jump to navigation

In praise of statistics April 28, 2006

Posted by Sharath Rao in statistics.

My potential alternate profession.

Okay, its not quite likely I would do that – but it would rank high among many choices I have. It brings two things I love – patterns among numbers and the game of cricket.

There have been several funny quotes on statistics and statisticians. I dont quite take it when people say that statistics dont reveal anything or that they can be used to prove anything.

Firstly, they reveal a lot if you know how to measure them. In Cricket, well, averages say a lot – there is no single better measure of a batsmen’s ability. If you say that Player X is inconsistent, you can pull up the variance. If you say that statistics dont reveal the circumstances of the innings – well, you can meausure much of that too – runs on the board, status of the match, status of the series, wickets, past string of low scores, home ground, strength of bowling attack, experience of player – there are a whole lot of measurables involved. What is required is that you be creative enough to decide what these measurables are. The problem then is not in statistics but your ability to use the tools.

Same is the reverse problem. Given some statistics, drawing conclusions requires you to know what exactly these numbers are, what assumptions were made and how accurate they are ( possible error etc. ). If you are smart, you cant be fooled ! If you can ask the right questions, you cant be mislead.

So what eventually is required is that you understand the domain – the better you understand the problem, the better is your ability to find measurables and also find loopholes when presented with a table/chart.

I dont ofcourse say – numbers are all there is, instead I am saying that there is more to it than meet the eye. While Americans and much of western civilization is about experiments, trials, measurements, interpretations and documentations, its not so in India where we have no respect for numbers ( okay, we discoverd the zero, but I am talking about data/statistics, not numerals ) – we are more happy to go by gutfeels ( which in some sense is probabilistic ( bayesian ) ) – the two often differ at times. But thats for another post.

For now my only take is that people do sumnarily dismiss statistics as a conspiracy hatched by statisticians are like those who dont like a particular car because they dont know how to drive !



No comments yet — be the first.

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 )

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

%d bloggers like this: