Saturday, September 5, 2015

Solving the Khan-undrum

The three Khans of Bollywood - each a superstar in his own right. Each having dominated the industry for so long that despite the rise of, arguably, better acting Khans like Irrfan and the long standing Saifs of the world, the trio doesn't need their first names for identification.

Battle lines have been drawn among the three on numerous occasions with a each of the Khans drawing a rather equal share of accolades and his own fair quota of brickbats. Yet, the question of who's the best of the three Khans has always got both Bollywood pundits and the audiences scratching their heads or each other's faces.

In such cases, I believe that the word of ten people holds more weight than that of one, and that of a hundred even more. So, indeed the opinion of thousands should hold enough weight to lead to some conclusion. Where will I find this, I thought. IMDB. The folks at IMDB have been kind enough to put up an updated copy of their database here - which in its excel form is extremely tough, yet not impossible, to sift through.

After a painstaking few days poring over excel sheets, I found what follows in this post. A lot of it is hardly surprising and merely proves what most of us have already believed. But quoting our flamboyant turbanator Mr. Sidhu (who was probably quoting someone else), "I use statistics like a drunk man uses a lamp-post; for support, not for illumination".

Here's a quick glance of the line-up:

I think one of the few things entirely dissimilar among the Khans is the number of marriages they have been part of

It is tough to find a set at such similar stages of life and career. Let's dive deeper with some very basic metrics.

The ratings table (cleaned up voice-overs & appearances as child-actor)
Salman & SRK have about double the number of movie as compared to Aamir. No surprises there - Mr. Perfectionist has always been more frugal with his role selection. In terms of the number of votes cast, Salman comes as a bit of a surprise (Note: That says quite a bit about the Salman fan base; it is not composed of a largely IMDB visiting populace). On this count, Aamir pulls away with some dazzling numbers: Just the votes on 3 Idiots, Taare Zameen Par, PK and Lagaan are equal to the total votes cast for Salman's 96 movies.

Sometimes less is, indeed, more
Aamir wins hands down on quality of votes as well with the average rating on his movies almost an entire point above that of SRK. The chart on the left should help by drilling down further into their montage of films.

Check out how SRK has done an equal or, perhaps, higher number of well-rated (7 & above) movies than Aamir, However, doing a lot of sub-par (< 5.0) movies hurt him. Salman, of course, went full throttle in that bracket and undid any good work he had managed to dole out. Interestingly, Aamir & Salman's highest rated movie is the same: Andaz Apna Apna (1994) with a rating of 8.8. Salman had to wait over 20 years to deliver another 8-above with Bajrangi Bhaijaan (2015) while Aamir dished out 7 others in the same period to take his total of 8-above's to 9 (Jo Jeeta Wahi Sikander in 1992 was another one)

SRK also has attained the 8-above marker an impressive 8 times. However, our analysis has yet not laid stress on recency: an important factor in an actor's current value. All 8 of SRK's 8-above's came prior to the current decade with 4 of them in the last millenium. Aamir, keeping in mind the niggardly fashion in which he went about his movies, has done better to have all of these but 3 in the 2000s. This cumulative rating average over the years chart should help you picture this fact better.

Aamir's ratings curve is a corporate presenter's dream - steady rise over the years
Cumulative rating over the years shows Aamir's steady rise
But it might be too soon to discount the Bhai fanaticism...
One glance at the ratings table in the beginning of the post tells you Salman garnered the lowest average rating. Yet, the moment you start weighing the average by votes everyone jumps up a rating point. The percentage jump, however, is the highest for Sallu bhai - an indicator that his fans are far more likely to vote when they like a movie than otherwise. This could be called higher fan intensity.

Who is the most versatile (or experimental) of them?
Who would've thought Salman would be the most diverse?
Apart from just the objective rating scores, I felt it was important to see the diversity of movie genres each of the Khans has featured in. Each movie on IMDB is likely to have more than one genre mapped to it. Since, it was nearly impossible to assign priority weights to each genre based on prominence in the movie, I have split weights equally (e.g. if a movie is listed as Comedy, Romance & Drama, that gets counted as 1/3rd of each genre).

Comedy and Romance find a larger place in Salman's portfolio of films than any of the other Khans. Remarkably, the King of romance SRK is lagging both the Khans in % of movies under the 'Romantic' genre. He, however, as expected, takes the cake as far as 'Drama' goes.

Lastly, to get who was the most diverse, I relied on my favourite statistical measure - the Herfindahl-Hirschman Index - taking the percentages in each genre for each individual only to find that Sallu bhai emerges as a winner in this round.

I also chanced upon some exciting tid-bits, a few of which are listed on the left. For the record, Salman has played 'Prem' 12 times so 'Prem - naam to suna hoga' would have been more apt. Also, the only tele-drama all three Khans have made guest appearances in is.. wait for it... Diya aur Baati hum.

I would, of course, not venture a claim to the 'Best' Khan for two reasons: 1) 'Best' is highly subjective and, 2) I have no intent of being lynched by the mob whose favourite I do not consider or get run over by a car.

Nevertheless, this analysis did prove quite a few things:
1) the money a movie/ star makes has absolutely no correlation to its rating,
2) being picky helps in improving quality of work, and
3) you can get data to talk in any way you please!