Tuesday, May 10, 2016

Visualising Kobe Bryant's Scintillating Career - Well, almost!

Kobe Bean Bryant, one of the most accomplished & decorated NBA stars, ended his 20 seasons spanning career this year and went out on a high scoring a startling '60 or above' (10 more than his good ol' mate Shaq asked him) in his last professional outing in Los Angeles Lakers' clothes.

His speed & accuracy over the course of his career synonymous of those of a brand of snake earned him the nickname Black Mamba.

I tried to dissect his career in terms of shots made & missed looking for patterns based on the dataset Kaggle made public which has all information on ~84% of shots made in his entire career . There are no huge revelations or weaknesses, indicating the sort of legend he is, but some of the highlights worth mentioning are summarised below:

1. Dunks are rarely missed & Kobe favours one side of the board to the other

The visual below shows the layout of the court and the dots indicate the position from which he has attempted his shots. The image on the left has all shots he scored off & the one on the right comprises of shots he missed.

  • A huge % of shots he has attempted have been Jump shots, which is a no-brainer. Looking beyond those jump shots (greyed out above), one can notice that the spread of shots is higher to the reader's left of the board in 'Shots Hit' visual and the spread in 'Shots Missed' visual is predominantly to the right. Possibly a little weakness that teams playing Lakers wouldn't have cared to miss?!
  • He rarely misses a dunk (marked in 'Brown') - There are literally no brown dots at all in the Shots Missed visual.
  • When he attempts a Bank shot, it is expected to go in at a higher rate (notice the many red dots in 'Hit' & a fewer % of them in 'Missed')
2.  Layups & turnaround shots may look great for a fan but they are costly for the players

Success rate of his shots (% shots scored out of all attempted) by shot type seems to go in hand with the observations above:
Slam & Driving dunks are at close to 100% success rate; Some of the risky ones like the Layup & a few Turnaround shots understandably have ~50% success rate.

3.  Very very small zones Kobe isn't comfortable shooting from

After spending some time on shot types, shot distances & shot positions, only one visual had anything to offer in terms of a possible highlight - The one corresponding to distance from which Kobe has attempted the shot.
The areas marked in circles above seem to be points from which no shot has gone in & these are not areas where he has not attempted shots from (these gaps are not seemingly large in the missed visual).  I am no seasoned basketball fan to find a reason to that but probably those are strong defence areas that anyone finds difficult to score from. If not, voila! I just noticed a little Achilles heel in Mamba's illustrious career.

4. Clutch is the word for Kobe

If scoring 60 on his 'announced' retirement game was not enough proof of his ability to clutch on to his nerves in moments that matter, the consistency of success rate he shows across seconds remaining (in the quarter I presumed as it had 0 to 60 as against 0 to 24 if it had been seconds remaining for a shot) & minutes remaining (12 minute sessions are what NBA's stipulated quarters are made of).

The deviations are pretty low if you look at Seconds remaining for Quarter end (understandably reducing hugely for 1 second but it's pretty good at even 2-5!!)
 There is a dubious low seen at '8 minutes' remaining which can't be ignored as an anomaly for the sheer number of data points available (~26,000 shots attempted). I couldn't find what could have caused that. Any possible reasons for that dip?!
5. Nobody can defy age, not even Kobe

Here is a look at his success rate by seasons & by months within seasons.
He has been consistently successful picking up a little after 2 seasons from debut and staying there for 4 seasons to follow. A seeming little dip later, he has made the comeback like legends always try to and has faded gradually in the last leg (indicated by the dipping tail towards the end)

More fascinating was how he starts off seasons (with the success rates gradually climbing across the first 4-5 months - Oct to Feb before gradually settling down). His success rates in the PlayOffs months is a little fascinating. It peaks in May & drops to the lowest in June. Is that indicative of the quality of opponents in the last leg of playoffs - the final? Possibly. My restricted NBA knowledge can get me only so far.

This was all I could muster in my quest for predicting Hit or Miss for the remaining 5000 shots Kaggle's competition dataset had. All through this work, the thought that kept knocking my brain cells was - How nice would it be to make a tribute to some legend in a sport I am a lot more closer with - Tennis! Where my analysis will be a lot more sound given the subject matter expertise, if I could call it that :) There is hope in the form of tennisabstract.com's dataset. As long as there is hope, all is well.