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

1,107,248 views ใƒป 2015-07-06

TED


์•„๋ž˜ ์˜๋ฌธ์ž๋ง‰์„ ๋”๋ธ”ํด๋ฆญํ•˜์‹œ๋ฉด ์˜์ƒ์ด ์žฌ์ƒ๋ฉ๋‹ˆ๋‹ค.

๋ฒˆ์—ญ: Jeonghyeon Yeon ๊ฒ€ํ† : Gemma Lee
00:12
My colleagues and I are fascinated by the science of moving dots.
0
12954
3583
๋™๋ฃŒ์™€ ์ €๋Š” ์›€์ง์ด๋Š” ์ ์˜ ๊ณผํ•™์— ํ‘น ๋น ์ ธ์žˆ์Šต๋‹ˆ๋‹ค.
00:16
So what are these dots?
1
16927
1150
์ด ์ ๋“ค์€ ๋ฌด์—‡์ผ๊นŒ์š”?
00:18
Well, it's all of us.
2
18101
1287
๊ธ€์„ธ์š”. ์šฐ๋ฆฌ ๋ชจ๋‘์ฃ .
00:19
And we're moving in our homes, in our offices, as we shop and travel
3
19412
5085
์šฐ๋ฆฌ๋Š” ์ง‘์—์„œ๋‚˜ ํšŒ์‚ฌ์—์„œ๋„ ์›€์ง์ž…๋‹ˆ๋‹ค.
์‹œ๋‚ด๋ฅผ ๋Œ๋ฉฐ ์‡ผํ•‘ํ•˜๊ฑฐ๋‚˜ ์ „ ์„ธ๊ณ„๋ฅผ ์—ฌํ–‰ํ•  ๋•Œ์ฒ˜๋Ÿผ ๋ง์ด์ฃ .
00:24
throughout our cities and around the world.
4
24521
2066
00:26
And wouldn't it be great if we could understand all this movement?
5
26958
3669
์ด ๋ชจ๋“  ์›€์ง์ž„์„ ์ดํ•ดํ•  ์ˆ˜ ์žˆ๋‹ค๋ฉด ๊ต‰์žฅํ•˜์ง€ ์•Š์„๊นŒ์š”?
00:30
If we could find patterns and meaning and insight in it.
6
30918
2890
์›€์ง์ž„ ์†์— ๋‹ด๊ธด ํŒจํ„ด๊ณผ ์˜๋ฏธ๋ฅผ ์ฐพ์•„ ์ดํ•ดํ•  ์ˆ˜ ์žˆ๋‹ค๋ฉด
00:34
And luckily for us, we live in a time
7
34259
1785
๋‹คํ–‰ํžˆ ์ž์‹ ์— ๋Œ€ํ•œ ์ •๋ณด๋ฅผ ์ˆ˜์ง‘ํ•˜๋Š”๋ฐ ๊ต‰์žฅํžˆ ๋Šฅ์ˆ™ํ•œ ์‹œ๋Œ€์— ์‚ด๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.
00:36
where we're incredibly good at capturing information about ourselves.
8
36068
4497
00:40
So whether it's through sensors or videos, or apps,
9
40807
3663
์„ผ์„œ๋‚˜ ๋น„๋””์˜ค๋‚˜ ์–ดํ”Œ์„ ํ†ตํ•œ๋‹ค๋ฉด
00:44
we can track our movement with incredibly fine detail.
10
44494
2809
์—„์ฒญ๋‚˜๊ฒŒ ์•„์ฃผ ๋ฏธ์„ธํ•˜๊ณ  ์ž์„ธํžˆ ์›€์ง์ž„์„ ๋‹ด์•„๋‚ผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
์›€์ง์ž„์„ ๊ฐ€์žฅ ์ž˜ ์ˆ˜์ง‘ํ•  ์ˆ˜ ์žˆ๋Š” ๊ณณ ์ค‘ ํ•˜๋‚˜๊ฐ€
00:48
So it turns out one of the places where we have the best data about movement
11
48092
5032
00:53
is sports.
12
53148
1208
์Šคํฌ์ธ ์ฃ .
00:54
So whether it's basketball or baseball, or football or the other football,
13
54682
5333
๋†๊ตฌ, ์•ผ๊ตฌ, ๋ฏธ์‹์ถ•๊ตฌ, ์ถ•๊ตฌ ๊ฒฝ๊ธฐ๋ผ๋ฉด
01:00
we're instrumenting our stadiums and our players to track their movements
14
60039
4402
๊ฒฝ๊ธฐ์žฅ์—์„œ ์„ ์ˆ˜๋“ค์˜ ์›€์ง์ž„์„ ํฌ์ฐฉํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
01:04
every fraction of a second.
15
64465
1313
๋งค์ดˆ ๋ชจ๋“  ์›€์ง์ž„์„ ๋ง์ด์ฃ .
01:05
So what we're doing is turning our athletes into --
16
65802
4382
๋™๋ฃŒ์™€ ์ œ๊ฐ€ ํ•˜๋Š” ์ผ์€ ์„ ์ˆ˜๋“ค์„
01:10
you probably guessed it --
17
70208
1959
์—ฌ๋Ÿฌ๋ถ„์ด ์•„๋งˆ ์ง์ž‘ํ•˜๊ณ  ์žˆ๋Š”
01:12
moving dots.
18
72191
1396
์›€์ง์ด๋Š” ์ ์œผ๋กœ ๋ณ€ํ™˜ํ•˜๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค.
01:13
So we've got mountains of moving dots and like most raw data,
19
73946
4934
์›€์ง์ด๋Š” ์ ์„ ์—„์ฒญ ๋งŽ์ด ๊ฐ€์ง€๊ณ  ์žˆ๋Š”๋ฐ ๋Œ€๋ถ€๋ถ„์˜ ์›๋ฐ์ดํ„ฐ ์ฒ˜๋Ÿผ
01:18
it's hard to deal with and not that interesting.
20
78904
2502
์ฒ˜๋ฆฌํ•˜๊ธฐ๋„ ์–ด๋ ต๊ณ  ํฅ๋ฏธ๋กญ์ง€๋„ ์•Š์€ ๊ฒƒ์ด์ฃ .
01:21
But there are things that, for example, basketball coaches want to know.
21
81430
3769
ํ•˜์ง€๋งŒ ์˜ˆ๋ฅผ ๋“ค๋ฉด ๋†๊ตฌ ์ฝ”์น˜๋“ค์ด ์•Œ๊ณ  ์‹ถ์–ดํ•˜๋Š” ๊ฒƒ๋“ค์ด ์žˆ์ฃ .
01:25
And the problem is they can't know them because they'd have to watch every second
22
85223
3810
๋งค ๊ฒฝ๊ธฐ๋ฅผ ๋ชจ๋‘ ์ง€์ผœ๋ณด๊ณ  ๊ธฐ์–ตํ–ˆ๋‹ค๊ฐ€ ์ฒ˜๋ฆฌํ•ด์•ผ ํ•˜๊ธฐ ๋•Œ๋ฌธ์— ์•Œ ์ˆ˜ ์—†์ฃ .
01:29
of every game, remember it and process it.
23
89057
2589
01:31
And a person can't do that,
24
91804
1930
์‚ฌ๋žŒ์ด ๊ทธ๊ฒƒ์„ ํ•  ์ˆ˜ ์—†์ฃ .
01:33
but a machine can.
25
93758
1310
ํ•˜์ง€๋งŒ ๊ธฐ๊ณ„๋Š” ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
01:35
The problem is a machine can't see the game with the eye of a coach.
26
95661
3410
๋ฌธ์ œ๋Š” ๊ธฐ๊ณ„๊ฐ€ ์ฝ”์น˜์˜ ๋ˆˆ์œผ๋กœ ๊ฒฝ๊ธฐ๋ฅผ ์ง€์ผœ๋ณผ ์ˆ˜ ์—†๋‹ค๋Š” ๊ฑฐ์ฃ .
01:39
At least they couldn't until now.
27
99363
2261
์ ์–ด๋„ ์ง€๊ธˆ๊นŒ์ง€๋Š”์š”.
01:42
So what have we taught the machine to see?
28
102228
2103
๊ธฐ๊ณ„๊ฐ€ ๋ณผ ์ˆ˜ ์žˆ๊ฒŒ ํ•˜๋ ค๋ฉด ๋ฌด์—‡์„ ๊ฐ€๋ฅด์ณ์•ผ ํ• ๊นŒ์š”?
01:45
So, we started simply.
29
105569
1787
๊ฐ„๋‹จํ•œ ๊ฒƒ๋ถ€ํ„ฐ ์‹œ์ž‘ํ–ˆ์–ด์š”.
01:47
We taught it things like passes, shots and rebounds.
30
107380
3799
ํŒจ์Šค, ์Š›, ๋ฆฌ๋ฐ”์šด๋“œ ๊ฐ™์€ ๊ฒƒ๋“ค์„ ๊ฐ€๋ฅด์ณค์ฃ .
01:51
Things that most casual fans would know.
31
111203
2541
๋ณดํ†ต ํŒฌ์ด๋ผ๋ฉด ๊ทธ ์ •๋„๋Š” ์•Œ ์ˆ˜ ์žˆ๋Š” ๊ฑฐ์ฃ .
01:53
And then we moved on to things slightly more complicated.
32
113768
2832
๊ทธ๋Ÿฐ ๋‹ค์Œ ์กฐ๊ธˆ ๋” ์–ด๋ ค์šด ๊ฒƒ์„ ๊ฐ€๋ฅด์ณค์ฃ .
01:56
Events like post-ups, and pick-and-rolls, and isolations.
33
116624
4588
ํฌ์ŠคํŠธ์—…, ํ”ฝ์•ค๋กค, ์•„์ด์†”๋ ˆ์ด์…˜ ๊ฐ™์€ ๊ธฐ์ˆ ์ด์ฃ .
02:01
And if you don't know them, that's okay. Most casual players probably do.
34
121377
3543
์ด๋Ÿฐ ๊ธฐ์ˆ ์„ ๋ชฐ๋ผ๋„ ๊ดœ์ฐฎ์Šต๋‹ˆ๋‹ค. ๋Œ€๋ถ€๋ถ„ ๋ณดํ†ต ํŒฌ๋“ค๋„ ์•„๋งˆ ๊ทธ๋Ÿด ๊ฑฐ์˜ˆ์š”.
02:05
Now, we've gotten to a point where today, the machine understands complex events
35
125560
5340
ํ˜„์žฌ๋Š” ๊ธฐ๊ณ„๊ฐ€ ๋ณต์žกํ•œ ๊ธฐ์ˆ ์„ ์ดํ•ดํ•˜๋Š” ์ˆ˜์ค€๊นŒ์ง€ ์™€ ์žˆ์Šต๋‹ˆ๋‹ค.
02:10
like down screens and wide pins.
36
130924
3073
๋‹ค์šด ์Šคํฌ๋ฆฐ๊ณผ ์™€์ด๋“œ ํ•€ ๊ฐ™์€ ๊ธฐ์ˆ  ๋ง์ด์ฃ .
02:14
Basically things only professionals know.
37
134021
2726
๊ธฐ๋ณธ์ ์œผ๋กœ ์ด๋Ÿฐ ๊ฒƒ๋“ค์€ ์ „๋ฌธ๊ฐ€๋งŒ ์•„๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค.
02:16
So we have taught a machine to see with the eyes of a coach.
38
136771
4388
์šฐ๋ฆฌ๋Š” ๊ธฐ๊ณ„๊ฐ€ ์ฝ”์น˜์˜ ๋ˆˆ์œผ๋กœ ๋ณผ ์ˆ˜ ์žˆ๋„๋ก ๊ฐ€๋ฅด์ณ ์™”์Šต๋‹ˆ๋‹ค.
02:22
So how have we been able to do this?
39
142009
1857
์šฐ๋ฆฌ๊ฐ€ ์–ด๋–ป๊ฒŒ ์ด๊ฒƒ์„ ํ•  ์ˆ˜ ์žˆ์—ˆ์„๊นŒ์š”?
02:24
If I asked a coach to describe something like a pick-and-roll,
40
144511
3118
์ œ๊ฐ€ ์ฝ”์น˜์—๊ฒŒ ํ”ฝ์•ค๋กค ๊ฐ™์€ ๊ฒƒ์„ ์„ค๋ช…ํ•ด ๋‹ฌ๋ผ๊ณ  ํ•˜๋ฉด
02:27
they would give me a description,
41
147653
1640
์ œ๊ฒŒ ์„ค๋ช…ํ•˜์ฃ .
๊ทธ๋Ÿฌ๋ฉด ์ „ ๊ทธ๊ฒƒ์„ ์•Œ๊ณ ๋ฆฌ์ฆ˜์œผ๋กœ ์ธ์ฝ”๋”ฉํ•˜๋ ค๋ฉด ์ฃฝ์„ ๋ง›์ด์ฃ .
02:29
and if I encoded that as an algorithm, it would be terrible.
42
149317
2856
02:33
The pick-and-roll happens to be this dance in basketball between four players,
43
153026
4278
ํ”ฝ์•ค๋กค์€ ๋†๊ตฌ์—์„œ 4๋ช… ์„ ์ˆ˜ ์‚ฌ์ด์—์„œ ์ด๋Ÿฐ ์‹์œผ๋กœ ์›€์ง์ด๋ฉฐ ์ผ์–ด๋‚ฉ๋‹ˆ๋‹ค.
02:37
two on offense and two on defense.
44
157328
1912
๋‘๋ช…์€ ๊ณต๊ฒฉํ•˜๊ณ  ๋‘ ๋ช…์€ ๋ฐฉ์–ดํ•˜๋ฉด์„œ ๋ง์ด์ฃ .
02:39
And here's kind of how it goes.
45
159486
1618
์˜ˆ๋ฅผ ๋“ค๋ฉด ์ด๋Ÿฐ ๊ฑฐ์˜ˆ์š”.
02:41
So there's the guy on offense without the ball
46
161128
2533
๊ณต ์—†์ด ๊ณต๊ฒฉํ•˜๋Š” ์„ ์ˆ˜๊ฐ€ ์žˆ์Šต๋‹ˆ๋‹ค.
02:43
the ball and he goes next to the guy guarding the guy with the ball,
47
163685
3209
๊ณต์„ ๊ฐ€์ง€๊ณ  ์žˆ๋Š” ์„ ์ˆ˜๋ฅผ ์ˆ˜๋น„ํ•˜๋Š” ์„ ์ˆ˜ ์˜†์„ ์ง€๋‚˜๊ฐ€๋‹ค๊ฐ€
02:46
and he kind of stays there
48
166918
1257
์ž ์‹œ ๊ฑฐ๊ธฐ์— ์žˆ๋‹ค๊ฐ€
02:48
and they both move and stuff happens, and ta-da, it's a pick-and-roll.
49
168199
3317
๊ทธ ์„ ์ˆ˜ ๋‘˜์ด ์›€์ง์˜€๋‹ค๊ฐ€ ์ฑ„์›Œ๋„ฃ์œผ๋ฉด์„œ ์งœ์ž”, ์ด๊ฒŒ ํ”ฝ์•ค๋กค์ด์ฃ .
02:51
(Laughter)
50
171540
2215
(์›ƒ์Œ)
02:53
So that is also an example of a terrible algorithm.
51
173779
2508
๊ทธ๊ฒƒ ์—ญ์‹œ ๋ณต์žกํ•œ ์•Œ๊ณ ๋ฆฌ์ฆ˜์˜ ์˜ˆ์ž…๋‹ˆ๋‹ค.
02:56
So, if the player who's the interferer -- he's called the screener --
52
176913
4204
๋ง‰์•„ ์ฃผ๋Š” ์„ ์ˆ˜๋ฅผ ์Šคํฌ๋ฆฌ๋„ˆ๋ผ๊ณ  ๋ถ€๋ฅด๋Š”๋ฐ
03:01
goes close by, but he doesn't stop,
53
181278
2872
๊ฐ€๊นŒ์ด ๋‹ค๊ฐ€๊ฐ€์ง€๋งŒ ๋ง‰์•„์ฃผ์ง€ ์•Š๋Š”๋‹ค๋ฉด
03:04
it's probably not a pick-and-roll.
54
184174
1765
์•„๋งˆ ํ”ฝ์•ค๋กค์€ ์•„๋‹Œ ๊ฒƒ ๊ฐ™๋„ค์š”.
03:06
Or if he does stop, but he doesn't stop close enough,
55
186560
3945
๊ทธ๊ฐ€ ๋ง‰์ง€๋งŒ, ๊ฐ€๊นŒ์ด์—์„œ ์ถฉ๋ถ„ํžˆ ๋ง‰์ง€ ์•Š๋Š”๋‹ค๋ฉด
03:10
it's probably not a pick-and-roll.
56
190529
1761
์•„๋งˆ ํ”ฝ์•ค๋กค์ด ์•„๋‹ ๊ฑฐ์˜ˆ์š”.
03:12
Or, if he does go close by and he does stop
57
192642
3237
๋˜๋Š” ๊ฐ€๊นŒ์ด ๋‹ค๊ฐ€๊ฐ€์„œ ๋ง‰๋Š”๋ฐ
03:15
but they do it under the basket, it's probably not a pick-and-roll.
58
195903
3324
๊ณจ๋Œ€ ์•„๋ž˜์„œ์—์„œ ๊ทธ๋ ‡๊ฒŒ ํ•œ๋‹ค๋ฉด ์•„๋งˆ ํ”ฝ์•ค๋กค์ด ์•„๋‹ ๊ฑฐ์˜ˆ์š”.
03:19
Or I could be wrong, they could all be pick-and-rolls.
59
199462
2524
์ œ๊ฐ€ ํ‹€๋ ธ๋‹ค๋ฉด ์ง€๊ธˆ๊นŒ์ง€ ๋งํ•œ ๋ชจ๋‘๊ฐ€ ํ”ฝ์•ค๋กค์ด๊ฒ ์ฃ .
03:22
It really depends on the exact timing, the distances, the locations,
60
202010
4568
์ •ํ™•ํ•œ ํƒ€์ด๋ฐ, ๊ฑฐ๋ฆฌ, ์œ„์น˜์— ๋”ฐ๋ผ ์ง„์งœ ๋‹ฌ๋ผ์ง‘๋‹ˆ๋‹ค.
03:26
and that's what makes it hard.
61
206602
1495
๊ทธ๋Ÿฌ๋‹ˆ ๊ทธ๊ฒŒ ํž˜๋“  ๊ฑฐ์ง€์š”.
03:28
So, luckily, with machine learning, we can go beyond our own ability
62
208579
4944
๋‹คํ–‰์ด ๊ธฐ๊ณ„๊ฐ€ ํ•™์Šต์„ ํ•˜๋ฉด์„œ ์šฐ๋ฆฌ๊ฐ€ ์•Œ๊ณ  ์žˆ๋Š” ๊ฒƒ์„ ์„ค๋ช…ํ•˜๋Š”๋ฐ
03:33
to describe the things we know.
63
213547
1743
์šฐ๋ฆฌ์˜ ๋Šฅ๋ ฅ์„ ๋„˜์–ด์„ค ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
03:35
So how does this work? Well, it's by example.
64
215314
2280
์ด๊ฒŒ ์–ด๋–ป๊ฒŒ ๊ฐ€๋Šฅํ• ๊นŒ์š”? ์˜ˆ๋ฅผ ๋“ค์–ด๋ณด์ฃ .
03:37
So we go to the machine and say, "Good morning, machine.
65
217759
2830
์šฐ๋ฆฌ๊ฐ€ ๊ธฐ๊ณ„์—๊ฒŒ ๊ฐ€์„œ "์ข‹์€ ์•„์นจ์ด๋„ค. ๊ธฐ๊ณ„ ์”จ.
03:41
Here are some pick-and-rolls, and here are some things that are not.
66
221077
3359
์—ฌ๊ธฐ ๋ช‡ ๊ฐ€์ง€ ํ”ฝ์•ค๋กค๊ณผ ์ด์ „์— ์—†๋˜ ๋ช‡ ๊ฐ€์ง€๊ฐ€ ์žˆ๋„ค.
03:44
Please find a way to tell the difference."
67
224720
2252
๋ฐฉ๋ฒ•์„ ์ฐพ์•„ ์ฐจ์ด์ ์„ ๋งํ•ด์คฌ์œผ๋ฉด ํ•ด."
03:47
And the key to all of this is to find features that enable it to separate.
68
227076
3707
์ด ๋ชจ๋“  ๊ฒƒ์˜ ํ•ต์‹ฌ์€ ํŠน์ง•์„ ์ฐพ์•„ ๊ทธ๊ฒƒ์„ ๋ถ„๋ฆฌํ•ด๋‚ด๋Š” ๊ฑฐ์ฃ .
03:50
So if I was going to teach it the difference
69
230807
2109
๋งŒ์•ฝ ์ œ๊ฐ€ ์‚ฌ๊ณผ์™€ ์˜ค๋ Œ์ง€์˜ ์ฐจ์ด์ ์„ ๊ฐ€๋ฅด์น˜๋ ค๊ณ  ํ•˜๋ฉด
03:52
between an apple and orange,
70
232940
1381
"์ƒ‰๊ณผ ๋ชจ์–‘์„ ์‚ฌ์šฉํ•˜๋Š” ๊ฒŒ ์–ด๋•Œ?"๋ผ๊ณ  ๋งํ•˜๊ฒ ์ฃ .
03:54
I might say, "Why don't you use color or shape?"
71
234345
2375
03:56
And the problem that we're solving is, what are those things?
72
236744
2943
์šฐ๋ฆฌ๊ฐ€ ํ’€๋ ค๋Š” ๋ฌธ์ œ๋Š” ๊ทธ๊ฒƒ๋“ค์ด ๋ญ”๊ฐ€๋ผ๋Š” ๊ฑฐ์ฃ .
03:59
What are the key features
73
239711
1247
์–ด๋–ค ํ•ต์‹ฌ๊ธฐ๋Šฅ์œผ๋กœ
04:00
that let a computer navigate the world of moving dots?
74
240982
3499
์ปดํ“จํ„ฐ๊ฐ€ ์›€์ง์ด๋Š” ์ ์˜ ์„ธ๊ณ„๋ฅผ ์ฒ˜๋ฆฌํ•  ์ˆ˜ ์žˆ์„๊นŒ์š”?
04:04
So figuring out all these relationships with relative and absolute location,
75
244505
4823
์ƒ๋Œ€์ ์ด๊ณ  ์ ˆ๋Œ€์ ์ธ ์œ„์น˜, ๊ฑฐ๋ฆฌ, ํƒ€์ด๋ฐ, ์†๋„๋กœ
04:09
distance, timing, velocities --
76
249352
1909
์ด ๋ชจ๋“  ๊ฒƒ์„ ์„ค๋ช…ํ•œ๋‹ค๋Š” ๊ฒƒ์ด
04:11
that's really the key to the science of moving dots, or as we like to call it,
77
251440
4928
๊ทธ๊ฒŒ "์›€์ง์ด๋Š” ์ ์˜ ๊ณผํ•™"์—์„œ ์ง„์งœ ํ•ต์‹ฌ์ด์ฃ .
ํ•™๊ณ„์—์„œ ์“ฐ๋Š” ๋ง๋กœ๋Š” ์‹œ๊ณต๊ฐ„์  ํŒจํ„ด ์ธ์‹์ด๋ผ๊ณ  ํ•˜์ฃ .
04:16
spatiotemporal pattern recognition, in academic vernacular.
78
256392
3344
04:19
Because the first thing is, you have to make it sound hard --
79
259925
2898
์—ฌ๋Ÿฌ๋ถ„์ด ๋ฐœ์Œํ•˜๊ธฐ ํž˜๋“ค๊ธฐ ๋•Œ๋ฌธ์—
04:22
because it is.
80
262847
1278
์„ธ๊ฒŒ ๋ฐœ์Œํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค.
04:24
The key thing is, for NBA coaches, it's not that they want to know
81
264410
3141
ํ•ต์‹ฌ์€ NBA์ฝ”์น˜๋“ค์ด ์•Œ๊ณ  ์‹ถ์–ดํ•˜๋Š” ๊ฒƒ์ด
04:27
whether a pick-and-roll happened or not.
82
267575
1922
ํ”ฝ์•ค๋กค์ด ์ผ์–ด๋‚ฌ๋Š”์ง€ ์•ˆ ์ผ์–ด๋‚ฌ๋Š”์ง€๊ฐ€ ์•„๋‹ˆ๋ผ
04:29
It's that they want to know how it happened.
83
269521
2076
์–ด๋–ป๊ฒŒ ์ผ์–ด๋‚˜๋Š๋ƒํ•˜๋Š” ๊ฑฐ์ฃ .
04:31
And why is it so important to them? So here's a little insight.
84
271621
2986
์™œ ๊ทธ๊ฒŒ ์ฝ”์น˜์—๊ฒŒ ๊ทธ๋ ‡๊ฒŒ ์ค‘์š”ํ• ๊นŒ์š”? ๋ฐ”๋กœ ์ง๊ด€๋•Œ๋ฌธ์ด์ฃ .
04:34
It turns out in modern basketball,
85
274631
1771
ํ˜„๋Œ€ ๋†๊ตฌ์—์„œ ๋ฐํ˜€์ง„ ๋ฐ”๋กœ๋Š”
04:36
this pick-and-roll is perhaps the most important play.
86
276426
2539
์ด๋Ÿฐ ํ”ฝ์•ค๋กค ๊ธฐ์ˆ ์ด ์•„๋งˆ ๊ฐ€์žฅ ์ค‘์š”ํ•œ ํ”Œ๋ ˆ์ด์ž…๋‹ˆ๋‹ค.
04:39
And knowing how to run it, and knowing how to defend it,
87
279065
2620
์–ด๋–ป๊ฒŒ ๋‹ฌ๋ฆฌ๊ณ  ๋ฐฉ์–ดํ•˜๋Š”์ง€ ์•„๋Š” ๊ฒƒ์ด
04:41
is basically a key to winning and losing most games.
88
281709
2670
๊ธฐ๋ณธ์ ์œผ๋กœ ๋Œ€๋ถ€๋ถ„์˜ ๊ฒฝ๊ธฐ์˜ ์ŠนํŒจ๋ฅผ ๊ฐ€๋ฅด๋Š” ์—ด์‡ ์ž…๋‹ˆ๋‹ค.
04:44
So it turns out that this dance has a great many variations
89
284403
3801
์ด ๊ธฐ์ˆ ์ด ๋งค์šฐ ๋งŽ์€ ๋ณ€ํ˜•์ด ์žˆ๋‹ค๋Š” ๊ฒƒ์„ ์•Œ์•„๋ƒˆ์ฃ .
04:48
and identifying the variations is really the thing that matters,
90
288228
3648
๋ณ€ํ˜•์„ ์•Œ์•„๋‚ด๋Š” ๊ฒƒ์ด ์ง„์งœ ์ค‘์š”ํ•œ ์ผ์ด์ฃ .
04:51
and that's why we need this to be really, really good.
91
291900
2529
์•„์ฃผ ์•„์ฃผ ์ž˜ ํ•ด๋‚ด๋ ค๋ฉด ์ด๊ฒƒ์ด ํ•„์š”ํ•œ ์ด์œ ์ฃ .
04:55
So, here's an example.
92
295228
1176
์˜ˆ๋ฅผ ๋“ค๋ฉด,
04:56
There are two offensive and two defensive players,
93
296428
2379
๋‘ ๊ณต๊ฒฉ ์„ ์ˆ˜์™€ ๋‘ ์ˆ˜๋น„ ์„ ์ˆ˜๊ฐ€ ์žˆ์ฃ .
04:58
getting ready to do the pick-and-roll dance.
94
298831
2152
ํ”ฝ์•ค๋กค ๊ธฐ์ˆ ์„ ํ•˜๋ ค๊ณ  ์ค€๋น„ํ•ฉ๋‹ˆ๋‹ค.
๊ณต์„ ๊ฐ€์ง„ ์„ ์ˆ˜๊ฐ€ ๊ทธ ๊ธฐ์ˆ ์„ ๋ฐ›์•„๋“ค์ผ ์ˆ˜๋„ ๊ฑฐ๋ถ€ํ•  ์ˆ˜๋„ ์žˆ์ฃ .
05:01
So the guy with ball can either take, or he can reject.
95
301007
2683
05:04
His teammate can either roll or pop.
96
304086
3001
๊ทธ์˜ ๋™๋ฃŒ๋Š” ๋’ค๋กœ ๋Œ์•„๋‚˜๊ฐ€๊ฑฐ๋‚˜ ์™ธ๊ณฝ์ชฝ์œผ๋กœ ๋น ์งˆ ์ˆ˜ ์žˆ์ฃ .
05:07
The guy guarding the ball can either go over or under.
97
307111
2986
์ˆ˜๋น„ ์„ ์ˆ˜๋Š” ์œ„์ชฝ ๋˜๋Š” ์•„๋ž˜์ชฝ์œผ๋กœ ๋›ธ ์ˆ˜ ์žˆ์ฃ .
๊ทธ์˜ ๋™๋ฃŒ๋Š” ๊ณต์„ ๊ฐ€์ง„ ์„ ์ˆ˜๋ฅผ ๋”ฐ๋ผ๊ฐ€๊ฑฐ๋‚˜ ์Šคํฌ๋ฆฌ๋„ˆ๋ฅผ ๋ง‰๊ฑฐ๋‚˜
05:10
His teammate can either show or play up to touch, or play soft
98
310121
4565
๋’ค์ชฝ์œผ๋กœ ๋น ์ง€๊ฑฐ๋‚˜
05:14
and together they can either switch or blitz
99
314710
2618
๋‘˜๋‹ค ์Šค์œ„์น˜๋‚˜ ๋ธ”๋ฆฌ์น˜๋ฅผ ํ•  ์ˆ˜ ์žˆ์ฃ .
05:17
and I didn't know most of these things when I started
100
317352
2659
์ œ๊ฐ€ ์‹œ์ž‘ํ•  ๋•Œ๋Š” ์ด๋Ÿฐ ๊ฒƒ๋“ค์„ ๋Œ€๋ถ€๋ถ„ ๋ชฐ๋ž์–ด์š”.
05:20
and it would be lovely if everybody moved according to those arrows.
101
320035
3920
๋ชจ๋‘๊ฐ€ ์ด๋Ÿฐ ํ™”์‚ดํ‘œ๋Œ€๋กœ ์›€์ง์ธ๋‹ค๋ฉด ์ข‹์•˜์„ ๊ฑฐ์˜ˆ์š”.
05:23
It would make our lives a lot easier, but it turns out movement is very messy.
102
323979
3905
๊ทธ๋Ÿฌ๋ฉด ์•„์ฃผ ์‰ฌ์› ์„ ํ…๋ฐ ์›€์ง์ž„์€ ๋งค์šฐ ๋ณต์žกํ–ˆ์ฃ .
์‚ฌ๋žŒ๋“ค์€ ๋งค์šฐ ์ž์ฃผ ์›€์ง์ด๊ณ  ๋ณ€ํ˜•๋„ ๋งŽ์•˜์ฃ .
05:28
People wiggle a lot and getting these variations identified
103
328047
5484
๋งค์šฐ ์ •๋ฐ€ํ•˜๊ฒŒ ํ™•์ธํ•ด ๋ฐ”๋กœ๋Š”์š”.
05:33
with very high accuracy,
104
333555
1303
05:34
both in precision and recall, is tough
105
334882
1868
์ •ํ™•ํ•˜๊ฒŒ ํšŒ์ƒํ•˜๊ธฐ๋Š” ํž˜๋“ค์—ˆ์–ด์š”.
05:36
because that's what it takes to get a professional coach to believe in you.
106
336774
3618
์—ฌ๋Ÿฌ๋ถ„์ด ๋ฏฟ์„ ๋งŒํ•œ ์ „๋ฌธ ์ฝ”์น˜๋ฅผ ๊ตฌํ•˜๋Š” ๋ฐ ์‹œ๊ฐ„์ด ๊ฑธ๋ ธ๊ธฐ ๋•Œ๋ฌธ์ด์ฃ .
05:40
And despite all the difficulties with the right spatiotemporal features
107
340416
3380
์ •ํ™•ํ•œ ์‹œ๊ณต๊ฐ„ ํŠน์ง•์„ ์ฐพ์•„๋‚ธ๋‹ค๋Š” ๊ฒƒ์ด ์–ด๋ ค์›€์—๋„ ๋ถˆ๊ตฌํ•˜๊ณ 
05:43
we have been able to do that.
108
343820
1474
๊ทธ๊ฒƒ์„ ํ•ด๋‚ผ ์ˆ˜ ์žˆ์—ˆ์Šต๋‹ˆ๋‹ค.
์ฝ”์น˜๊ฐ€ ๊ธฐ๊ณ„์˜ ๋Šฅ๋ ฅ์„ ๋ฏฟ๊ณ  ์ด๋Ÿฐ ๋ณ€ํ˜•์„ ํ™•์ธํ–ˆ์ฃ .
05:45
Coaches trust our ability of our machine to identify these variations.
109
345318
3927
05:49
We're at the point where almost every single contender
110
349478
3533
๊ฑฐ์˜ ๋ชจ๋“  ๊ฒฝ์Ÿ์ž๋“ค์ด
05:53
for an NBA championship this year
111
353035
1623
์˜ฌํ•ด NBA ์„ ์ˆ˜๊ถŒ์„ ์œ„ํ•ด
05:54
is using our software, which is built on a machine that understands
112
354682
4408
์ด ์†Œํ”„ํŠธ์›จ์–ด๋ฅผ ์ด์šฉํ•˜๋ ค๊ณ  ํ•˜๊ณ  ์žˆ์ฃ .
๋†๊ตฌ์˜ ์›€์ง์ด๋Š” ์ ์„ ์ดํ•ดํ•˜๋Š” ๊ธฐ๊ณ„๋กœ ๋งŒ๋“  ์†Œํ”„ํŠธ์›จ์–ด ๋ง์ด์—์š”.
05:59
the moving dots of basketball.
113
359114
1634
06:01
So not only that, we have given advice that has changed strategies
114
361872
5153
๊ทธ๊ฒƒ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ์ „์ˆ ์„ ๋ฐ”๊พธ๋Š” ์กฐ์–ธ์„ ํ•ด ์ค˜
06:07
that have helped teams win very important games,
115
367049
3352
ํŒ€์„ ๋„์™€ ๋งค์šฐ ์ค‘์š”ํ•œ ๊ฒฝ๊ธฐ์—์„œ ์Šน๋ฆฌํ•˜๊ฒŒ ํ–ˆ์ฃ .
06:10
and it's very exciting because you have coaches who've been in the league
116
370425
3732
๋งค์šฐ ๋“ค๋–ด๋˜ ์ด์œ ๋Š” 30๋…„๋™์•ˆ ๋ฆฌ๊ทธ์—์„œ ์žˆ์—ˆ๋˜ ์ฝ”์น˜๊ฐ€
06:14
for 30 years that are willing to take advice from a machine.
117
374181
3067
๊ธฐ๊ณ„์—์„œ ๋‚˜์˜จ ์กฐ์–ธ์„ ๊ธฐ๊บผ์ด ์ˆ˜์šฉํ–ˆ๊ธฐ ๋•Œ๋ฌธ์ด์ฃ .
06:17
And it's very exciting, it's much more than the pick-and-roll.
118
377874
2906
ํ”ฝ์•ค๋กค์„ ์•Œ์•˜์„ ๋•Œ๋ณด๋‹ค ํ›จ์”ฌ ๋” ์‹ ๋‚ฌ์–ด์š”.
06:20
Our computer started out with simple things
119
380804
2076
์ปดํ“จํ„ฐ๋Š” ๊ฐ„๋‹จํ•œ ๊ฒƒ๋ถ€ํ„ฐ ์‹œ์ž‘ํ•ด์„œ
06:22
and learned more and more complex things
120
382904
2064
์ ์  ๋” ๋ณต์žกํ•œ ๊ฒƒ์„ ํ•™์Šตํ–ˆ๊ณ 
06:24
and now it knows so many things.
121
384992
1561
์ง€๊ธˆ์€ ๋งŽ์€ ๊ฒƒ์„ ์•Œ๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.
06:26
Frankly, I don't understand much of what it does,
122
386577
2835
์†”์งํžˆ ์ „ ์ปดํ“จํ„ฐ๊ฐ€ ํ•˜๋Š” ์ผ์„ ๋‹ค ์ดํ•ดํ•˜์ง€ ๋ชปํ•˜์ง€๋งŒ
06:29
and while it's not that special to be smarter than me,
123
389436
3715
์ €๋ณด๋‹ค ๋” ๋˜‘๋˜‘ํ•ด์ง„๋‹ค๊ณ  ๊ทธ๋ ‡๊ฒŒ ํŠน๋ณ„ํ•ด์ง€์ง€ ์•Š๊ฒ ์ง€๋งŒ
06:33
we were wondering, can a machine know more than a coach?
124
393175
3644
๊ถ๊ธˆํ•˜๊ธฐ๋Š” ํ•˜๋„ค์š”. ๊ธฐ๊ณ„๊ฐ€ ์ฝ”์น˜๋ณด๋‹ค ๋” ๋งŽ์ด ์•Œ ์ˆ˜ ์žˆ์„๊นŒ์š”?
06:36
Can it know more than person could know?
125
396843
2055
์‚ฌ๋žŒ๋ณด๋‹ค ๋” ๋งŽ์ด ์•Œ ์ˆ˜ ์žˆ์„๊นŒ์š”?
06:38
And it turns out the answer is yes.
126
398922
1745
๋ฐํ˜€์ง„ ๋ฐ”๋กœ๋Š” "๋„ค."์ž…๋‹ˆ๋‹ค.
06:40
The coaches want players to take good shots.
127
400691
2557
์ฝ”์น˜๋Š” ์„ ์ˆ˜๋“ค์ด ์Š›์„ ์ž˜ ์˜๊ธฐ ๋ฐ”๋ž๋‹ˆ๋‹ค.
06:43
So if I'm standing near the basket
128
403272
1651
์ œ๊ฐ€ ๊ณจ๋Œ€ ๊ทผ์ฒ˜์— ์„œ ์žˆ๊ณ 
06:44
and there's nobody near me, it's a good shot.
129
404947
2166
๊ทผ์ฒ˜์— ์•„๋ฌด๋„ ์—†๋‹ค๋ฉด ์Š›ํ•˜๊ธฐ ์ข‹์ฃ .
์ˆ˜๋น„์ˆ˜์— ๋‘˜๋Ÿฌ์‹ธ์—ฌ ๊ณจ๋Œ€ ๋ฉ€๋ฆฌ ์„œ์žˆ๋‹ค๋ฉด ๋Œ€๊ฐœ ์Š›ํ•˜๊ธฐ ์ข‹์ง€ ์•Š์ฃ .
06:47
If I'm standing far away surrounded by defenders, that's generally a bad shot.
130
407137
3940
06:51
But we never knew how good "good" was, or how bad "bad" was quantitatively.
131
411101
4876
ํ•˜์ง€๋งŒ ๊ฒฐ์ฝ” ์•Œ์ง€ ๋ชปํ•˜์ฃ .
์ข‹์€ ๊ฒŒ ์–ผ๋งˆ๋‚˜ ์ข‹์€์ง€ ๋‚˜์œ ๊ฒŒ ์–ผ๋งˆ๋‚˜ ๋‚˜์œ์ง€๋ฅผ์š”.
06:56
Until now.
132
416209
1150
ํ˜„์žฌ๊นŒ์ง€๋Š”์š”.
06:57
So what we can do, again, using spatiotemporal features,
133
417771
3058
ํ•  ์ˆ˜ ์žˆ๋Š” ๊ฒƒ์€ ๋‹ค์‹œ ์‹œ๊ณต๊ฐ„์ ์ธ ํŠน์ง•์„ ์ด์šฉํ•ด์„œ
07:00
we looked at every shot.
134
420853
1374
์Š›์„ ๋ชจ๋“  ๋ฐฉํ–ฅ์—์„œ ๋ณด์•˜์Šต๋‹ˆ๋‹ค.
07:02
We can see: Where is the shot? What's the angle to the basket?
135
422251
3005
์Š›์„ ์–ด๋””์„œ ํ–ˆ๊ณ  ๊ณจ๋Œ€๊นŒ์ง€ ๊ฐ๋„๊ฐ€ ์–ด๋–ค์ง€ ์•Œ ์ˆ˜ ์žˆ์—ˆ์ฃ .
07:05
Where are the defenders standing? What are their distances?
136
425280
2762
์ˆ˜๋น„์ˆ˜๊ฐ€ ์–ด๋”” ์„œ์žˆ๋Š”์ง€? ๊ทธ ๊ฑฐ๋ฆฌ๋Š” ์–ผ๋งˆ์ธ์ง€?
๊ทธ๋“ค๊ณผ์˜ ๊ฐ๋„๋Š” ์–ผ๋งˆ์ธ์ง€?
07:08
What are their angles?
137
428066
1331
07:09
For multiple defenders, we can look at how the player's moving
138
429421
2977
๋‹ค์ˆ˜์˜ ์ˆ˜๋น„์ˆ˜๊ฐ€ ์–ด๋–ป๊ฒŒ ์›€์ง์ด๋Š”์ง€ ๋ณผ ์ˆ˜ ์žˆ์—ˆ์ฃ .
07:12
and predict the shot type.
139
432422
1433
์Š› ํƒ€์ž…๋„ ์˜ˆ์ƒํ•  ์ˆ˜ ์žˆ์—ˆ๊ณ ์š”.
07:13
We can look at all their velocities and we can build a model that predicts
140
433879
4074
๊ณต์˜ ์†๋„๋ฅผ ๋ณด๊ณ  ์˜ˆ์ธก ๋ชจ๋ธ์„ ๋งŒ๋“ค ์ˆ˜ ์žˆ์—ˆ์ฃ .
07:17
what is the likelihood that this shot would go in under these circumstances?
141
437977
4052
์ด ์Š›์ด ์ด๋Ÿฐ ์ƒํ™ฉ์—์„œ ๋‚˜์˜ฌ ๊ฐ€๋Šฅ์„ฑ์ด ์–ผ๋งˆ๋‚˜ ๋ ๊นŒ?
07:22
So why is this important?
142
442188
1500
์ด๊ฒŒ ์™œ ์ค‘์š”ํ• ๊นŒ์š”?
07:24
We can take something that was shooting,
143
444102
2803
์Š›์„ ํ•ด์„ํ•  ๋•Œ
07:26
which was one thing before, and turn it into two things:
144
446929
2680
์ „์—๋Š” ํ•˜๋‚˜๋กœ ๋ณด์•˜๋˜ ๊ฒƒ์„ ์ง€๊ธˆ ๋‘˜๋กœ ๋‚˜๋ˆ„์–ด ํ•ด์„ํ•˜์ฃ .
07:29
the quality of the shot and the quality of the shooter.
145
449633
2651
์Š›์˜ ์œ ํšจ์œจ๊ณผ ์Šˆํ„ฐ์˜ ์ž์งˆ์ž…๋‹ˆ๋‹ค.
07:33
So here's a bubble chart, because what's TED without a bubble chart?
146
453680
3262
๋ฒ„๋ธ”์ฐจํŠธ๊ฐ€ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๊ฑฐ ์—†์ด๋Š” ํ…Œ๋“œ ๊ฐ•์—ฐ์„ ํ•  ์ˆ˜ ์—†์ฃ .
07:36
(Laughter)
147
456966
1014
(์›ƒ์Œ)
07:38
Those are NBA players.
148
458004
1311
NBA ์„ ์ˆ˜๋“ค์ž…๋‹ˆ๋‹ค.
07:39
The size is the size of the player and the color is the position.
149
459339
3120
ํฌ๊ธฐ๋Š” ์„ ์ˆ˜๋“ค์˜ ์ฒด๊ฒฉ์ด๊ณ  ์ƒ‰์€ ์„ ์ˆ˜๋“ค์˜ ํฌ์ง€์…˜์ž…๋‹ˆ๋‹ค.
07:42
On the x-axis, we have the shot probability.
150
462483
2132
X์ถ•์—๋Š” ์Š›์„ฑ๊ณต๋ฅ ์ž…๋‹ˆ๋‹ค.
07:44
People on the left take difficult shots,
151
464639
1953
์™ผ์ชฝ์— ์žˆ๋Š” ์„ ์ˆ˜๋Š” ์Š›์„ ์„ฑ๊ณตํ•˜๊ธฐ ํž˜๋“ค์ฃ .
07:46
on the right, they take easy shots.
152
466616
2229
์˜ค๋ฅธ์ชฝ์—๋Š” ์žˆ๋Š” ์„ ์ˆ˜๋Š” ์Š›์„ ์„ฑ๊ณตํ•˜๊ธฐ ์‰ฝ์ฃ .
07:49
On the [y-axis] is their shooting ability.
153
469194
2057
Y์ถ•์€ ์Š›์„ ์  ์ˆ˜ ์žˆ๋Š” ๋Šฅ๋ ฅ์ž…๋‹ˆ๋‹ค.
07:51
People who are good are at the top, bad at the bottom.
154
471275
2562
์œ„์—๋Š” ๋Šฅ๋ ฅ์ด ์ข‹์€ ์„ ์ˆ˜, ์•„๋ž˜๋Š” ๋Šฅ๋ ฅ์ด ๋‚˜์œ ์„ ์ˆ˜์ž…๋‹ˆ๋‹ค.
07:53
So for example, if there was a player
155
473861
1760
์˜ˆ๋ฅผ ๋“ค๋ฉด, ๋Œ€์ฒด๋กœ ์Š›์˜ 47%๋ฅผ ๋„ฃ๋Š” ์„ ์ˆ˜๊ฐ€ ์žˆ๋Š”๋ฐ
07:55
who generally made 47 percent of their shots,
156
475621
2097
07:57
that's all you knew before.
157
477718
1389
๊ทธ๊ฒŒ ์—ฌ๋Ÿฌ๋ถ„์ด ์•„๋Š” ์ „๋ถ€์ฃ .
07:59
But today, I can tell you that player takes shots that an average NBA player
158
479345
4850
ํ•˜์ง€๋งŒ ์˜ค๋Š˜ ๋ง์”€๋“œ๋ฆด ์ˆ˜ ์žˆ๋Š” ๊ฒƒ์€ NBA ํ‰๊ท  ์„ ์ˆ˜๋“ค์ด
08:04
would make 49 percent of the time,
159
484219
1961
๊ทธ ๋‹น์‹œ ์Š›์„ฑ๊ณต๋ฅ ์€ 49%์ด๊ณ 
08:06
and they are two percent worse.
160
486204
1684
์Š› ์˜๋Š” ๋Šฅ๋ ฅ์€ -2%๋ผ๋Š” ๊ฒƒ์ด์ฃ .
08:08
And the reason that's important is that there are lots of 47s out there.
161
488266
4515
์ค‘์š”ํ•œ ์‚ฌ์‹ค์€ 47%์ธ ์„ ์ˆ˜๊ฐ€ ๋งŽ๋‹ค๋Š” ๊ฑฐ์ฃ .
08:13
And so it's really important to know
162
493714
2549
์•Œ์•„์•ผ ํ•  ์ •๋ง ์ค‘์š”ํ•œ ๊ฒƒ์€
08:16
if the 47 that you're considering giving 100 million dollars to
163
496287
3956
1์–ต ๋‹ฌ๋Ÿฌ๋ฅผ ์ฃผ๊ธฐ๋กœ ํ•œ 47%์˜ ์„ ์ˆ˜๊ฐ€
08:20
is a good shooter who takes bad shots
164
500267
3055
์ข‹์€ ์Šˆํ„ฐ์ง€๋งŒ ์Š›์„ ์ž˜ ๋ชป ๋˜์ง€๋Š”์ง€
08:23
or a bad shooter who takes good shots.
165
503346
2397
๋‚˜์œ ์Šˆํ„ฐ์ง€๋งŒ ์Š›์„ ์ž˜ ๋˜์ง€๋Š”์ง€ ์•„๋Š” ๊ฑฐ์ฃ .
08:27
Machine understanding doesn't just change how we look at players,
166
507130
3333
๊ธฐ๊ณ„๊ฐ€ ์ดํ•ดํ•˜๊ณ  ์žˆ๋Š” ๊ฒƒ์ด ์„ ์ˆ˜๋ฅผ ๋ณด๋Š” ๋ฐฉ๋ฒ•์„ ๋ณ€ํ™”์‹œํ‚ค์ง€ ๋ชปํ•˜์ง€๋งŒ
08:30
it changes how we look at the game.
167
510487
1858
๊ฒฝ๊ธฐ๋ฅผ ๋ฐ”๋ผ๋ณด๋Š” ๋ฐฉ๋ฒ•์„ ๋ณ€ํ™”์‹œํ‚ต๋‹ˆ๋‹ค.
08:32
So there was this very exciting game a couple of years ago, in the NBA finals.
168
512369
3755
๊ฐ€์žฅ ํฅ๋ฏธ์ง„์ง„ํ–ˆ๋˜ ๋ช‡ ๋…„ ์ „ NBA ๊ฒฐ์Šน ๊ฒฝ๊ธฐ์—์„œ
๋งˆ์ด์• ๋ฏธ๋Š” 3์ ์„ ๋’ค์ง€๊ณ  20์ดˆ๊ฐ€ ๋‚จ์•˜์—ˆ์Šต๋‹ˆ๋‹ค.
08:36
Miami was down by three, there was 20 seconds left.
169
516148
3207
08:39
They were about to lose the championship.
170
519379
2025
์šฐ์Šน์„ ๋†“์น˜๊ธฐ ์ผ๋ณด ์ง์ „์ด์—ˆ์ฃ .
08:41
A gentleman named LeBron James came up and he took a three to tie.
171
521428
3341
๋ฅด๋ธŒ๋ก  ์ œ์ž„์Šค ์„ ์ˆ˜๊ฐ€ 3์ ์Š›์„ ์ด ๋™์ ์„ ๋งŒ๋“ค๋ ค๊ณ  ํ–ˆ์œผ๋‚˜
08:44
He missed.
172
524793
1198
์‹คํŒจํ–ˆ์ฃ .
๋™๋ฃŒ์„ ์ˆ˜ ํฌ๋ฆฌ์Šค ๋ณด์‰ฌ๊ฐ€ ๋ฆฌ๋ฐ”์šด๋“œ๋ฅผ ์žก์•˜๊ณ 
08:46
His teammate Chris Bosh got a rebound,
173
526015
1837
08:47
passed it to another teammate named Ray Allen.
174
527876
2159
๊ณต์„ ๋™๋ฃŒ์„ ์ˆ˜ ๋ ˆ์ด ์•Œ๋ Œ์—๊ฒŒ ํŒจ์Šคํ–ˆ์Šต๋‹ˆ๋‹ค.
3์ ์„ ์„ฑ๊ณตํ–ˆ๊ณ  ์—ฐ์žฅ์ „์œผ๋กœ ๋“ค์–ด๊ฐ”์ฃ .
08:50
He sank a three. It went into overtime.
175
530059
1919
๊ฒฝ๊ธฐ์— ์ด๊ฒผ๊ณ  ์ฑ”ํ”ผ์–ธ์‹ญ์—์„œ ์šฐ์Šนํ–ˆ์ฃ .
08:52
They won the game. They won the championship.
176
532002
2096
๊ฐ€์žฅ ํฅ๋ฏธ์ง„์ง„ํ–ˆ๋˜ ๋†๊ตฌ๊ฒฝ๊ธฐ ์ค‘ ํ•˜๋‚˜์˜€์Šต๋‹ˆ๋‹ค.
08:54
It was one of the most exciting games in basketball.
177
534122
2444
08:57
And our ability to know the shot probability for every player
178
537438
3429
๋งค ์ˆœ๊ฐ„ ๋ชจ๋“  ์„ ์ˆ˜์˜ ์Š›์„ฑ๊ณต๋ฅ ๊ณผ
09:00
at every second,
179
540891
1188
๋งค ์ˆœ๊ฐ„ ๋ฆฌ๋ฐ”์šด๋“œ๋ฅผ ๋”ฐ๋‚ผ ๊ฐ€๋Šฅ์„ฑ์„ ํŒŒ์•…ํ•˜๋Š” ๋Šฅ๋ ฅ์„
09:02
and the likelihood of them getting a rebound at every second
180
542103
2956
์ „์—๋Š” ๋ถˆ๊ฐ€๋Šฅํ–ˆ๋˜ ๋ฐฉ๋ฒ•์œผ๋กœ ์„ค๋ช…ํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋˜์—ˆ์ฃ .
09:05
can illuminate this moment in a way that we never could before.
181
545083
3443
09:09
Now unfortunately, I can't show you that video.
182
549618
2668
ํ˜„์žฌ, ์•ˆํƒ€๊น๊ฒŒ๋„ ์—ฌ๋Ÿฌ๋ถ„์—๊ฒŒ ๊ทธ ๋น„๋””์˜ค๋ฅผ ๋ณด์—ฌ์ค„ ์ˆ˜ ์—†๋„ค์š”.
09:12
But for you, we recreated that moment
183
552310
4493
ํ•˜์ง€๋งŒ, ์—ฌ๋Ÿฌ๋ถ„์„ ์œ„ํ•ด ๊ทธ ์ˆœ๊ฐ„์„ ์žฌํ˜„ํ–ˆ์Šต๋‹ˆ๋‹ค.
09:16
at our weekly basketball game about 3 weeks ago.
184
556827
2336
3์ฃผ ์ „ ์ฃผ๋ง ๋†๊ตฌ๊ฒฝ๊ธฐ์—์„œ ๋ง์ด์ฃ .
09:19
(Laughter)
185
559279
2167
(์›ƒ์Œ)
09:21
And we recreated the tracking that led to the insights.
186
561573
3410
์ง๊ด€์„ ๋ณด์—ฌ์ฃผ๋Š” ๊ฒฝ๊ธฐ๋ฅผ ์žฌํ˜„ํ–ˆ์ฃ .
09:25
So, here is us. This is Chinatown in Los Angeles,
187
565199
4255
์ด๊ณณ์ž…๋‹ˆ๋‹ค. ๋กœ์Šค์•ค์ ค๋ ˆ์Šค ์ฐจ์ด๋‚˜ํƒ€์šด์ด์ฃ .
09:29
a park we play at every week,
188
569478
1564
๋งค์ฃผ ๊ฒฝ๊ธฐ๋ฅผ ํ–ˆ๋˜ ๊ณต์›์ž…๋‹ˆ๋‹ค.
09:31
and that's us recreating the Ray Allen moment
189
571066
2231
๋ ˆ์ด ์•Œ๋ Œ์ด ํ”Œ๋ ˆ์ดํ–ˆ๋˜ ์ˆœ๊ฐ„๊ณผ
09:33
and all the tracking that's associated with it.
190
573321
2229
๊ทธ๊ฒƒ๊ณผ ๊ด€๋ จ๋œ ๋ชจ๋“  ๊ฒƒ์„ ์žฌํ˜„ํ–ˆ์Šต๋‹ˆ๋‹ค.
09:36
So, here's the shot.
191
576772
1517
์Š›์žฅ๋ฉด์ž…๋‹ˆ๋‹ค.
09:38
I'm going to show you that moment
192
578313
2516
๊ทธ ์ˆœ๊ฐ„์„ ์—ฌ๋Ÿฌ๋ถ„์—๊ฒŒ ๋ณด์—ฌ์ฃผ๋ ค๊ณ  ํ•ฉ๋‹ˆ๋‹ค.
09:40
and all the insights of that moment.
193
580853
2587
๊ทธ ์ˆœ๊ฐ„์˜ ์ง๊ด€๊นŒ์ง€์š”.
09:43
The only difference is, instead of the professional players, it's us,
194
583464
3730
์ฐจ์ด์ ์ด๋ผ๊ณค ์ „๋ฌธ ์„ ์ˆ˜๋“ค์ด ์•„๋‹ˆ๋ผ ์ œ ๋™๋ฃŒ๋ผ๋Š” ๊ฑฐ์ฃ .
09:47
and instead of a professional announcer, it's me.
195
587218
2618
์ „๋ฌธ์ ์ธ ์•„๋‚˜์šด์„œ๊ฐ€ ์•„๋‹ˆ๋ผ ์ œ๊ฐ€ ํ•˜๊ณ  ์žˆ๋‹ค๋Š” ๊ฑฐ์ฃ .
09:49
So, bear with me.
196
589860
1477
์ž ์‹œ ์ฐธ์•„์ฃผ์„ธ์š”.
09:53
Miami.
197
593153
1150
๋งˆ์ด์• ๋ฏธ
09:54
Down three.
198
594671
1150
3์ ์ด ์ง€๊ณ  ์žˆ๋Š” ์ƒํ™ฉ
09:56
Twenty seconds left.
199
596107
1150
20์ดˆ๊ฐ€ ๋‚จ์•˜๋„ค์š”.
09:59
Jeff brings up the ball.
200
599385
1198
์ œํ”„๊ฐ€ ๊ณต์„ ๋ชฐ๊ณ  ์˜ค๋„ค์š”.
10:02
Josh catches, puts up a three!
201
602656
1535
์กฐ์‰ฌ๊ฐ€ ๋ฐ›์•„ 3์  ์Š›!
10:04
[Calculating shot probability]
202
604631
1849
[์Š› ์„ฑ๊ณต๋ฅ  ๊ณ„์‚ฐ ์ค‘]
10:07
[Shot quality]
203
607278
1150
[์Š›์˜ ์œ ํšจ์œจ]
10:09
[Rebound probability]
204
609048
1785
[๋ฆฌ๋ฐ”์šด๋“œ ์„ฑ๊ณต๋ฅ ]
10:12
Won't go!
205
612373
1173
์žก์ง€ ๋ชปํ–ˆ๋„ค์š”.
10:13
[Rebound probability]
206
613570
1446
[๋ฆฌ๋ฐ”์šด๋“œ ์„ฑ๊ณต๋ฅ ]
10:15
Rebound, Noel.
207
615777
1256
๋…ธ์—˜์ด ๋ฆฌ๋ฐ”์šด๋“œํ–ˆ์Šต๋‹ˆ๋‹ค.
10:17
Back to Daria.
208
617057
1150
๋‹ค๋ฆฌ์•„์—๊ฒŒ ํŒจ์Šค
10:18
[Shot quality]
209
618509
3365
[์Š›์˜ ์œ ํšจ์œจ]
10:22
Her three-pointer -- bang!
210
622676
1620
3์  ์Š›, ๊ณจ!
10:24
Tie game with five seconds left.
211
624320
2197
5์ดˆ๋ฅผ ๋‚จ๊ธด ์ƒํ™ฉ
10:26
The crowd goes wild.
212
626880
1618
๊ด€์ค‘์ด ํฅ๋ถ„ํ•˜๊ธฐ ์‹œ์ž‘ํ•ฉ๋‹ˆ๋‹ค.
10:28
(Laughter)
213
628522
1659
(์›ƒ์Œ)
10:30
That's roughly how it happened.
214
630205
1547
๋Œ€์ถฉ ์ด๋Ÿฐ ์‹์œผ๋กœ ์ผ์–ด๋‚˜์ฃ .
10:31
(Applause)
215
631776
1151
(๋ฐ•์ˆ˜)
10:32
Roughly.
216
632951
1175
๋Œ€์ถฉ์ด์š”.
10:34
(Applause)
217
634150
1531
(๋ฐ•์ˆ˜)
10:36
That moment had about a nine percent chance of happening in the NBA
218
636121
5484
๊ทธ๋Ÿฐ ์ˆœ๊ฐ„์€ NBA์—์„œ ์ผ์–ด๋‚  ํ™•๋ฅ ์ด ์•ฝ 9%๋ผ๊ณ  ํ•˜์ฃ .
10:41
and we know that and a great many other things.
219
641629
2261
์•„์‹œ๋‹ค์‹œํ”ผ ๋‹ค๋ฅธ ๊ฒƒ๋“ค๋„ ๋งค์šฐ ๋งŽ์ฃ .
10:43
I'm not going to tell you how many times it took us to make that happen.
220
643914
3491
์–ผ๋งˆ๋‚˜ ๋งŽ์ด ์ผ์–ด๋‚˜๋Š”์ง€ ๋งํ•˜๋ ค๊ณ  ํ•˜๋Š” ๊ฒŒ ์•„๋‹™๋‹ˆ๋‹ค.
10:47
(Laughter)
221
647429
1747
(์›ƒ์Œ)
10:49
Okay, I will! It was four.
222
649200
1872
์ข‹์•„์š”. ๋งํ•ด์ค„๊ฒŒ์š”. ๋„ค ๋ฒˆ ์ •๋„์ฃ .
10:51
(Laughter)
223
651096
1001
(์›ƒ์Œ)
10:52
Way to go, Daria.
224
652121
1165
์ž˜ํ–ˆ์–ด. ๋‹ค๋ฆฌ์•„
10:53
But the important thing about that video
225
653647
4263
ํ•˜์ง€๋งŒ ๊ทธ ๋น„๋””์˜ค์—์„œ ์ค‘์š”ํ•œ ๊ฒƒ๊ณผ
10:57
and the insights we have for every second of every NBA game -- it's not that.
226
657934
4568
๋ชจ๋“  NBA๊ฒฝ๊ธฐ์˜ ๋งค ์ดˆ๋งˆ๋‹ค ํŒŒ์•…ํ•˜๋ ค๊ณ  ํ•œ ์ง๊ด€์€ ๊ทธ๊ฒŒ ์•„๋‹ˆ์ฃ .
11:02
It's the fact you don't have to be a professional team to track movement.
227
662639
3929
์‚ฌ์‹ค ์›€์ง์ž„์„ ์ถ”์ ํ•˜๊ธฐ ์œ„ํ•ด ์ „๋ฌธ์ ์ธ ํŒ€์ผ ํ•„์š”๋Š” ์—†์Šต๋‹ˆ๋‹ค.
11:07
You do not have to be a professional player to get insights about movement.
228
667083
3657
์›€์ง์ž„์— ๋Œ€ํ•œ ์•ˆ๋ชฉ์„ ํ‚ค์šฐ๋ ค๊ณ  ์ „๋ฌธ์ ์ธ ์„ ์ˆ˜์ผ ํ•„์š”๋Š” ์—†์ฃ .
11:10
In fact, it doesn't even have to be about sports because we're moving everywhere.
229
670764
3858
์‚ฌ์‹ค ๊ตณ์ด ์Šคํฌ์ธ ์ผ ํ•„์š”๋Š” ์—†์ฃ . ์šฐ๋ฆฌ๋Š” ๋ชจ๋“  ์ˆœ๊ฐ„ ์›€์ง์ด๋‹ˆ๊นŒ์š”.
11:15
We're moving in our homes,
230
675654
2369
์ง‘์— ๊ฐ€๊ฑฐ๋‚˜
11:21
in our offices,
231
681428
1205
ํšŒ์‚ฌ์— ๊ฐ€๊ฑฐ๋‚˜
11:24
as we shop and we travel
232
684238
2690
์‡ผํ•‘์„ ํ•˜๊ณ 
์—ฌํ–‰ํ•˜์ฃ .
11:29
throughout our cities
233
689318
1253
๋„์‹œ ์—ฌ๊ธฐ ์ €๊ธฐ๋ฅผ
11:32
and around our world.
234
692065
1618
๊ทธ๋ฆฌ๊ณ  ์ „ ์„ธ๊ณ„๋ฅผ
11:35
What will we know? What will we learn?
235
695270
2295
์šฐ๋ฆฌ๋Š” ๋ฌด์—‡์„ ์•Œ๊ณ  ๋ฐฐ์› ์„๊นŒ์š”?
11:37
Perhaps, instead of identifying pick-and-rolls,
236
697589
2305
์•„๋งˆ ํ”ฝ์•ค๋กค์„ ๊ฐ์ง€ํ•˜๋Š” ๋Œ€์‹ 
11:39
a machine can identify the moment and let me know
237
699918
3010
๋”ธ์˜ ์ฒซ๋ฐœ์„ ๋‚ด๋”›๋Š” ์ˆœ๊ฐ„์„ ๊ฐ์ง€ํ•ด ์•Œ๋ ค์ค„ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
11:42
when my daughter takes her first steps.
238
702952
2059
11:45
Which could literally be happening any second now.
239
705035
2536
๋ง ๊ทธ๋Œ€๋กœ ํ˜„์žฌ ์ผ์–ด๋‚˜๊ณ  ์žˆ๋Š” ๋ชจ๋“  ์ˆœ๊ฐ„์„ ๋ง์ด์ฃ .
11:48
Perhaps we can learn to better use our buildings, better plan our cities.
240
708140
3697
์•„๋งˆ ๊ฑด๋ฌผ์„ ๋” ์ž˜ ์ด์šฉํ•˜๊ณ 
๋„์‹œ ์„ค๊ณ„๋ฅผ ๋” ์ž˜ ํ•  ์ˆ˜ ์žˆ์„ ๊ฒ๋‹ˆ๋‹ค.
11:52
I believe that with the development of the science of moving dots,
241
712362
4173
์›€์ง์ด๋Š” ์ ์˜ ๊ณผํ•™ ๋ฐœ์ „์„ ํ†ตํ•ด
11:56
we will move better, we will move smarter, we will move forward.
242
716559
3643
์šฐ๋ฆฌ๋Š” ๋” ์ข‹์€ ์ชฝ์œผ๋กœ ๋” ์˜๋ฆฌํ•˜๊ฒŒ, ์•ž์œผ๋กœ ๋‚˜์•„๊ฐˆ ๊ฑฐ๋ผ ๋ฏฟ์Šต๋‹ˆ๋‹ค
12:00
Thank you very much.
243
720607
1189
๋Œ€๋‹จํžˆ ๊ฐ์‚ฌํ•ฉ๋‹ˆ๋‹ค.
12:01
(Applause)
244
721820
5045
(๋ฐ•์ˆ˜)
์ด ์›น์‚ฌ์ดํŠธ ์ •๋ณด

์ด ์‚ฌ์ดํŠธ๋Š” ์˜์–ด ํ•™์Šต์— ์œ ์šฉํ•œ YouTube ๋™์˜์ƒ์„ ์†Œ๊ฐœํ•ฉ๋‹ˆ๋‹ค. ์ „ ์„ธ๊ณ„ ์ตœ๊ณ ์˜ ์„ ์ƒ๋‹˜๋“ค์ด ๊ฐ€๋ฅด์น˜๋Š” ์˜์–ด ์ˆ˜์—…์„ ๋ณด๊ฒŒ ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๊ฐ ๋™์˜์ƒ ํŽ˜์ด์ง€์— ํ‘œ์‹œ๋˜๋Š” ์˜์–ด ์ž๋ง‰์„ ๋”๋ธ” ํด๋ฆญํ•˜๋ฉด ๊ทธ๊ณณ์—์„œ ๋™์˜์ƒ์ด ์žฌ์ƒ๋ฉ๋‹ˆ๋‹ค. ๋น„๋””์˜ค ์žฌ์ƒ์— ๋งž์ถฐ ์ž๋ง‰์ด ์Šคํฌ๋กค๋ฉ๋‹ˆ๋‹ค. ์˜๊ฒฌ์ด๋‚˜ ์š”์ฒญ์ด ์žˆ๋Š” ๊ฒฝ์šฐ ์ด ๋ฌธ์˜ ์–‘์‹์„ ์‚ฌ์šฉํ•˜์—ฌ ๋ฌธ์˜ํ•˜์‹ญ์‹œ์˜ค.

https://forms.gle/WvT1wiN1qDtmnspy7