The Math Behind Basketball's Wildest Moves | Rajiv Maheswaran | TED Talks

Prevodilac: Vesna Radovic Lektor: Mile Živković

My colleagues and I are fascinated by the science of moving dots.

Moje kolege i mene fascinira nauka o pokretnim tačkama.

So what are these dots?

Kakve su to tačke?

Well, it's all of us.

Pa, to smo svi mi.

And we're moving in our homes, in our offices, as we shop and travel

Krećemo se u našim domovima i kancelarijama

dok kupujemo i putujemo po gradovima i svetu.

throughout our cities and around the world.

And wouldn't it be great if we could understand all this movement?

Zar ne bi bilo sjajno da možemo da razumemo sve ove pokrete?

If we could find patterns and meaning and insight in it.

Da imamo uvid u njihove šeme i značenja.

And luckily for us, we live in a time

A srećom po nas, živimo u vremenu

where we're incredibly good at capturing information about ourselves.

u kom nam neverovatno dobro ide da pratimo informacije o sebi.

So whether it's through sensors or videos, or apps,

Bilo da je to putem senzora, snimaka ili aplikacija,

we can track our movement with incredibly fine detail.

možemo vrlo detaljno pratiti naše pokrete.

So it turns out one of the places where we have the best data about movement

Ispada da je jedno od mesta gde imamo najviše podataka o pokretu

is sports.

sport.

So whether it's basketball or baseball, or football or the other football,

Bilo da je to košarka, bejzbol, ili američki fudbal ili onaj drugi fudbal,

we're instrumenting our stadiums and our players to track their movements

postavljamo instrumente na stadione i igrače da bismo pratili njihovo kretanje

every fraction of a second.

u svakom deliću sekunde.

So what we're doing is turning our athletes into --

Ono što radimo je da pretvaramo naše sportiste u -

you probably guessed it --

verovatno već pretpostavljate -

moving dots.

pokretne tačke.

So we've got mountains of moving dots and like most raw data,

Tako imamo gomilu pokretnih tački, i slično većini neobrađenih podataka,

it's hard to deal with and not that interesting.

teško je raditi sa njima, a nije baš ni zanimljivo.

But there are things that, for example, basketball coaches want to know.

Ali postoje stvari koje, na primer, košarkaški treneri žele da znaju.

And the problem is they can't know them because they'd have to watch every second

A problem je što ne mogu da ih znaju zato što bi morali da gledaju svaki sekund

of every game, remember it and process it.

svake utakmice, sve to zapamte i obrade.

And a person can't do that,

A čovek to ne može,

but a machine can.

ali mašina može.

The problem is a machine can't see the game with the eye of a coach.

Problem je što mašina ne vidi utakmicu očima trenera.

At least they couldn't until now.

Bar do sada nije mogla.

So what have we taught the machine to see?

Šta smo to naučili mašinu da vidi?

So, we started simply.

Pa, počeli smo sa osnovama.

We taught it things like passes, shots and rebounds.

Naučili smo je stvarima kao što su pasovi, šutevi i skok pod košem.

Things that most casual fans would know.

Stvarima koje zna većina prosečnih obožavalaca.

And then we moved on to things slightly more complicated.

A potom smo prešli na malo komplikovanije stvari.

Events like post-ups, and pick-and-rolls, and isolations.

Događaje poput postapova, pik end rola i presinga.

And if you don't know them, that's okay. Most casual players probably do.

Ako ne znate te pojmove, u redu je. Većina prosečnih igrača zna.

Now, we've gotten to a point where today, the machine understands complex events

E, sad, došli smo do tačke da danas mašina razume kompleksne događaje

like down screens and wide pins.

poput akcija i vajd pinova.

Basically things only professionals know.

Praktično, ono što znaju samo profesionalci.

So we have taught a machine to see with the eyes of a coach.

Zapravo, naučili smo mašinu da vidi očima trenera.

So how have we been able to do this?

Kako smo ovo uspeli?

If I asked a coach to describe something like a pick-and-roll,

Ako pitam trenera da opiše nešto poput pik end rola,

they would give me a description,

on mi objasni šta je to,

i ako bih to ubacio u algoritam, izgledalo bi užasno.

and if I encoded that as an algorithm, it would be terrible.

The pick-and-roll happens to be this dance in basketball between four players,

Stvar je u tome što je pik end rol u košarci ples između četiri igrača,

two on offense and two on defense.

dva u napadu i dva u odbrani.

And here's kind of how it goes.

I evo kako to otprilike ide.

So there's the guy on offense without the ball

Jedan momak je u napadu bez lopte

the ball and he goes next to the guy guarding the guy with the ball,

i on ide do momka koji čuva drugog momka sa loptom

and he kind of stays there

i otprilike ostaje tamo,

obojica se pomeraju i nešto se dešava, i ta-da, to je pik end rol.

and they both move and stuff happens, and ta-da, it's a pick-and-roll.

(Laughter)

(Smeh)

So that is also an example of a terrible algorithm.

To je takođe, jedan primer užasnog algoritma.

So, if the player who's the interferer -- he's called the screener --

Tako, ako igrač koji vrši udvajanje - on se zove bloker -

goes close by, but he doesn't stop,

prilazi blizu, ali se ne zaustavlja,

it's probably not a pick-and-roll.

to verovatno nije pik end rol.

Or if he does stop, but he doesn't stop close enough,

Ili ako se zaustavi, ali se ne zaustavi dovoljno blizu,

it's probably not a pick-and-roll.

to verovatno opet nije pik end rol.

Or, if he does go close by and he does stop

Ili, ako priđe blizu i zaustavi se,

but they do it under the basket, it's probably not a pick-and-roll.

ali to uradi pod košem, to verovatno nije pik end rol.

Or I could be wrong, they could all be pick-and-rolls.

Ili ja možda grešim, možda je sve to pik end rol.

It really depends on the exact timing, the distances, the locations,

To zaista zavisi od preciznog tajminga, udaljenosti, lokacije,

and that's what makes it hard.

i to je ono što otežava stvari.

So, luckily, with machine learning, we can go beyond our own ability

Srećom, sa mašinskim učenjem, možemo da idemo dalje od naše sposobnosti

to describe the things we know.

da opišemo stvari koje znamo.

So how does this work? Well, it's by example.

Kako ovo funkcioniše? Pa, po primeru.

So we go to the machine and say, "Good morning, machine.

Odemo do mašine i kažemo, "Dobro jutro, mašino.

Here are some pick-and-rolls, and here are some things that are not.

Evo nekih pik end rolova, evo nekih stvari koje nisu.

Please find a way to tell the difference."

Molim te, nađi način da napraviš razliku."

And the key to all of this is to find features that enable it to separate.

I ključ za sve ovo je pronaći svojstva koja joj omogućavaju da to raščlani.

So if I was going to teach it the difference

Tako, ako bih hteo da je naučim razlici

between an apple and orange,

između jabuke i pomorandže,

I might say, "Why don't you use color or shape?"

mogao bih da kažem: "Što ne uzmeš boju ili obllk?"

And the problem that we're solving is, what are those things?

Problem koji rešavamo je, šta su to te stvari?

What are the key features

Koje su ključne stavke

that let a computer navigate the world of moving dots?

koje kompjuteru omogućavaju da upravlja svetom pokretnih tački?

So figuring out all these relationships with relative and absolute location,

Razumevši sve ove veze sa relativnom i apsolutnom lokacijom,

distance, timing, velocities --

udaljenost, tajming, brzina -

that's really the key to the science of moving dots, or as we like to call it,

to je zaista ključ nauke o pokretnim tačkama ili kako mi volimo da zovemo,

spatiotemporal pattern recognition, in academic vernacular.

spaciotemporalna šema prepoznavanja, akademskim žargonom govoreći.

Because the first thing is, you have to make it sound hard --

Jer kao prvo, mora da zvuči teško -

because it is.

jer to i jeste.

The key thing is, for NBA coaches, it's not that they want to know

Ključna stvar, za NBA trenere, nije to da žele da znaju

whether a pick-and-roll happened or not.

da li je došlo do pik end rola ili ne.

It's that they want to know how it happened.

Oni žele da znaju kako se odvijao.

And why is it so important to them? So here's a little insight.

A zašto je to njima tako važno? Evo malog uvida.

It turns out in modern basketball,

Izgleda da je u modernoj košarci

this pick-and-roll is perhaps the most important play.

ovaj pik end rol možda najvažniji deo igre.

And knowing how to run it, and knowing how to defend it,

Znati kako treba da se izvede, i kako da se odbrani,

is basically a key to winning and losing most games.

je suštinski ključ pobede ili poraza u većini utakmica.

So it turns out that this dance has a great many variations

Tako ispada da ovaj ples ima mnoge varijacije

and identifying the variations is really the thing that matters,

i identifikovanje tih varijacija je ono što je stvarno važno,

and that's why we need this to be really, really good.

i zato nam je potrebno da ovo bude baš, baš dobro.

So, here's an example.

Evo jednog primera.

There are two offensive and two defensive players,

Imamo dva igrača u napadu i dva igrača u odbrani,

getting ready to do the pick-and-roll dance.

spremni su da izvedu pik end rol ples.

So the guy with ball can either take, or he can reject.

Igrač sa loptom može ili prihvatiti, ili odbiti.

His teammate can either roll or pop.

Saigrač se može ili saviti ili otvoriti.

The guy guarding the ball can either go over or under.

Momak koji čuva loptu može ići iznad ili ispod.

His teammate can either show or play up to touch, or play soft

Njegov saigrač može ili da se otkrije ili da igra do kontakta, ili bez kontakta

and together they can either switch or blitz

i zajedno mogu ili da se zamene ili napadnu

and I didn't know most of these things when I started

a nisam znao većinu ovih stvari kada sam počinjao

and it would be lovely if everybody moved according to those arrows.

i bilo bi divno kada bi se svi pomerali u skladu sa ovim strelicama.

It would make our lives a lot easier, but it turns out movement is very messy.

To bi umnogome olakšalo naše živote, ali izgleda da su naši pokreti zbrkani.

People wiggle a lot and getting these variations identified

Ljudi se mnogo meškolje i dobijanje ovih identifikovanih varijacija

with very high accuracy,

sa veoma velikom tačnošću,

both in precision and recall, is tough

i u preciznosti i povlačenju, je teško

because that's what it takes to get a professional coach to believe in you.

jer zbog toga je potrebno da imaš profesionalnog trenera koji veruje u tebe.

And despite all the difficulties with the right spatiotemporal features

I uprkos svim poteškoćama sa tačnim spaciotemporalnim karakteristikama,

we have been able to do that.

mi smo to uradili.

Coaches trust our ability of our machine to identify these variations.

Treneri veruju mogućnostima naših mašina da identifikuju ove varijacije.

We're at the point where almost every single contender

Došli smo do tačke gde skoro svaki kandidat

for an NBA championship this year

za NBA šampionat ove godine

is using our software, which is built on a machine that understands

koristi naš softver, koji je ugrađen u mašinu koja razume

the moving dots of basketball.

pokretne tačke u košarci.

So not only that, we have given advice that has changed strategies

I ne samo to, davali smo savete koji menjaju strategije

that have helped teams win very important games,

koje pomažu timovima da dobiju veoma važne utakmice,

and it's very exciting because you have coaches who've been in the league

a to je vrlo uzbudljivo jer imate trenere koji su u ligi

for 30 years that are willing to take advice from a machine.

i po 30 godina i koji su spremni da prihvate savet od mašine.

And it's very exciting, it's much more than the pick-and-roll.

I to je vrlo uzbudljivo, to je mnogo više od pik end rola.

Our computer started out with simple things

Naš kompjuter je počeo sa prostim stvarima

and learned more and more complex things

i učio sve komplikovanije stvari

and now it knows so many things.

tako da sada zna dosta toga.

Frankly, I don't understand much of what it does,

Iskreno, ni ja ne razumem dosta toga što on radi,

and while it's not that special to be smarter than me,

i dok nije toliko teško biti pametniji od mene,

we were wondering, can a machine know more than a coach?

pitali smo se, može li neka mašina da zna više od trenera?

Can it know more than person could know?

Može li da zna više od nego čovek?

And it turns out the answer is yes.

I izgleda da je odgovor da.

The coaches want players to take good shots.

Treneri žele da im igrači imaju dobar šut.

So if I'm standing near the basket

Ako ja stojim blizu koša

and there's nobody near me, it's a good shot.

i nema nikoga u blizini, to je dobra pozicija.

If I'm standing far away surrounded by defenders, that's generally a bad shot.

Ako stojim daleko okružen odbranom, to je generalno loša pozicija.

But we never knew how good "good" was, or how bad "bad" was quantitatively.

Međutim, nikad kvantitativno nismo znali koliko je stvarno dobra ili loša.

Until now.

Do sada.

So what we can do, again, using spatiotemporal features,

Opet, ono što možemo uraditi, koristeći spaciotemporalne podatke,

we looked at every shot.

je da pogledamo svaki šut.

We can see: Where is the shot? What's the angle to the basket?

Možemo videti: Odakle ide? Pod kojim uglom je od koša?

Where are the defenders standing? What are their distances?

Gde stoji odbrana? Na kojoj su udaljenosti?

What are their angles?

Pod kojim su oni uglom?

For multiple defenders, we can look at how the player's moving

Za više odbrambenih igrača, možemo videti kako se igrač kreće

and predict the shot type.

i predvideti vrstu šuta.

We can look at all their velocities and we can build a model that predicts

Možemo videti svačiju brzinu i možemo napraviti model koji predviđa

what is the likelihood that this shot would go in under these circumstances?

koja je verovatnoća da će ovaj šut ući pod ovim okolnostima?

So why is this important?

Zašto je ovo važno?

We can take something that was shooting,

Možemo uzeti neki šut,

which was one thing before, and turn it into two things:

ranije gledan kao celina, i podeliti ga na dve stvari:

the quality of the shot and the quality of the shooter.

kvalitet šuta i kvalitet igrača.

So here's a bubble chart, because what's TED without a bubble chart?

I ovde imamo grafikon sa mehurićima, jer šta bi bio TED bez toga?

(Laughter)

(Smeh)

Those are NBA players.

Ovo su NBA igrači.

The size is the size of the player and the color is the position.

Veličina je veličina igrača a boja je njegova pozicija.

On the x-axis, we have the shot probability.

Na x-osi, imamo verovatnoću pogotka.

People on the left take difficult shots,

Ljudi s leva šutiraju iz teške pozicije,

on the right, they take easy shots.

sa desna, iz lakih pozicija.

On the [y-axis] is their shooting ability.

Na y-osi je njihov procenat pogotka.

People who are good are at the top, bad at the bottom.

Ljudi koji su dobri su pri vrhu, loši pri dnu.

So for example, if there was a player

Tako na primer, vidite igrača

who generally made 47 percent of their shots,

koji generalno ima 47 procenat šuta,

that's all you knew before.

i to je sve što ste znali.

But today, I can tell you that player takes shots that an average NBA player

Ali danas, mogu da vam kažem da igrač šutira onako kako bi prosečan NBA igrač

would make 49 percent of the time,

šutirao 49 posto vremena,

and they are two percent worse.

i gore je za dva procenta.

And the reason that's important is that there are lots of 47s out there.

A razlog što je ovo važno je što ovde ima dosta onih sa 47%.

And so it's really important to know

I tako je vrlo važno znati

if the 47 that you're considering giving 100 million dollars to

da li je onaj sa 47 kome razmišljate da platite 100 miliona dolara

is a good shooter who takes bad shots

dobar šuter iz teških pozciija,

or a bad shooter who takes good shots.

ili loš šuter koji pogađa lake koševe.

Machine understanding doesn't just change how we look at players,

Mašina ne utiče na naš pogled na igrače,

it changes how we look at the game.

već utiče na naš pogled na igru.

So there was this very exciting game a couple of years ago, in the NBA finals.

Tako je pre par godina, u NBA finalu, bila jedna vrlo zanimljiva utakmica.

Miami was down by three, there was 20 seconds left.

Majami je gubio sa tri razlike, 20 sekundi pre kraja.

They were about to lose the championship.

Bili su na pragu da izgube titulu.

A gentleman named LeBron James came up and he took a three to tie.

Gospodin po imenu LeBron Džejms je ušao i pucao trojku za izjednačenje.

He missed.

Promašio je.

His teammate Chris Bosh got a rebound,

Njegov saigrač Kris Boš skače i brani,

passed it to another teammate named Ray Allen.

pruža loptu saigraču, Reju Alenu.

He sank a three. It went into overtime.

On ubacuje za tri. Idu produžeci.

They won the game. They won the championship.

Pobedili su. Osvojili su šampionat.

It was one of the most exciting games in basketball.

To je bila jedna od najuzbudljivijih utakmica.

And our ability to know the shot probability for every player

I to što možemo da znamo verovatnoću pogotka svakog igrača

at every second,

u svakoj sekundi,

and the likelihood of them getting a rebound at every second

i verovatnoću skoka pod košem u svakoj sekundi

can illuminate this moment in a way that we never could before.

može da rasvetli ovaj trenutak na način na koji nikad ranije nismo mogli.

Now unfortunately, I can't show you that video.

Nažalost, sada vam ne mogu pokazati taj snimak.

But for you, we recreated that moment

Ali, za vas smo ponovili tu situaciju

at our weekly basketball game about 3 weeks ago.

na našoj nedeljnoj utakmici pre oko tri sedmice.

(Laughter)

(Smeh)

And we recreated the tracking that led to the insights.

I ponovo smo oživeli putanju koja je dovela do tog uvida.

So, here is us. This is Chinatown in Los Angeles,

I, evo nas. Ovo je Kineska četvrt u Los Anđelesu,

a park we play at every week,

park gde igramo svake nedelje,

and that's us recreating the Ray Allen moment

i evo ga ponovo trenutak Reja Alena

and all the tracking that's associated with it.

i svi pokreti u vezi sa tim.

So, here's the shot.

I evo šuta.

I'm going to show you that moment

Pokazaću vam taj trenutak

and all the insights of that moment.

i sve uvide u taj trenutak.

The only difference is, instead of the professional players, it's us,

Jedina razlika je što smo umesto profesionalnih igrača ovde mi,

and instead of a professional announcer, it's me.

a umesto profesionalnog komentatora, tu sam ja.

So, bear with me.

Pa ćete morati da me istrpite.

Miami.

Majami.

Down three.

Minus tri.

Twenty seconds left.

Još dvadeset sekundi.

Jeff brings up the ball.

Džef donosi loptu.

Josh catches, puts up a three!

Džoš je hvata, ubacuje trojku!

[Calculating shot probability]

[Računa se verovatnoća pogotka]

[Shot quality]

[Kvalitet šuta]

[Rebound probability]

[Mogućnost odbrane]

Won't go!

Neće ući!

[Rebound probability]

[Verovatnoća odbrane]

Rebound, Noel.

Brani Noel.

Back to Daria.

Vraća do Darije.

[Shot quality]

[Kvalitet šuta]

Her three-pointer -- bang!

Za tri poena - bam!

Tie game with five seconds left.

Izjednačenje pet sekundi pre kraja.

The crowd goes wild.

Publika je u transu.

(Laughter)

(Smeh)

That's roughly how it happened.

Otprilike tako nekako.

(Applause)

(Aplauz)

Roughly.

Otprilike.

(Applause)

(Aplauz)

That moment had about a nine percent chance of happening in the NBA

Taj momenat je imao šansu od devet procenata da se desi u NBA

and we know that and a great many other things.

i znamo to kao i mnogo drugih stvari.

I'm not going to tell you how many times it took us to make that happen.

Neću vam reći iz koliko pokušaja nam je ovo uspelo.

(Laughter)

(Smeh)

Okay, I will! It was four.

Okej, ipak hoću! Četiri puta.

(Laughter)

(Smeh)

Way to go, Daria.

Svaka čast, Darija.

But the important thing about that video

Ali ono što je važno u vezi sa ovim snimkom

and the insights we have for every second of every NBA game -- it's not that.

i uvidima koje imamo za svaki sekund svake NBA utakmice - nije to.

It's the fact you don't have to be a professional team to track movement.

To je činjenica da ne morate biti profesionalni tim da bi pratili kretanje.

You do not have to be a professional player to get insights about movement.

Ne morate biti profesionalni igrač da biste imali uvid u pokrete.

In fact, it doesn't even have to be about sports because we're moving everywhere.

U stvari, ne mora uopšte da se radi o sportu jer se mi krećemo svuda.

We're moving in our homes,

Krećemo se u našim domovima,

in our offices,

u kancelarijama,

as we shop and we travel

dok kupujemo i putujemo

throughout our cities

po gradu

and around our world.

ili po svetu.

What will we know? What will we learn?

Šta ćemo znati? Šta ćemo naučiti?

Perhaps, instead of identifying pick-and-rolls,

Možda, umesto identifikovanja pick-and-rolla,

a machine can identify the moment and let me know

mašina može da identifikuje trenutak i da me obavesti

when my daughter takes her first steps.

kada moja ćerka prohoda.

Which could literally be happening any second now.

Što bukvalno može da se desi svakog trenutka.

Perhaps we can learn to better use our buildings, better plan our cities.

Možda možemo bolje da koristimo zgrade, da bolje planiramo gradove.

I believe that with the development of the science of moving dots,

Verujem da ćemo se sa razvojem nauke pokretnih tački,

we will move better, we will move smarter, we will move forward.

bolje kretati, pametnije kretati, kretati napred.

Thank you very much.

Hvala vam mnogo.

(Applause)

(Aplauz)