could see some reduction. All the other contenders are way behind at this point.

On the ODI side, there is simply no comparison. Tendulkar is so far ahead that there is no credible rival for the top spot (runs, centuries) for the near future.

Sehwag’s Strange Statistics

Posted on by

In his early career days, when people used to talk about Sehwag, some compared his physique and batting style to Tendulkar, implying that he was a poor imitation. Over the years, Sehwag has created a different legacy for himself even outdoing Tendulkar in some cases with his triple centuries and almost-borderline manic batting in huge innings.

What was always struck me as strange is Sehwag’s great Test average (52.50) compared to his terrible ODI average (33.98). A few have pointed out that the different ODI fielding settings may have something to do with this. Plus Sehwag’s brain-fade when he tries one too many times to assert his authority over some bowler.

But even Sehwag’s Test average seems strange. He has the most Test double-centuries by an Indian. He could have had the most Test triple-centuries if he had made a few more runs yesterday. And when he gets a century, he invariably makes 150+. So why is he still in the low fifties when he could be in the sixties or upper fifties?

To understand that, I compared Sehwag with another great batsman of the same era, who has a similar average and played roughly the same number of Tests. Sehwag has 6248 runs at 52.50 in 123 innings. Kumar Sangakkara has 7545 runs at 55.47 in 147 innings. Not exactly the same, but will do for our purposes. Here is the breakdown of their innings:

sehwag's innings

sangakkara's innings

One of the first things you will catch is that Sehwag has a much higher percentage of scores below 25 (46% versus 41%) and below 50 (71% versus 64%). Sehwag has 30 scores under 10, while Sangakkara over more innings has 28. But once you start looking at scores above 50, Sehwag looks much more attractive. He has been able to convert more fifties into hundreds and those hundreds into huge scores that have made a big difference to India’s results.

So for the opposition, Sehwag is easier to get early. But once he goes past that fifty, start saying your prayers because he is going to make you pay dearly.

India’s Test History

Posted on by

India just marked their 100th Test victory with a thumping win against Sri Lanka. It was interesting reading the statistics of India’s Test history.

India is the nation that has taken the most number of matches (432) to reach 100 Test wins. The previous entry was Pakistan at 320 Tests. Sri Lanka have 60 wins in 191 matches and surely should make it before 300. Of course, they had a lot of help from Zimbabwe and Bangladesh, but still.

The only other nation likely to ever reach 100 wins is New Zealand who have 66 wins from 354 matches and are sure to beat India’s record. Bangladesh (3 wins from 61 matches) has a loooooooong way to go!

India has made a lot of strides in this decade, with more wins abroad than in the rest of their history. Many reasons, but a few I can think of: Sourav Ganguly’s captaincy, Rahul Dravid’s form abroad, Kumble turning around his overseas form, India holding their own against Australia in Australia, the decline of the West Indies, England no longer a threat, Bangladesh, Zimbabwe.

On the other hand, India has had a few setbacks at home. They lost a series against Australia after decades. They have conceded matches against South Africa, England and Pakistan. More aggression at home would have made Win #100 come earlier. We haven’t seen a whitewash for sometime now.

Winning IPL Scores

Posted on by

Twenty20 being a recent innovation, teams and players are still trying to come to terms with strategies and tactics necessary to win the matches. One element of this is what score is good enough? Obviously, one should try to make as large a score as possible and defend even the weakest targets, but what is a par score. To some extent, people have used 150 as the par T20 score. If you exceed that, you are well-placed. If you didn’t, well, your bowlers better perform well.

But I decided to do a little more poking around. I took all the IPL group matches so far, eliminating the ones that were affected by rain (five of them – Punjab v Delhi, Punjab v Kolkata, Mumbai v Rajasthan, Chennai v Kolkata, Chennai v Punjab) and the tied match (Kolkata v Rajasthan). So totally 50 matches. Statistically, it is not the best sample because it only consider one year’s matches in one country (South Africa). But for starters, it will do for our purposes. I will try to update it with a bigger T20 set (including last year’s IPL and international matches) at a later time and blog about any discrepancies.

OK, so take a look at this chart. This plots the score ranges (made by teams batting first) and the success %age.

winning-ipl-scores

As would be intuitively evident, the higher the score, the greater success percentage. The key point is that below the 160 mark, you lose more matches than you win. But then the succes rate spikes. You are twice as much likely to win a match if you are over 180 than you would be if you make between 140 and 160. That is a huge change in fortunes. (No team made a score less than 100 batting first, though Rajasthan came close. It would probably be 0% win percentage)

I also took a look at the winning and losing scores, using the “Median” to avoid the effects of outliers and here is what we get (using rounded values).

Winning Score: The median winning score is 164/6. The team would win by 19 runs. That implies that teams chasing the target usually end up around 145, but in practice, the median chase only ends up at the 135-run mark.

Losing Score: The median losing score is 145/7. The team typically loses by 6 wickets, but this usually happens with only 5 balls to spare. So there have been quite a lot of last over finishes.

There is a slight difference in the losing target and the losing chase (145 vs. 135). This suggests that faced with a tough target, some teams may have already given up the chase, thus decelerating, or perhaps, more likely, they were forced to by the loss of more wickets.

In conclusion, it looks like a par score is between 145 and 164. Splitting the difference, we get 154. Which is almost exactly the score made by Delhi against Deccan today. But with Gilchrist in the murderous mood that he was in, they probably needed a lot more than that.

A Cricketing View Joins Rediff and Cricinfo in Statistics Hell

Posted on by

bubble

Kartikeya Date at A Cricketing View writes a good cricket blog, but he seems to have caught the statistics virus that is going around. We previously blogged about how first Rediff and then Cricinfo seem to be carelessly using statistics to prove whether one team or player is fractionally better than the others.

Karthikeya’s methodology is to consider the most recent home and away series for each of the 8 sides against the others, use a home rating and an away rating and then award a win bonus for a side which plays in another side’s country. Sounds familiar? Exactly! It is the poor man’s version of the Rediff ratings.

Unfortunately, Karthikeya’s methodology makes it even worse. For instance, he only seems to be taking Test victories and losses into account. This creates the problem that some teams have played a lot more Tests than others. For example, England has played 51 Tests against 32 Tests by New Zealand.

Why is this a problem? In statistics, it has to do with the law of large numbers. When you have short series (the 1-off and 2-Test series) played by New Zealand, there is a greater chance that the wins or losses are uncharacteristic of their actual talent. In a longer series, the better team generally prevails and the margins are validated. This is one of the reasons why it is easier to beat a team in Twenty20 than in Tests, because in the longer format, the minor battles play less of a part in the overall game.

Most of the other problems we discussed about Rediff’s ratings crop up here again. The ratings do not consider the following important factors:

  • Margin of victories in the Tests or clinging to a draw.
  • Dead rubber victories/losses.
  • The fact that team composition changes over time.
  • The rain factor.
  • Who won the toss on a dicey pitch.
  • Umpiring errors and sheer luck. (Sydney, anyone?)

The other BIG problem with the home-away ratings is that it is going to be obsolete for the next several years. The obvious case is Pakistan where no one is going to play for the foreseeable future. Sri Lanka and Bangladesh are not far behind. With the move of the IPL to South Africa, India could also end up in the same boat. And of course, India and Pakistan will not be playing against one another soon.

There is room for a qualitative analysis of the series wins and victories. Quantitatively, it makes no sense, whatsoever. Giving a single number to represent a team’s strength based on just home and series wins is just incredibly wrong. What say you, Karthikeya?

[Photo licensed from kubina]

Will Ponting Overtake Tendulkar?

Posted on by

Teams fight against each other to win Test and ODI series, but there are other competitions going on, especially in the individual record space. One of these is who will remain on the Mount Everest of cricket, i.e., who will score the most runs and most 100s. Currently, Tendulkar leads both charts with almost 12.5 K runs and 41 centuries. Here is a table of his nearest active competition and how younger they are when compared to Tendulkar.

Current Player Tests Runs Avg 100s Less Tendulkar
Tendulkar 157 12499 54.58 41 -
Ponting 131 10948 56.43 37 1y 8m
Dravid 132 10575 52.35 26 4m OLDER
Kallis 131 10175 54.41 30 2y 6m
Chanderpaul 119 8502 50.01 21 1y 4m
Jayawardene 102 8251 53.23 25 4 yrs

Ponting, Dravid and Kallis are within striking distance of Tendulkar (surprisingly, they have played almost the exact number of Tests). We can eliminate Dravid as he is 4 months older than Tendulkar and presumably will retire earlier. He will also play the same number of matches as Tendulkar, unless Tendulkar has a sustained poor run of form and is dropped while Dravid keeps playing. At this moment, that does not seem very likely. Apart from form, the Board will have to deal with angry public opinion if Tendulkar is ever dropped. Not gonna happen!

Ponting and Kallis have roughly the same shot at overtaking Tendulkar. If  they all retire at the same age as Tendulkar retires, they will have 1.6 and 2.5 years respectively to overtake his total which would be around 2000 runs more. In fact, they may have more time because if they are really close to Tendulkar, they may stick around for a year or so longer. If they maintain their current average, they would need roughly 20-25 Tests to overtake Tendulkar. Last year, the big teams played 12-15 Tests. So it is definitely achievable.

So the main questions are: Will Ponting and Kallis maintain their batting towards the end of their career, or will they have a sudden end like Hayden? Both of them have fallen from their previous near-60 averages and are nowhere near hitting the 1000 runs per year they used to. Ponting has more room here, because a winning Australia will help him retain his place on the basis of captaincy. With the victory in SA, he is back at the top again. Kallis has only his batting form to fall back upon.

The success of Twenty20 may disrupt some of these calculations. More international T20 and ODI matches mean less time for Tests. Nowadays, Test series are down from 5 and 6 Test series to 2 and 3 Test series. IPL and other local T20 tournaments can reduce the incentive of players to continue playing Tests. Injuries are more problematic at older ages. Here, Ponting is worse off. Being the captain, he is forced to be active in all the three forms of the game, whereas Kallis could end his ODI commitment earlier. Tendulkar has already quit T20 and only plays some of India’s ODI matches – he may be saving himself for Tests and one last World Cup.

That being said, last year proves that teams can play several Tests in one year. Australia and South Africa are popular teams and, like the popular kids in school, get to play with every one. Both Ponting and Kallis are exceptionally talented and can raise their game, and their cricket boards will accommodate their desire to hit a target more than other players. In any case, Ponting has a real chance of catching up to Tendulkar’s century count. He is just 4 behind. Unless Tendulkar has a spectacular year, that record is Ponting’s for the taking.

Chanderpaul, too, can be eliminated. He is too far behind, even if he has been racking up huge scores and crossed the 50-barrier recently. It is surprising that he has even scored so many runs after playing so low down the order. Jayawardene is another story. 4 years to hit 4000 runs. Possible, especially with Sri Lanka’s schedule leading them to more matches with the low-ranked Test teams. Whether Jayawardene’s appetite will be enough is a bigger question. Resigning the captaincy after the India defeat showed that he does not have the stomach for handling adversity.

The New Australian Team

Posted on by

Australia have made a confident start to their effort to regain the No.1 Test ranking from South Africa by first posting an above-average Wanderers score of 466 and then reducing the South Africans to 85 for 3. The Test could still go the other way, but there is no disputing Australia’s dominance at this moment. But what may not be so evident from the scorecard is that the Australians are a new-look team with 3 players making their debut and two others with experience of 5 Tests or less.

Take a look at the following chart of the last 3 Tests (including the current one) played by Australia. I have ignored the first Test of the previous series against South Africa as the team was essentially unchanged.

                                  Total Tests   Avg. Tests/Player
2nd Test vs. SA (in Australia)           464               42.18
3rd Test vs. SA (in Australia)           373               33.91 (2 debutants)
1st Test vs. SA (in SA)                  276               25.09 (3 debutants)

In comparison, the South African team has a total experience of 630 Tests in the current Test, almost twice the experience, and only Duminy has played less than 15 Tests. Of course, that would not count for much, if the new Australian players are very talented. Let us take a look at the five players who may either be the future of Australian cricket or just be having their 15 minutes of fame.

Marcus North: With a century on debut, the 18th Australian to do so, North’s stock is pretty high. He has played for various teams, including Australia A. North has a good record in first-class matches with almost 9000 runs and 22 centuries.

Philip Hughes: Heaven for some, hell for others. Hughes made a golden duck in his first Test innings, but this 20-year old is extremely talented, being the youngest man to score a Pura Cup final century.

Benjamin Hilfenhaus: Another golden duck on debut, he denied Johnson his maiden century. He must have been relieved when he scalped Amla off his 2nd ball in Test cricket. He has been around in international ODIs and T20′s since 2007.

Andrew McDonald: An all-rounder who made his debut in Australia’s previous Test, he has to improve upon his duck and so-far wicketless performance in this Test. His first Test was not great either. Most likely to dropped at this moment in time.

Peter Siddle: Compared to the rest, he is a veteran. Took 13 wickets in the previous series against SA and helped win the Sydney Test. Young, hasn’t played much in first-class cricket, but has shown enough to be playing Tests.

Triple Century-Makers in Cricket Tests

Posted on by

Younis Khan, playing the captain’s knock of a lifetime, scored a triple century to help Pakistan avoid a follow-on against Sri Lanka and overhaul the enormous Lankan total of 644/7. While the pitch is a graveyard for bowlers, fast and spin, that should not detract from the enormous achievement of batting for over 2 days against a good opponent. There have been flat tracks before, there will be flat tracks in the future, but very few people have managed to get a three in the hundreds column. In over 1900 Tests, there have only been 23 triple centuries.

Here are the century makers (the bolded scores were the record at the time)

  1. Andy Sandham, 325, England v WI, Kingston, 1930
  2. Don Bradman:
    1. 334, Australia v England, Leeds, 1930
    2. 304, Australia v England, Leeds, 1934
  3. Wally Hammond, 336*, England v NZ, 1933
  4. Len Hutton, 364, England v Australia, The Oval, 1938
  5. Hanif Mohammad, 337, Pakistan v WI, Bridgetown, 1958
  6. Garfield Sobers, 365*, WI v Pakistan, Kingston, 1958
  7. Bob Simpson, 311, Australia v England, Manchester, 1964
  8. John Edrich, 310*, England v NZ, Leeds, 1965
  9. Bob Cowper, 307, Australia v England, Melbourne, 1966
  10. Lawrence Rowe, 302, WI v England, Bridgetown, 1974
  11. Graham Gooch, 333, England v India, Lords, 1990
  12. Brian Lara
    1. 375, WI v England, St. John’s, 1994
    2. 400*, WI v England, St. John’s, 2004
  13. Sanath Jayasuriya, 340, Sri Lanka v India, 1997
  14. Mark Taylor, 334*, Australia v Pakistan, Peshawar, 1998
  15. Inzamam-ul-Haq, 329, Pakistan v NZ, Lahore, 2002
  16. Matthew Hayden, 380, Australia v Zimbabwe, Perth, 2003
  17. Virender Sehwag
    1. 309, India v Pakistan, Multan, 2004
    2. 319, India v South Africa, Chennai, 2008
  18. Chris Gayle, 317, WI v South Africa, St. John’s, 2005
  19. Mahela Jayawardene, 374, Sri Lanka v South Africa, Colombo, 374
  20. Younis Khan, 313, Pakistan v Sri Lanka, Karachi, 2009

Only three batsmen have hit two triple centuries: Don Bradman, Brian Lara and Virender Sehwag. Other than Lara, no one with over 10,000 runs has hit a triple century. At the other end of the spectrum, Andy Sandham scored just 875 runs in his entire career.

England leads the way with 5 triple century-makers, followed by Australia and West Indies with four each, Pakistan with three, Sri Lanka with two and India with Sehwag. New Zealand is unlucky to miss out with Crowe getting out on 299. South Africa are also yet to open their account.

St. John’s and Leeds have the most triple centuries (3 each). England has conceded the most triple centuries (7) while Australia, perhaps not surprisingly, only just one. Except for Hanif Mohammad’s rearguard match-saving effort against the West Indies, all the triple centuries were made in the first innings of the respective teams.

Graham Gooch is the only person to hit both a triple century and a ton in the same match. Virender Sehwag hit the fastest 300 off just 278 balls. He may still do it again. Wonder who the next victim will be.

England Team Keeps Fans Confused

Posted on by

antigua

England have bounced back from their spectacular low of 51 all out in the first Test against the West Indies with a monumental total of 566. With almost all the England players cashing in with 2 centuries and 3 fifties, this was a team effort and everyone deserves credit from being able to exorcise the ghosts of the Sabina Park debacle.

When a team makes a score ten times higher than their previous outing, we have to look closely at their patterns of performance. So, we took a look at the recent performances of England and compared it with the West Indies. To no one’s surprise, England is much stronger on its batting performances than the West Indies. The English team has an average of 36 runs per wicket in its last 13 Test matches (including the Jamaica Test), while the West Indies have an average of 31.68 per wicket. That is almost 43 runs per innings.

However, England’s scores have a standard deviation of 18 runs per innings, which means that they are as likely to score 180 as they are to score 540. WI have a standard deviation of 11 runs, which means their capability is more between 200 and 420. Obviously, teams change in composition (with players retiring or dropped) and they play against different countries, but it is still very striking. The West Indians, while performing poorly, are performing consistently. England, on the other hand, can put up huge scores, but are also prone to collective failures.

Even without the help of statistics, we see England running the gamut from an imperious 593/8 declared against an opponent no less than the South Africans, but also slumping to 110 all out against the Kiwis. The lowest score for West Indies was 139 and their highest score was 408, significantly lower than what England could achieve as a team. In fact, if they overhaul England’s first innings in the current Test, that would be a statistical outlier, almost 50% more than their high score in more than a year.

England should not have to bat again in this match, but the West Indians are pretty good at sticking around. And if the weather plays a part, they may still escape unscathed.

[Photo licensed from z_dead]