An invisible, powerful force is lifting
professional basketball to new heights.
Transforming how this multi-billion dollar
sport is played.
In elite sport, the difference
between success and failure
is often the finest of margins.
Go Rockets, woo!
The Houston Rockets are
one of the top teams
in NBA basketball.
They boast some of the
sport's biggest stars,
including the NBA's Most
Valuable Player in 2018.
Offensively what we were trying to do,
defensively what we were trying to do,
so the focus is trying
to be on the same page
every position for every game.
In the past decade,
the Rockets have risen
from mid-table mediocrity
to serious NBA Championship contenders.
It's not just big names that have
fueled this dramatic
ascent, it's big data.
The data that we go over
is definitely important,
you know, I think it gives
us a little bit of an edge.
To keep a balance of data
and instinct is a big thing.
The Rockets' recent success
owes much to their pioneering decision
to start crunching data about
every aspect of their game.
And this is the man responsible.
Analytics has really permeated everything,
our edge has allowed us
to win the most games,
65 last year.
Computer scientist Daryl Morey
is the sport's foremost
data and statistics guru
among NBA bosses.
You can break winning
down into two things,
one is how many points do
you get per possession,
and then how do you get extra possessions?
Ten years ago, Daryl set out
to find that winning formula.
The Rockets were one of four NBA teams
to install a pioneering
video tracking system,
which mined raw data from games.
What they discovered changed
the way teams tried to win.
The data revealed which shots provided
the best bang for buck, two-point dunks,
and three-pointers, shot from
outside the three point line,
rather than long-range two
point shots from inside it.
It's pretty dramatic,
how powerful the three point shot is.
You only have to make a third
of your three point shots
to be worth a half of your two.
Even within that, there's an extra miss
you might rebound offensively.
Face up, pulls the trigger, buries it!
In the 1990s, long two point shots
from just inside the three
point line were common.
But Daryl's analysis
showed that statistically,
these shots provided the worst return.
The number of attempted three
point shots has increased
every year for the past decade.
In the 2017-18 season,
the Rockets made more
three pointers than any
other team in NBA history.
And this was a major reason they won more
games than any of their rivals.
Every percentage matters in sports,
and extracting those percentages
is what's really become
much easier with data.
We're in a pretty big
transformational stage in sport.
Professor Rajiv Maheswaran
co-founded Second Spectrum.
The analytics company gathers and codes
a vast range of increasingly granular data
for all 30 NBA teams.
It feels like two chess pieces.
Cameras now track and record
3D spacial data for every
player and ball movement
at 25 frames per second.
Machine learning technology
uses this huge volume of data
to produce interactive visualizations,
allowing teams to analyze the minutiae
of their performances, and
achieve marginal gains on court.
At first we taught the
machine about 20 terms,
right now it knows over
500 basketball terms.
So the machine is essentially
the integrated knowledge
of all the elite basketball coaches.
There is a particular focus
on all important data
around player movement,
and the probability of making a shot.
What we can do with the machine is to say
look at all the shots like this one,
and tell me the chance
that the shot really could
go in either for this player
or for an average player.
And you bring in all the
things that a person can see,
not just where they're standing,
but how they're moving,
what type of shot they're trying to take,
where are the defenders?
And what we can do is put a number
to what previously was a feeling,
or something qualitative.
Data analysis has even changed
the type of players that successful teams,
like the Rockets, have.
Players today are on average
leaner and more agile.
Having skilled players,
even if they're a little smaller,
is more important than having bigger guys
who are less skilled.
When it comes to recruiting new players
from the college draft each season,
poring over data on player performance
has given the Rockets a winning edge.
We've been able to eke
out about a 5% edge,
and that actually turns out
to be a massive advantage.
Pro basketball's finale for the NBA title.
Basketball has constantly changed,
but it's about to enter a brave new world,
where data could be court-side
in the hands of coaches,
helping to swing a game as it happens.
The data's always getting
faster and faster,
the data used to take hours,
we've got it down to minutes,
and very soon it'll be down to seconds.
It's basically going to become
functionally real-time very, very soon.
I don't think we've scratched the surface
of what's possible with data.
