00:00 do you bring this trick out at parties
00:03 oh no it's a terrible party trick here
00:06 three point one four one five nine two
00:08 six five three five eight nine seven
00:09 eight this is grant gusman he watched an
00:12 old video of mine about how we think
00:14 that there are two systems of thought
00:16 system two is the conscious slow
00:18 effortful system and system one is
00:20 subconscious fast and automatic
00:23 to explore how these systems work in his
00:25 own head grant decided to memorize a
00:27 hundred digits of pi
00:29 three four four six then he just kept
00:32 going he has now memorized twenty three
00:34 thousand digits of pie in preparation to
00:37 challenge the north american record five
00:39 four nine three zero three eight one
00:40 nine six that's two hundred
00:47 i have wanted to make a video about
00:49 experts for a long time
00:56 this is magnus carlson the five-time
00:58 world chess champion he's being shown
01:01 chessboards and asked to identify the
01:03 game in which they occurred uh this
01:05 looks an awful lot like
01:19 now i'm going to play through an opening
01:22 stop me when you recognize the game and
01:25 if you can tell me who was playing black
01:32 i'm sure you've seen this opening before
01:34 okay it's going to be arnold
01:40 how can he do this it seems like
01:42 superhuman ability well decades ago
01:45 scientists wanted to know what makes
01:47 experts like chess masters special do
01:49 they have incredibly high iqs much
01:52 better spatial reasoning than average
01:54 bigger short-term memory spans well it
01:56 turns out that as a group chess masters
01:59 are not exceptional on any of these
02:02 but one experiment showed how their
02:04 performance was vastly superior to
02:06 amateurs in 1973 william chase and
02:10 herbert simon recruited three chess
02:12 players a master an a player who's an
02:14 advanced amateur and a beginner
02:17 a chess board was set up with around 25
02:19 pieces positioned as they might be
02:21 during a game and each player was
02:23 allowed to look at the board for five
02:26 then they were asked to replicate the
02:27 setup from memory on a second board in
02:31 the players could take as many
02:32 five-second peaks as they needed to get
02:34 their board to match from just the first
02:37 look the master could recall the
02:39 positions of 16 pieces the a player
02:41 could recall eight and the beginner only
02:45 the master only needed half the number
02:47 of peaks as the a player to get their
02:50 but then the researchers arranged the
02:52 board with pieces in random positions
02:54 that would never arise in a real game
02:57 and now the chess master performed no
02:59 better than the beginner
03:01 after the first look all players
03:02 regardless of rank could remember the
03:04 location of only three pieces the data
03:07 are clear chess experts don't have
03:09 better memory in general but they have
03:11 better memory specifically for chess
03:13 positions that could occur in a real
03:15 game the implication is what makes chess
03:17 masters special is that they have seen
03:19 lots and lots of chess games and over
03:22 that time their brains have learned
03:23 patterns so rather than seeing
03:25 individual pieces at individual
03:27 positions they see a smaller number of
03:29 recognizable configurations
03:32 this is called chunking what we have
03:34 stored in long-term memory allows us to
03:36 recognize complex stimuli as just one
03:39 thing for example you recognize this as
03:42 pi rather than a string of six unrelated
03:44 numbers or meaningless squiggles for
03:47 there's a wonderful sequence i like a
03:52 which to me means stephen curry number
03:56 173 games which is the record back in
04:01 at its core expertise is about
04:05 magnus carlson recognizes chess
04:07 positions the same way we recognize
04:10 and recognition leads directly to
04:12 intuition if you see an angry face you
04:16 have a pretty good idea of what's going
04:19 chess masters recognize board positions
04:21 and instinctively know the best move
04:25 i know what to do i don't have to
04:32 to develop the long-term memory of an
04:34 expert takes a long time ten thousand
04:37 hours is the rule of thumb popularized
04:39 by malcolm gladwell but ten 000 hours of
04:41 practice by itself is not sufficient
04:44 there are four additional criteria that
04:48 and in areas where these criteria aren't
04:50 met it's impossible to become an expert
04:54 so the first one is many repeated
04:56 attempts with feedback
04:57 tennis players hit hundreds of four
04:59 hands in practice chess players play
05:02 thousands of games before their
05:03 grandmasters and physicists solve
05:06 thousands of physics problems
05:08 each one gets feedback the tennis player
05:10 sees whether each shot clears the net
05:12 and is in or out the chess player either
05:14 wins or loses the game and the physicist
05:17 gets the problem right or wrong
05:19 but some professionals don't get
05:21 repeated experience with the same sorts
05:23 of problems political scientist philip
05:25 tetlock picked 284 people who make their
05:28 living commenting or offering advice on
05:30 political and economic trends this
05:32 included journalists foreign policy
05:34 specialists economists and intelligence
05:36 analysts over two decades he peppered
05:39 them with questions like would george
05:41 bush be re-elected would apartheid in
05:43 south africa end peacefully would quebec
05:46 secede from canada and would the dot-com
05:50 in each case the pundits rated the
05:52 probability of several possible outcomes
05:54 and by the end of the study tetlock had
06:02 pretty terribly these experts most of
06:05 whom had postgraduate degrees performed
06:08 worse than if they had just assigned
06:09 equal probabilities to all the outcomes
06:12 in other words people who spend their
06:14 time and earn their living studying a
06:15 particular topic produce poorer
06:17 predictions than random chance even in
06:20 the areas they knew best experts were
06:22 not significantly better than
06:23 non-specialists the problem is most of
06:26 the events they have to predict are
06:28 one-offs they haven't had the experience
06:30 of going through these events or very
06:31 similar ones many times before
06:33 even presidential elections only happen
06:35 infrequently and each one in a slightly
06:38 different environment so we should be
06:40 wary of experts who don't have repeated
06:43 experience with feedback
06:46 the next requirement is a valid
06:48 environment one that contains
06:49 regularities that make it at least
06:51 somewhat predictable a gambler betting
06:54 at the roulette wheel for example may
06:55 have thousands of repeated experiences
06:57 with the same event and for each one
06:59 they get clear feedback in the form of
07:01 whether they win or lose but you would
07:03 rightfully not consider them an expert
07:05 because the environment is low validity
07:08 a roulette wheel is essentially random
07:10 so there are no regularities to be
07:13 in 2006 legendary investor warren
07:16 buffett offered to bet a million dollars
07:18 that he could pick an investment that
07:20 would outperform wall street's best
07:21 hedge funds over a 10-year period hedge
07:24 funds are pools of money that are
07:25 actively managed by some of the
07:26 brightest and most experienced traders
07:28 on wall street they use advanced
07:30 techniques like short selling leverage
07:32 and derivatives in an attempt to provide
07:34 outsized returns and consequently they
07:37 charge significant fees
07:39 one person took buffett up on the bet
07:41 ted cites of protege partners for his
07:44 investment he selected five hedge funds
07:46 well actually five funds of hedge funds
07:49 so in total a collection of over 200
07:53 warren buffett took a very different
07:55 approach he picked the most basic boring
07:58 investment imaginable a passive index
08:00 fund that just tracks the weighted value
08:02 of the 500 biggest public companies in
08:04 america the s p 500
08:06 they started the bet on january 1st 2008
08:09 and immediately things did not look good
08:11 for buffett it was the start of the
08:13 global financial crisis and the market
08:15 tanked but the hedge funds could change
08:18 their holdings and even profit from
08:20 market falls so they lost some value but
08:22 not as much as the market average
08:25 the hedge funds stayed ahead for the
08:26 next three years but by 2011 the s p 500
08:29 had pulled even and from then on it
08:32 wasn't even close the market average
08:34 surged leaving the hedge funds in the
08:36 dust after 10 years buffett's index fund
08:39 125.8 percent to the hedge funds 36
08:44 now the market performance was not
08:46 unusual over this time at eight and a
08:48 half percent annual growth it nearly
08:50 matches the stock market's long run
08:51 average so why did so many investment
08:54 professionals with years of industry
08:56 experience research at their fingertips
08:57 and big financial incentives to perform
09:00 failed to beat the market well because
09:02 stocks are a low validity environment
09:04 over the short term stock price
09:06 movements are almost entirely random
09:08 so the feedback although clear and
09:10 immediate doesn't actually reflect
09:11 anything about the quality of the
09:15 it's closer to a roulette wheel than to
09:19 over a 10-year period around 80 percent
09:21 of all actively managed investment funds
09:24 fail to beat the market average and if
09:26 you look at longer time periods
09:27 underperformance rises to 90 percent and
09:30 before you say well that means 10 of
09:32 managers have actual skill consider that
09:35 just through random chance some people
09:36 would beat the market anyway portfolios
09:38 picked by cats or throwing darts have
09:41 been shown to do just that and in
09:43 addition to luck there are nefarious
09:45 practices from insider trading to pump
09:48 now i don't mean to say there are no
09:50 expert investors i mean warren buffett
09:51 himself is a clear example but the vast
09:54 majority of stock pickers and active
09:56 investment managers do not demonstrate
09:58 expert performance because of the low
10:01 validity of their environment
10:03 brief side note if we know that stock
10:05 picking will usually yield worse results
10:07 over the long term and that what active
10:10 managers charge and fees is rarely
10:11 compensated for in improved performance
10:14 then why is so much money invested in
10:16 individual stocks mutual funds and hedge
10:20 well let me answer that with a story
10:22 there was an experiment carried out with
10:24 rats and humans where there's a red
10:25 button and a green button that can each
10:29 eighty percent of the time the green
10:30 button lights up and twenty percent of
10:32 the time the red button lights up but
10:34 randomly so you can never be sure which
10:37 button will light and the task for the
10:39 subject either rat or human is to guess
10:41 beforehand which button will light up by
10:43 pressing it for the rap if they guess
10:45 right they get a bit of food and if they
10:47 guess wrong a mild electric shock
10:49 the rat quickly learns to press only the
10:52 green button and accept the 80 win
10:54 percentage humans on the other hand
10:57 usually press the green button but once
10:59 in a while they try to predict when the
11:01 red light will go on and as a result
11:03 they guess right only 68 percent of the
11:06 we have a hard time accepting average
11:07 results and we see patterns everywhere
11:09 including in randomness so we try to
11:12 beat the average by predicting the
11:13 pattern but when there is no pattern
11:16 this is a terrible strategy
11:18 even when there are patterns you need
11:20 timely feedback in order to learn them
11:23 and youtube knows this which is why
11:24 within the first hour after posting a
11:26 video they tell you how its performance
11:28 compares to your last 10 videos there's
11:31 even confetti fireworks when the video
11:33 is number one i know it seems like a
11:35 silly thing but you have no idea how
11:37 powerful a reward this is and how much
11:39 youtuber effort is spent chasing this
11:41 supercharged dopamine hit
11:44 to understand the difference between
11:45 immediate and delayed feedback
11:47 psychologist daniel kahneman contrasts
11:49 the experiences of anesthesiologists and
11:51 radiologists anesthesiologists work
11:54 alongside the patient and get feedback
11:56 straight away is the patient unconscious
11:58 with stable vital signs with this
12:00 immediate feedback it's easier for them
12:02 to learn the regularities of their
12:04 environment radiologists on the other
12:06 hand don't get rapid feedback on their
12:08 diagnoses if they get it at all this
12:10 makes it much harder for them to improve
12:12 radiologists typically correctly
12:14 diagnose breast cancer from x-rays just
12:16 70 percent of the time
12:19 delayed feedback also seems to be a
12:20 problem for college admissions officers
12:22 and recruitment specialists
12:24 after admitting someone to college or
12:26 hiring someone at a big company you may
12:28 never or only much later find out how
12:30 they did this makes it harder to
12:32 recognize the patterns in ideal
12:34 candidates in one study richard melton
12:36 tried to predict the grades of freshmen
12:38 at the end of their first year of
12:39 college a set of 14 counselors
12:42 interviewed each student for 45 minutes
12:44 to an hour they also had access to high
12:46 school grades several aptitude tests and
12:48 a four-page personal statement for
12:51 comparison melton created an algorithm
12:53 that used as input only a fraction of
12:56 the information just high school grades
12:58 and one aptitude test nevertheless the
13:00 formula was more accurate than 11 of the
13:05 study was reported alongside over a
13:07 dozen similar results across a variety
13:09 of other domains from predicting who
13:11 would violate parole to who'd succeed in
13:13 pilot training if you've ever been
13:15 denied admission to an educational
13:17 institution or turned down for a job it
13:19 feels like an expert has considered your
13:21 potential and decided that you don't
13:23 have what it takes to succeed you know i
13:25 was rejected twice from film school and
13:27 twice from a drama program so it's
13:29 comforting to know that the gatekeepers
13:31 at these institutions aren't great
13:32 predictors of future success
13:34 so if you're in a valid environment and
13:36 you get repeated experience with the
13:38 same events with clear timely feedback
13:40 for each attempt will you definitely
13:42 become an expert in 10 000 hours or so
13:45 the answer unfortunately is no because
13:47 most of us want to be comfortable
13:50 for a lot of tasks in life we can become
13:52 competent in a fairly short period of
13:53 time take driving a car for example
13:56 initially it's pretty challenging it
13:57 takes up all of system 2 but after 50
14:00 hours or so it becomes automatic system
14:02 one takes over and you can do it without
14:04 much conscious thought after that more
14:06 time spent driving doesn't improve
14:09 performance if you wanted to keep
14:10 improving you would have to try driving
14:12 in challenging situations like new
14:14 terrain higher speeds or in difficult
14:16 weather now i have played guitar for 25
14:18 years but i'm not an expert because i
14:20 usually play the same songs it's easier
14:22 and more fun but in order to learn you
14:25 have to be practicing at the edge of
14:26 your ability pushing beyond your comfort
14:29 zone you have to use a lot of
14:30 concentration and methodically
14:32 repeatedly attempt things you aren't
14:36 practice everything exactly as it is and
14:38 exactly as it's written
14:41 but at just such a speed that you have
14:44 to think about and know exactly where
14:46 you are and what your fingers are doing
14:47 and what it feels like this is known as
14:50 deliberate practice and in many areas
14:52 professionals don't engage in deliberate
14:54 practice so their performance doesn't
14:55 improve in fact sometimes it declines
14:58 if you're experiencing chest pain and
15:00 you walk into a hospital would you
15:01 rather the doctor is a recent graduate
15:04 or someone with 20 years experience
15:06 researchers have found that diagnostic
15:08 skills of medical students increase with
15:09 their time in medical school which makes
15:11 sense the more cases you've seen with
15:13 feedback the better you are at spotting
15:14 patterns but this only works up to a
15:16 point when it comes to rare diseases of
15:19 the heart or lungs doctors with 20 years
15:21 experience were actually worse at
15:23 diagnosing them than recent graduates
15:25 and that's because they haven't thought
15:26 about those rare diseases in a long time
15:28 so they're less able to recognize the
15:30 symptoms only after a refresher course
15:32 could the doctors accurately diagnose
15:36 and you can see the same effect in chess
15:38 the best predictor of skill level is not
15:40 the number of games or tournaments
15:41 played but the number of hours dedicated
15:44 to serious solitary study players spend
15:47 thousands of hours alone learning chess
15:49 theory studying their own games and
15:51 those of others and they play through
15:52 compositions which are puzzles designed
15:54 to help you recognize tactical patterns
15:57 in chess as in other areas it can be
15:59 challenging to force yourself to
16:00 practice deliberately and this is why
16:02 coaches and teachers are so valuable
16:05 they can recognize your weaknesses and
16:06 assign tasks to address them to become
16:09 an expert you have to practice for
16:11 thousands of hours in the uncomfortable
16:13 zone attempting the things you can't do
16:17 true expertise is amazing to watch to me
16:20 it looks like magic but it isn't at its
16:22 core expertise is recognition and
16:25 recognition comes from the incredible
16:27 amount of highly structured information
16:28 stored in long-term memory to build that
16:31 memory requires four things a valid
16:33 environment many repetitions timely
16:36 feedback and thousands of hours of
16:38 deliberate practice when those criteria
16:40 are met human performance is astonishing
16:44 and when it's not you get people we
16:46 think of as experts who actually aren't
16:53 if you want to become a stem expert you
16:56 have to actively interact with problems
16:58 that's what you can do with brilliant
17:00 the sponsor of this video check out this
17:02 course on computer science where you can
17:03 uncover the optimal strategy for finding
17:05 a key in a room and you quickly learn
17:08 how your own strategy can be replicated
17:10 in a neural network logic is another
17:12 great course that i find challenges me
17:14 mentally you go from thinking you
17:15 understand something to actually getting
17:17 it and if it feels difficult that's a
17:19 good thing it means you're getting
17:20 pushed outside your comfort zone this is
17:23 how brilliant facilitates deliberate
17:25 practice and if you ever get stuck a
17:27 helpful hint is always close at hand so
17:29 don't fall into the trap of just getting
17:31 comfortable and doing what you know how
17:32 to do build in the habit of being
17:35 uncomfortable and regularly learning
17:37 something new that is the way to
17:39 lifelong learning and growth so i invite
17:41 you to check out the courses over at
17:42 brilliant.org veritasium and i bet you
17:45 will find something there that you want
17:46 to learn plus if you click through right
17:48 now brilliant are offering 20 off an
17:50 annual premium subscription to the first
17:52 200 people to sign up so i want to thank
17:54 brilliant for supporting veritasium and
17:56 i want to thank you for watching
![Welp, Here's what the Fed *Just* Said [Fed FOMC Minutes]](/_next/image?url=https%3A%2F%2Fi.ytimg.com%2Fvi%2Fl8u23oq1E9E%2Fhqdefault.jpg&w=3840&q=75)
