Showing posts with label Sachin Tendulkar. Show all posts
Showing posts with label Sachin Tendulkar. Show all posts

Friday, March 9, 2012

My tribute to Rahul Dravid: What if he debuted with Sachin

Rahul Dravid decided to call it a day and thus comes the end of an illustrious career. A lot has been written and more will be written in days to come, but those who have witnessed Rahul's batting know that no amount of runs can do justice to the man and his career. I often take the refuge of 'numbers' when it comes to define a career because memory gets clouded and lot of prejudices enter in our evaluation. 

This is specially true in the case of Rahul who all through his career remained in the shadow of Sachin Tendulkar. They say that he did not mind that, but what choice did he have when the whole country decided to remain mesmerized with Sachin (for few reasons of course).

Dravid's first test match somehow defined his career. A glorious knock of 95 at the Lords on his debut, but Ganguly another debutant scored a century and took all the limelight. This has been the fate of Rahul all through career. Judgement of Indian cricket masses is strongly clouded by their emotions and not one really pays attention to numbers. For some reason we treat Zaheer Khan a great bowler (in Indian context of course) but we forget to compare his record with Javagal Srinath, who is not considered a great bowler.  

So I tried to look through the number again compared Rahul Dravid with our benchmark of greatness Sachin Tendulkar when they played together. I hate to bring this comparison by just numbers so I made some graphics. I asked a simple question: when Sachin and Rahul batted together in an innings in a test match who scored more runs. I do not want to calculate averages, that tells nothing, I will only give you the raw data in a pictorial manner so you can judge with less emotions and more objectivity. Here is how to read the figure
Sachin Tendulkar vs Rahul Dravid.
Rahul is marginally better if at all when they play together.


Panel A: Innings by innings runs scored by Sachin Tendulkar. The dark line is smoothed version of the same data. 
Panel B: Innings by innings runs scored by Rahul Dravid. The dark line is smoothed version of the same data. 
Panel C: Probability distribution of runs scored by Sachin (brown line) and Rahul (blue line). This graph shows the probability of scoring some number  of runs in an innings. It turns out that there is no difference between the two when they batted in the same innings. Moreover, it is all noise, as the distribution is an exponential (thick gray line), characteristic of a class of stochastic processes (Poisson Process).  
Panel D: Scatter plot of runs scored by Sachin (Y-axis) and Rahul (X-axis) when they played together. If the point is above the gray line then Sachin scored more runs, if it is below the gray line Rahul scored more runs. It turns out that Rahul outscored Sachin 130 times while Sachin outscored Rahul 118 times. But there is not much difference in the runs they scored:  Rahul (12981) scored 395 runs more than Sachin  (12586). Ideally we would want that both score together but that was so rare.
Panel E: The scatter data in panel D is presented as probability in pseudocolors. 
Panel F: Innings by inning difference in the score of Rahul and Sachin. Positive number means Rahul scored more and negative number means Sachin scored more. 
Panel G: The distribution of run difference shown in panel E is a gaussian noise. 
Panel H: Autocorrelation of run scoring by Sachin (brown line) and Rahul (blue line). Once again, as revealed by panel C, the run scoring by the two players is a stochastic process. Although for Rahul there is some increased chance of repeating his performance (failure or success) every 5 matches (I am not sure if it is significant). 

What does this analysis tell us? Because I considered the matches when they played together, I automatically corrected by all the conditions, bowlers (they bat at no. 3 and 4) and other effects. SO when we look this analysis I do not see any difference between Sachin and Rahul. If I have to chose perhaps I will go for Rahul given that he scored 395 more runs. But I may not play them together because chances that they both will score good amount of runs is rather rare.

This great similarity between the two giants of Indian cricket and or World Cricket in 2000s makes me ask why Rahul remained in Sachin's shadow? What if Rahuls had debuted slightly before Sachin Tendulkar, would be worshipping Rahul and not Sachin? 

Perhaps Sachin is so popular for his early start and his explosive batting in limited over cricket, something Rahul was never allowed play like Sachin. I will do that analysis some other day. But here is a humble request, please, respect Rahul for what he has achieved, his numbers tell a very good story and they very conclusively say if at all Rahul Dravid is better than Sachin Tendulkar. Rahul is a humble man when he says that the next generation is well equipped, we all know the vacuum he is leaving at the number 3.

right arm over
Arvind

Saturday, August 13, 2011

The great indian batting line up of 2000s

India boast about their great batting line which is studded with Virender Sehwag (Runs: 7694, Avg. 52.69, 100s: 22, 50s: 27), Rahul Dravid (Runs: 12616, Avg. 52.56, 100s: 34, 50s: 60), Sachin Tendulkar (Runs 14851, Avg. 56.25, 100s 51, 50s 60) , Sourav Ganguly (Runs: 7212, Avg. 42.17, 100s: 16, 50s: 35), VVS Laxman (Runs: 8302, Avg.: 46.64, 100s: 16, 50s: 54).


In the 2000s India routinely fielded at least three of these five players in their batting line up and on occasions they all played. So I looked into the data when at least 3, 4 or all five played in a match and checked how much they each scored. Specifically I was interested in knowing when they all got to bat (three, four or five of them), did they all manage to score at least 50. The only way a team can benefit from its stars is if they all bat well together, otherwise individual stars are not enough to brighten the prospects of a team (of course Lara can do but even he managed to win only few Test matches on his own).

In the three figures below I have separated the data into first and second innings of Indian batting. This does not refer to the innings in the match. The gray boxes means that the corresponding player played in the match but did not get to bat in that innings. Light blue box means that the corresponding player made less than 50 runs and dark blue color means that the player made more than 50 runs. 

For collaborative batting efforts we are interested in knowing how often and how regularly at least three of then scored 50+ score in the same inning. The number are shown on the right of each subplot.

When at least three of them played together only in 11.43% cases all three managed to score 50+ runs, which is clearly a small fraction. Even more surprising is the fact that only 25% times two of the three made 50+ scores. These number drop in the second inning when only in 6% cases all three made 50+ scores.

Performance of five most successful modern Indian batsmen when at least three of them played together.

When at least four of them played together, in the first innings they never made 50+ score together (figure below). In 31% cases two of them managed to get 50+ in the first inning. The second innings remained a solo performance case when in 41% cases at least one scored a 50+.

Performance of five most successful modern Indian batsmen when at least four of them played together.
When all the five played together, media boasted about the strength of the Indian batting but never did all five made 50+ together. Playing all of them at the same time meant that at least two of them scored 50+ in about 40% cases in the first innings. Second innings again showcased individual performances.

Performance of five most successful modern Indian batsmen when all five played together.

This analysis shows that even though India has produced some very good batsmen in last two decades or so, but then have only rarely clicked together. I am not asking all of them to score runs in the same match but it is expected that at least three of the five score in the same inning, and unfortunately this has happened only 3% times in the first inning and 12% times in the second innings... but certainly only often enough that India could really exploit the batting abilities of these genius who as of now seem to like to work alone....to increase their individual scores and records...

right arm over
Arvind

Wednesday, May 4, 2011

Greatness of Sachin Tendulkar lies in his longevity

What makes Sachin Tendulkar a truly great player? There is no objective answer to this question. Every cricketer and cricket fan will have their own answers and augments.  In this age of data we have a better chance to identify components of Sachin's greatness. Below I provide one such analysis.

I have plotted few descriptors of batting for 50 top run getter in both Test matches and ODIs. The seven descriptors I am using are
Not Out [panel A]
Cumulative Runs [panel B]
Highest Score [panel C]
Batting Average [panel D]
Centuries [panel E]
Fifties [panel F]
Zeros [panel G]

Specifically I am interested in knowing if there is any correlation between any of the above descriptors and number of matches played. It is kind of obvious that if you are a batsman then more matches you play more runs and centuries you will get. 

In the two figures below the top one belongs to ODI matches and the lower one belongs to Test matches. Each player is marked with a colored circle. The color for each player is mentioned on the far right.

Indeed cumulative Runs scored has a correlation of 0.89 (ODI) and 0.84 (Test) with number of matches played [panels B in the figures below].
Same holds for number of 50s with a correlation of 0.74 (ODI) 0.89 (Tests) [panels F in the figures below].

In test matches even number of 100s scored is highly correlated with number of matches (correlation 0.84).  This is not so true for ODI (correlation 0.57). This analysis shows that Sachin's centuries in test matches are more or less predicted by his longevity. However, his count of ODI centuries is indeed a signature of his class and greatness.

The analysis also shows that in ODIs there is a weak but significant correlation between longevity and batting average. Similar trend is seen in Test matches but that more because of Sir Don Bradman.

Other descriptors are not correlated with longevity.

In both the pictures Sachin's stats are marked with a green circle and a small blue circle. For a comparison in test matches I have marked Don Bradman with a blue circle.

In summary, Sachin's runs and centuries and fifities in Test matches and runs and fifties in the ODIs are just a simple case of statistics. Any of the player shown in the figure would have matched that feat. Sachin's true brilliance shows up in the fact that he has scored so many centuries in ODIs.

Overall, this data suggests that largely Sachin's greatness is nothing but his longevity.




right arm over
Arvind