Cricinfo.com – plenty of cric, not so much info

The international cricket season isn’t far away, so doubtless we can soon expect a steady diet of stories like these, carving up a whole pile of numbers and telling us that the reason a cricketer / cricket team is succeeding / failing is… well, just about anything.

S Rajesh of cricinfo, though, is obsessed with the idea that performance is affected by the location of a match – whether it’s the ground, the country, one’s home country, one’s home country (again), or even the continent. (Indicative quote: “Clearly, Hafeez’s problem has been facing the new ball outside Asia.” Ummm… yeah. Clearly.)

His theories may be right. But there is so much wrong with the way he tests them that I’m not really sure where to begin. The formula seems to be something like this:

Pose a question, attributing something to a single variable: could Sri Lanka’s recent success be linked to the colour of their shoelaces?

Throw in some commentary, knocking down counterpoints as we go: Sri Lanka have won 6 of their last 7 matches whilst wearing egg-shell white laces, their only loss coming by a single run against a rampant Indian team, which – it must be noted – also had a number of players in egg-shell white laces.

Time for some tables, completely ignoring sample size, strength of opponent (apart from occasionally deleting all matches involving Zimbabwe or Bangladesh – but not 1980s Australia or 1990s England?), comparison to existing player averages and every other available variable: average runs scored in egg-shell, orange, green and pink laces… the egg-shell stats are slightly better!

Conclusion: this has been fascinating. Egg-shell laces – could there be something in it? Only time will tell.

And yet, Rajesh keeps on pumping out the articles, with scarcely a dissenting voice to be found – because, no matter which variable you choose to look at and (try to) measure, there’s always going to be a player or team perched impressively on top of the pile, ready to be lauded or critiqued in an article as per the formula.

To look at just one actual example, Rajesh, ahem, “examines” the idea of whether batting in one day internationals is as difficult in England as it’s reputed to be. He rolls out the data tables and, sure enough, the run-rate per over for matches played in England is up near the top.

This would work well if all the matches being referred to featured the same teams, or at least the same distribution of teams in similar combinations. But am I going out on too much of a limb to say that most of the matches played in England will have featured… you know… England?

Likewise West Indies in the West Indies. South Africa in South Africa. You get the drift.

Now, I have no idea whether England’s featuring role in these matches (as both batsmen and bowlers) will push the run rate up or down – but it’ll certainly do something (and Rajesh, that something will skew your numbers right off, mate).

One way to overcome that would be to look at all of the individual players and measure how quickly they scored in England, relative to their run-rate in other matches; and then, to robustify the whole thing, calculate confidence intervals to find out if your results really truly do mean something. This will all be very time-consuming and complicated, but might yield some interesting stuff. Sounds like a task for someone with vast resources and data available – a bit like cricinfo, perhaps?

Ok. One last thing to finish off:

Imitation is flattery.* Rajesh, I salute you with the offering below, researched and delivered in your inimitable style:

Some players excel on their home ground or in their home country, some in Asia or elsewhere, some at certain times of the year or at certain periods in their lives. But when Greg Chappell played test cricket on grounds whose name started with the letter B – whether it was at the Brisbane Cricket Ground, Bridgetown or Birmingham – he just freakin’ rocked the joint.

top 30 players

Top 30 player averages at test venues whose names begin with the letter B. (Minimum runs: 500)

You can say what you like about Michael Clarke, but one thing can’t be argued: as has been clear in his performances from Bangalore to… um… Birmingham (again), he is better than previous captain Ricky Ponting when it comes to scoring test runs at grounds starting with the letter B.

Thanks for the inspiration, Mr Rajesh. Case closed.

*Actually it’s not in this instance. But anyway.
Advertisements

Click here to like, or tell me whether this is genius or gibberish

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s