00:00:00
we're really excited to have our next
00:00:01
speaker here nate silver is the
00:00:03
editor-in-chief of 538.
00:00:05
um and he uh
00:00:07
uh came to our poker game recently oh
00:00:10
nate came to our game come on out nate
00:00:12
ran the game over
00:00:19
who was the big who's the big loser that
00:00:21
night i think wait who is the big loser
00:00:23
because nate was the big winner right
00:00:24
well i he was the big winner i was
00:00:27
basically trying i think
00:00:29
i think i donated a little bit to uh
00:00:31
nate silver's new uh model y
00:00:36
i think you got the model s coming to
00:00:37
the game um this was the largest poker
00:00:41
win you've had in your life largest cash
00:00:43
game win yeah by
00:00:45
a factor of five
00:00:47
two two okay yeah uh well we're happy to
00:00:49
donate um you're getting by the way your
00:00:52
speaker fee i just canceled it okay um
00:00:54
but uh ladies and gentlemen uh nate
00:00:56
silver on the mind of a gambler cool
00:01:03
[Music]
00:01:05
rain
00:01:06
[Music]
00:01:09
man we open source
00:01:13
[Music]
00:01:17
i'm really excited to talk to you all
00:01:18
today um i'm working on a book
00:01:21
about gambling risk and rationality for
00:01:24
the term critical thinking to
00:01:25
rationality um
00:01:27
it's partly involved talking to people
00:01:28
about global societal risks so talking
00:01:31
to people about the risk of nuclear war
00:01:33
which is like not very cheery
00:01:34
conversations by the way
00:01:37
but also people who take risk for a
00:01:39
living or at the very least manage risk
00:01:42
and have real skin in the game they're
00:01:43
not just some actuary getting a salary
00:01:45
they're they're assessing risk and
00:01:46
making big decisions financially mostly
00:01:49
based on on that assessment um
00:01:52
so i've started out talking to poker
00:01:53
players and sports bettors i'm a former
00:01:56
poke player myself i still play
00:01:57
recreation a lot i'm currently um
00:02:00
i think number 301 in the global poker
00:02:02
rankings not that i look that up every
00:02:03
morning
00:02:04
uh
00:02:06
but mostly it's the examples of talking
00:02:07
to other
00:02:08
fantastic poker players and sports
00:02:10
bettors about their process and i'm also
00:02:12
talking now more
00:02:13
two vcs
00:02:15
two founders to hedge fund managers for
00:02:17
example um
00:02:18
there are a lot of similarities there's
00:02:20
a certain type of person that's been
00:02:22
talking about today
00:02:23
personality traits modes of thinking it
00:02:26
even affects maybe their political views
00:02:27
a little bit and you see a lot of things
00:02:29
in common
00:02:31
so i gave you someone's context already
00:02:33
this is a work in progress i mean the
00:02:34
book is kind of halfway
00:02:36
zero percent written about 50 research
00:02:39
so things may change i'm evolving and
00:02:40
learning and correcting things um
00:02:43
but it's fun to write a book it's fun to
00:02:45
have great conversations with the
00:02:47
besties and other people and really kind
00:02:48
of experience this firsthand so i'm
00:02:50
going to start with some things that i
00:02:51
think are relatively straightforward and
00:02:54
obvious and then get into things that
00:02:55
are maybe less so toward the end but one
00:02:57
really obvious one is that successful
00:02:59
gamblers think probabilistically it's
00:03:01
not just a matter of understanding
00:03:03
probability but being
00:03:05
comfortable with uncertainty being
00:03:07
willing to take bets on an uncertain
00:03:10
basis that's kind of table sticks for
00:03:12
even getting into
00:03:13
any field involving risk
00:03:16
in poker at least you probably also need
00:03:18
a fair amount of mathematical ability
00:03:20
there are now these things called
00:03:22
solvers in poker that come up with a
00:03:24
game theory optimal strategy for every
00:03:26
situation
00:03:27
it's a very complex situation it's way
00:03:29
too hard to memorize you have to kind of
00:03:31
develop a mathematical intuition behind
00:03:33
it um but even players with rare
00:03:36
exception phil hellmuth might be the one
00:03:38
exception actually right
00:03:40
but even players who are known as being
00:03:42
field players or exploitive players
00:03:44
people who rely on tells and psychology
00:03:47
and table talk i talked to daniel
00:03:49
negroni for example one of the probably
00:03:51
five best book players of all time and
00:03:52
he told me that about four years ago he
00:03:54
realized that these solver kids as he
00:03:56
calls them the math quizzes were were
00:03:58
better than him at least at certain
00:04:00
forms of poker and he totally remade his
00:04:02
game to look at computer solutions to be
00:04:04
more studious and mathematical and it's
00:04:06
improved his results quite a bit
00:04:09
um in the vc field i don't know you if
00:04:11
you're necessarily doing as much
00:04:12
modeling as i might in an election model
00:04:14
for example but at the very least people
00:04:16
understand the intuition and the
00:04:18
intuition is
00:04:19
you need both but i think the intuition
00:04:21
is more important than the application
00:04:24
number two this one ought to be really
00:04:26
obvious um but it needs to be said
00:04:29
successful gamblers are highly
00:04:30
competitive things that you hear from
00:04:32
people
00:04:33
over and over again is i want to kind of
00:04:35
prove people wrong i want to outsmart
00:04:38
the competition
00:04:40
i want to show who's best it's almost
00:04:41
more this motivation than financial
00:04:44
motivation i think
00:04:45
um it's partly a matter of you can't
00:04:47
settle for being average as a gambler
00:04:50
the average poker player loses money to
00:04:51
the house so does the average sport's
00:04:53
better so therefore you have to deviate
00:04:55
from average in some ways um
00:04:58
but kind of one conclusion i've come to
00:04:59
talking to people for this book is that
00:05:01
human beings are
00:05:02
are a competitive species right to get
00:05:05
people to get along in society is pretty
00:05:07
hard you know one good thing about
00:05:09
capitalism is that you have some managed
00:05:11
type of competition democracy is kind of
00:05:13
a type of competition and that's
00:05:14
probably a better way to kind of
00:05:16
challenge these conflicting urges we
00:05:18
have um than things that are too
00:05:20
centrally planned but certainly people
00:05:22
in this room people i talk to are on the
00:05:24
high end of the competitive scale
00:05:28
successful poker players and sports
00:05:29
bettors have to care a lot about small
00:05:31
edges it's the nature of placing bets in
00:05:34
a field where you lose money on average
00:05:35
so
00:05:36
a very good sports bettor might win 55
00:05:39
or so percent of his or her bets um
00:05:42
you have to win 52 and a half percent to
00:05:44
make up for the house cut for the rake
00:05:46
or the vig as different names um that
00:05:49
means you can't afford to make one
00:05:52
percent errors right you really have to
00:05:53
squeeze every last drop of expected
00:05:56
value or ev out of any situation that
00:05:58
you're in
00:06:00
this makes people very very very
00:06:02
detail-oriented poker players tend to be
00:06:03
very studious about their craft they
00:06:05
really kind of love poker you can't just
00:06:08
get the big things right in poker you
00:06:09
have to get the small things right too
00:06:11
it's too competitive a field right now
00:06:14
this might not be as true by the way
00:06:17
of vc
00:06:18
the average investment makes money it
00:06:20
may not make a above market return but
00:06:22
it'll make money which is not true for
00:06:24
the average poker hand that you play um
00:06:27
by the way casinos too if you talk to
00:06:28
casino executives you might think you
00:06:30
walk into a casino and everything that
00:06:32
you do is maximize them to make the
00:06:34
maximum amount of revenue possible
00:06:36
not really right they're pretty good at
00:06:37
the big stuff they protect their
00:06:39
downside risk a lot um but they're not
00:06:42
maximized as much as you think because
00:06:43
when you have like a license to make
00:06:44
money you don't have to get the small
00:06:46
stuff right um but in poker where only
00:06:49
10 of players probably make money in the
00:06:51
long run you have to be very very
00:06:52
detail-oriented
00:06:55
number four you see here a a polygram
00:06:58
thread about elon actually
00:07:00
successful gamblers are contrarian at
00:07:02
least to some degree and again i'm kind
00:07:04
of repeating myself a little bit but
00:07:05
there's there's no edge if someone if
00:07:07
everyone does the same thing in the same
00:07:10
way you have to have some angle that you
00:07:12
think makes you better than what the
00:07:13
market thinks and the market's pretty
00:07:15
good so it's a hard thing to manage
00:07:18
i think risk takers tend to view
00:07:19
themselves as outsiders where they are
00:07:22
or not is probably an open question
00:07:25
but they think of themselves as thinking
00:07:26
differently from the rest of the world
00:07:28
potentially this can also make them
00:07:30
potentially like a little bit annoying
00:07:32
like outside of investing and gambling
00:07:34
context right where if you're kind of
00:07:36
always the person who says well actually
00:07:38
um that can get kind of pushed back in
00:07:40
social settings for example or on
00:07:41
twitter
00:07:43
but you can't kind of settle for having
00:07:44
the average view because the average
00:07:46
view doesn't do better than than the
00:07:47
market
00:07:48
um there is a fine line between being
00:07:51
ahead of the curve
00:07:52
and a permanent contrarian um
00:07:55
in sports betting there's a line that's
00:07:56
constantly mispriced then i can just bet
00:07:58
that i don't want that
00:08:00
market to correct itself i want to keep
00:08:02
exploiting that error in pricing right
00:08:04
um
00:08:05
in vc i mean you might want to find a
00:08:07
founder or a sector or a company that's
00:08:09
undervalued by the market but eventually
00:08:11
you want
00:08:12
that sector to grow you want to attract
00:08:15
more talent you want to attract future
00:08:16
fundraising rounds so you kind of are
00:08:17
more being
00:08:19
ahead of the curve by
00:08:21
by half a year to a year
00:08:22
i think in some ways it's analogous to
00:08:25
real estate we're trying to kind of
00:08:26
anticipate where the market will be in
00:08:28
the future and that's a subtly different
00:08:30
skill than being maybe contrarian per se
00:08:32
um but related at least
00:08:35
number five and again this might be a
00:08:37
little bit redundant but successful
00:08:38
gamblers take shots and this is very
00:08:41
intuitive in vc where you have a field
00:08:43
where you might um might only have a
00:08:46
unicorn or decacorn you know one percent
00:08:48
of the time attempts at the time of
00:08:50
course you have a whole portfolio it may
00:08:51
not be that risky in the aggregate
00:08:54
but you have to be willing to make
00:08:55
decisions under uh under pressure and be
00:08:58
somewhat fearless as a result right
00:09:01
timing is important you want to be first
00:09:03
to market in poker you have a finite
00:09:05
amount of time you can spend on a hand
00:09:07
in practice
00:09:08
um sometimes people take these risks in
00:09:10
a calculated way right i talked to um
00:09:13
sam beckman fried for the book for
00:09:14
example and he's like yeah i knew this
00:09:16
would have like a only a 50 chance of
00:09:18
succeeding but the payoff was so high i
00:09:20
would take it some people just kind of
00:09:21
more intuitively just kind of want to
00:09:23
put skin in the game want to take risks
00:09:25
get some excitement or pleasure from it
00:09:27
again we're talking about the very
00:09:28
extreme
00:09:29
tales here too right um the people who
00:09:33
are super uber successful maybe are
00:09:34
almost taking on more risk than they
00:09:36
maybe should you can instead like live a
00:09:38
very happy and complacent life as an
00:09:40
investment banker or something but
00:09:41
people who are true entrepreneurs that
00:09:43
think at scale and think and think very
00:09:45
very big
00:09:47
i should say one challenge of writing
00:09:48
the book is that talking only to
00:09:50
successful people you risk survivorship
00:09:52
bias right so kind of which people fell
00:09:54
along by the wayside
00:09:56
if you know unsuccessful
00:09:58
investors or gamblers then then send
00:10:00
them my way i need to talk to some of
00:10:01
them as well
00:10:04
number six successful gamblers
00:10:06
especially in poker work well with
00:10:08
incomplete
00:10:09
information
00:10:10
um
00:10:11
one thing that differentiates kind of
00:10:13
people in academia and academia is very
00:10:15
useful for basic research and for
00:10:17
advancing science and things like that
00:10:18
right but like but you don't get months
00:10:21
and months and years and years to
00:10:22
develop a hypothesis when there's actual
00:10:25
money on the line by that time the
00:10:26
people who were faster will already have
00:10:28
moved
00:10:29
so the skill is kind of knowing what
00:10:31
information
00:10:32
was available at the time you made the
00:10:34
decision i think there is a limit to how
00:10:36
useful like an after action review might
00:10:38
be if you kind of pretend as though you
00:10:40
had perfect hindsight right we can say
00:10:42
what would we have done about covid now
00:10:44
if we had known all the things we know
00:10:46
now about it back in february 2020 and
00:10:48
it would be an easier problem to solve
00:10:50
but like that's not how the real world
00:10:51
works so decision making under
00:10:52
uncertainty is i think a understudied
00:10:54
and under-appreciate
00:10:56
appreciated topic um
00:10:58
you can't be too much of a control freak
00:11:00
certainly in a game like poker um you
00:11:02
can lose your temper very easily
00:11:04
emotional management's a big part of
00:11:06
people who are successful in these
00:11:07
fields um
00:11:08
but you know all the value is in making
00:11:11
the hard decisions right um hard because
00:11:13
their edge case is hard because
00:11:15
you have some information but not others
00:11:16
and you're going to get some things
00:11:17
wrong but otherwise if it's easy then
00:11:19
everyone else is already going to do it
00:11:20
there's no excess returns to be made
00:11:24
i think related to this is that
00:11:25
successful gamblers focus on process
00:11:27
much more than results
00:11:30
one thing i found good for the book and
00:11:31
refreshing is
00:11:33
people if you ask them kind of what
00:11:34
makes you tick what's your advantage
00:11:36
they're quite thoughtful and deliberate
00:11:38
about it they step back and think about
00:11:40
what skills do i possess what's the
00:11:41
bigger kind of problem that i'm trying
00:11:43
to solve um again emotional displacement
00:11:47
is a big part of this uh i'm not saying
00:11:49
that you're supposed to be totally
00:11:51
emotionless i mean emotion can help to
00:11:53
guide intuition we'll talk about in a
00:11:54
bit um but it does mean you have to be
00:11:56
willing to like lose poker hand lose 10
00:11:59
000 bucks and then play the next hand as
00:12:01
best as you can right or have an
00:12:03
investment or many investments that go
00:12:06
sour uh and still be on your kind of
00:12:08
best state of mind for making future
00:12:10
investments
00:12:12
it's pretty hard
00:12:14
most people aren't very good at this at
00:12:15
all
00:12:16
you can't go too far and sometimes make
00:12:18
excuses there are lots of different ways
00:12:20
to get unlucky in poker or investing um
00:12:23
but having that focus on kind of what
00:12:25
was my thought process um it seems
00:12:28
intuitive people in this audience maybe
00:12:30
but it's not intuitive for the average
00:12:31
person it differentiates differentiates
00:12:33
successful gamblers quite a bit
00:12:37
this next slide here you may have seen
00:12:38
this video online where
00:12:41
there are a bunch of girls bouncing
00:12:43
basketballs and they ask you to count
00:12:44
how many times the basketball passed
00:12:46
back and forth um and during the video
00:12:49
kind of unbeknownst to the watcher a
00:12:51
gorilla walks on stage like things
00:12:53
change colors weird things happen
00:12:55
and the point is to try and say okay
00:12:57
well you were so focused on um counting
00:12:59
the number of basketballs being passed
00:13:00
that you ignored the gorilla in the room
00:13:03
so to speak right um
00:13:05
i think though that if you're instructed
00:13:07
to count the number of passes that is
00:13:09
actually a sign of intelligence that you
00:13:10
do focus on that right if you're an air
00:13:12
traffic controller
00:13:14
and you're trying to like land planes at
00:13:15
miami airport and you get a bunch of
00:13:17
text messages from your teenage daughter
00:13:18
right you shouldn't be distracted by
00:13:20
that you just kind of focus on the task
00:13:21
at hand
00:13:22
but the point is that people have an
00:13:24
amazing ability
00:13:26
to vary what they're paying attention to
00:13:28
especially in the moment
00:13:30
and in pokers is very important because
00:13:32
most of the time poker is pretty boring
00:13:35
right good players fold uh 70 of their
00:13:38
hands for the most part right a lot of
00:13:40
decisions are fairly automatic but maybe
00:13:42
once an hour or once a day you'll have a
00:13:45
decision for tens of thousands of
00:13:47
dollars or where you'll be knocked out
00:13:49
of the tournament or win the tournament
00:13:50
right and so having that kind of um
00:13:52
knowing where you focus your attention
00:13:54
is an underrated skill that i think
00:13:55
people who are at least in poker are
00:13:57
pretty good at
00:13:59
um by the way one thing i've noticed is
00:14:00
that you used to kind of think of poker
00:14:02
players as being kind of quite
00:14:04
schlubby and sloppy and some of them
00:14:06
still are but more and more the good
00:14:07
players are focusing on physical and
00:14:09
mental filness fitness it is like a
00:14:12
demanding game it's a little bit
00:14:14
demanding physically to kind of keep
00:14:16
your composure um and not give away
00:14:19
information from your from your mental
00:14:21
state and to really concentrate is like
00:14:23
actually a lot of work i mean they're
00:14:24
kind of famous studies of chess players
00:14:26
and now magnus carlson will lose you
00:14:28
know five pounds over the course of a
00:14:29
chess match just thinking which might
00:14:31
seem ridiculous but like but if you're
00:14:33
really playing
00:14:34
poker at your a game at your a plus game
00:14:36
which you don't achieve very often i
00:14:38
often find that when i am able to do
00:14:39
that rarely i'm completely wiped out
00:14:42
afterward it is a fiscally demanding
00:14:44
game but in general kind of being having
00:14:47
your ear to the ground looking out for
00:14:48
opportunities and being flexible enough
00:14:50
to take advantage of those opportunities
00:14:51
is an underrated skill set
00:14:55
number nine uh successful gamblers apply
00:14:57
practiced intuition so um i'm 44. one of
00:15:01
the things about becoming middle-aged
00:15:03
you start to kind of
00:15:04
uh potentially trust your gut
00:15:07
a lot more which can be which can be
00:15:08
risky in some ways i think people can in
00:15:10
matters that happen only occasionally
00:15:12
like a presidential election people
00:15:13
think well my gut is that so-and-so will
00:15:15
win
00:15:16
you don't have enough practice at
00:15:17
forecasting elections to really for that
00:15:19
to be very useful right we have one
00:15:20
election every four years
00:15:22
if you're a poker player though you're
00:15:24
playing tens of thousands of hands every
00:15:26
year if you're a vc you're hearing
00:15:29
thousands of pitches every year right um
00:15:32
and part of what experience does is kind
00:15:34
of take things that were once hard and
00:15:35
then become more automatic right i mean
00:15:38
you'll hear people say oh i made a
00:15:40
decision
00:15:41
very quickly right and i
00:15:43
make staff decisions you hear someone
00:15:44
give a pitch um and you decide maybe in
00:15:47
30 minutes whether or not to fund them
00:15:49
or not you hear examples like that um
00:15:51
but that doesn't really reflect the fact
00:15:53
that this person has spent years and
00:15:54
years and years honing their insight and
00:15:57
listening to pitches that that produces
00:15:59
value down the line so it's intuition of
00:16:01
a sword but it's practiced refined
00:16:04
well-calibrated intuition that involves
00:16:07
working really hard and and learning
00:16:08
right i mean in poker you kind of go
00:16:10
through cycles where um
00:16:11
where you kind of are grinding and get
00:16:13
better at something you feel like you
00:16:14
achieve some condition of mastery
00:16:16
then you kind of slip you get complacent
00:16:18
maybe you tilt a little bit
00:16:20
having the ability to like step back and
00:16:21
re-evaluate your game like negroni or
00:16:24
someone and say look i have to make some
00:16:26
adjustments here
00:16:27
maybe i'm doing playing barely ever have
00:16:29
but now i learn different flaws that i
00:16:31
have right my opponents are getting
00:16:33
better too
00:16:34
that bottom of this chart here where you
00:16:35
re-evaluate and grind back up and expect
00:16:38
that you're going to have to keep doing
00:16:40
that for your whole life frankly
00:16:42
differentiates successful people i think
00:16:46
number 10 is one of the more subtle ones
00:16:48
i think but successful gamblers
00:16:50
are structured thinkers they want to
00:16:52
understand
00:16:54
the process behind the outcome again in
00:16:56
poker it's too complex a game
00:16:58
just to purely learn through
00:17:00
memorization
00:17:01
and rote skill right you want to
00:17:03
understand how things adapt what the
00:17:05
underlying kind of curvatures are and
00:17:06
everything um
00:17:08
heuristics or rules of thumb are are
00:17:10
pretty useful right like in poker
00:17:12
heuristic might be that you should um
00:17:14
call with your medium strength hands
00:17:16
right your best hands you raise
00:17:18
um your worst hands you fold or check
00:17:20
right and then you kind of call with
00:17:21
your medium strength hands but they're
00:17:22
acceptance of that and heuristics only
00:17:24
go so far
00:17:25
um but the kind of flip side of this is
00:17:27
like so at this conference there's a lot
00:17:29
of talk about
00:17:30
investing and gambling but also politics
00:17:32
um
00:17:33
i do think when you're kind of the
00:17:34
highly structured philosophical thinker
00:17:37
um
00:17:38
that politics is very frustrating right
00:17:41
people in the other world i work in
00:17:42
which is like you know i run a website i
00:17:45
talk to people in politics talk to
00:17:46
academics who are trying to get on tv
00:17:48
and stuff like that right
00:17:50
it's a lot of very hypocritical and ad
00:17:52
hoc thinking
00:17:53
in politics right it's instrumental
00:17:55
reasoning toward trying to win an
00:17:57
argument at that moment in that day and
00:17:59
so you can see why someone um like elon
00:18:02
musk might get very
00:18:03
fed up
00:18:04
with the kind of hypocritical nature of
00:18:06
political discourse that's another
00:18:07
conversation we might have in the q a
00:18:08
but um but it does lead to some
00:18:10
personality clashes i think
00:18:14
the last one here which is maybe a
00:18:15
little bit counter-intuitive but i've
00:18:17
kind of learned i think from observation
00:18:19
is that the most successful gamblers
00:18:21
don't actually care that much about
00:18:23
money
00:18:24
um
00:18:25
you know they tend to be for the most
00:18:27
part very generous they're always going
00:18:29
to pick up the tab they're going to
00:18:30
spend a lot on good wine and things like
00:18:32
that right um they're in it really for
00:18:35
the love of the game and when they get
00:18:36
bored they tend to quit
00:18:39
again this might differentiate the 0.01
00:18:43
or 0.001 percent from like the one
00:18:45
percent um certainly you want to manage
00:18:47
your money carefully in poker but still
00:18:50
when you have people whose net worth
00:18:51
fluctuates by the amount that poker
00:18:53
players does like poker is basically
00:18:55
like you know three states right super
00:18:57
rich somewhat rich and dead broke like
00:19:00
not a lot in between when that's the
00:19:02
nature of your field and you've been in
00:19:04
those states for different parts of your
00:19:06
life and many good poke players have had
00:19:08
periods where they're broke you tend to
00:19:10
recognize the ephemeral nature of
00:19:11
success and to celebrate life in the
00:19:13
moment a bit more so
00:19:15
that's all thank you very much look
00:19:17
forward to the q a now
00:19:24
ladies and gentlemen nate silver i don't
00:19:25
know if there was this in the
00:19:26
introduction that you run 538 like does
00:19:28
everyone know that the site
00:19:31
so like during the election cycle
00:19:33
nate does the best job of anyone on
00:19:35
earth
00:19:36
gathering together polling data
00:19:38
and then using statistical models to
00:19:40
represent a distribution or a set of
00:19:42
expected outcomes and making that known
00:19:44
on his site in a way that's super i
00:19:46
think informative better than anything
00:19:48
else that's out there so thanks for that
00:19:50
now i get really addicted to them
00:19:51
leading up to elections i'm like i gotta
00:19:53
go in and check what what nate's changed
00:19:54
on the site every day and i know a lot
00:19:56
of people did
00:19:57
then what i observed on twitter which i
00:19:59
think speaks to your first point here
00:20:01
was a lot of people followed your
00:20:03
prediction on uh the trump election
00:20:05
right and they're like
00:20:06
um nate made a prediction that trump was
00:20:08
gonna win but what you said was that
00:20:10
there was a 43 chance that trump would
00:20:13
win and therefore people assumed he was
00:20:14
going to lose and that the nature of a
00:20:17
lot of people's understanding of things
00:20:18
is very deterministic in binary it's
00:20:20
like win or lose as opposed to hey the
00:20:23
future is not
00:20:25
binary the future is not deterministic
00:20:27
it's there's a set of outcomes that may
00:20:28
occur and you can get a better sense of
00:20:30
what that set is you make that
00:20:31
determination enough times you'll you'll
00:20:33
be right most of the time you'll be
00:20:34
wrong some of the time
00:20:36
do you think most so my question for you
00:20:38
based on how i saw people react
00:20:40
negatively yeah to your polling which
00:20:42
was crazy because it was like dude he
00:20:44
gave a really good set of statistical
00:20:45
models
00:20:46
no people blamed you
00:20:49
but my question is like do people have
00:20:50
the ability to think probabilistically
00:20:52
because you say you need to think
00:20:54
probabilistically or is that a minority
00:20:55
of people that can really get themselves
00:20:58
there because so many people just want a
00:20:59
yes or no answer they want to say this
00:21:01
is the future this isn't they don't want
00:21:02
to think about a distribution of
00:21:04
outcomes which is really the way the
00:21:05
world works i mean i think people think
00:21:07
probably all the time in real life right
00:21:09
if you're like running late for a
00:21:10
meeting you have to decide
00:21:12
should i speed what's the risk of
00:21:14
getting into an accident what's the risk
00:21:15
of getting pulled over or things like
00:21:16
that i think in um they don't realize
00:21:18
they're doing that
00:21:19
yeah i mean they don't they don't know
00:21:21
they're doing that like i mean i think
00:21:22
part of the answer is that
00:21:23
i think politics kind of breaks
00:21:25
people's brain right um because like
00:21:28
before
00:21:29
the 2016 election we got a lot of
00:21:31
criticism from like liberals for being
00:21:34
too bullish on trump the market price
00:21:36
was about 15 we have it 30 right so we
00:21:39
are very bullish on trump relative to to
00:21:42
other people but like
00:21:43
people thought we were like playing with
00:21:44
the numbers to try to i don't know
00:21:47
effect turnout somehow or or try to
00:21:49
cover our asses or things like that and
00:21:51
so um so i just don't think that like
00:21:54
politics teaches people to think
00:21:56
critically and it's like yeah part of
00:21:57
that is like they're not thinking
00:21:58
probabilistically but it's more like
00:22:00
people just want to see their kind of
00:22:02
priors confirmed um what was it like
00:22:05
when
00:22:06
people blamed you for trump winning
00:22:08
yeah there was a lot of that right yeah
00:22:10
yeah i mean and why did you pick trump
00:22:12
as president they literally i think they
00:22:14
pointed to you and basically said you
00:22:15
tipped the election like you you it's
00:22:17
almost as if like you convinced people
00:22:19
to go and like how did you how did you
00:22:21
deal with that
00:22:22
you just kind of have to say like [Â __Â ]
00:22:23
you and move on with your life i mean i
00:22:24
don't like i've been very lucky
00:22:28
in my life overall right including that
00:22:29
like
00:22:30
things that i think are not that
00:22:31
remarkable like you know
00:22:33
saying oh the polls are gonna be right
00:22:34
usually which they were in 2008 and
00:22:36
2012. like i think we got way too much
00:22:38
credit for that but you can't like i
00:22:40
mean again we're all poker players here
00:22:41
right so it's kind of like if you kind
00:22:43
of get your money in
00:22:45
really good right and you make a flush
00:22:47
or something like it you kind of have to
00:22:49
have that attitude about it yeah
00:22:52
so you were able to just basically shut
00:22:53
it out and say this and and what about
00:22:55
just the fact that like you live inside
00:22:57
of an organization now 530 it is owned
00:23:00
by abc's maybe senior year
00:23:03
um
00:23:03
was that before the election
00:23:05
that was we were we were
00:23:07
espn at the time of 2016 election now
00:23:08
we're abc news currently
00:23:11
and so does that change sort of like
00:23:13
um
00:23:14
like how do you make sure that you can
00:23:16
actually just do your job no matter how
00:23:18
unpopular the answer may be or the
00:23:20
distribution of outcomes may be
00:23:23
okay so why am i working on this book
00:23:25
it's partly because i want to diversify
00:23:26
away from just uh
00:23:28
first of all i think
00:23:30
forecasting elections is an interesting
00:23:32
problem but it's not like the most
00:23:34
interesting problem right but no i'm not
00:23:36
going to spend the rest of my life just
00:23:37
like being like the election forecasting
00:23:39
guy it's it's um
00:23:41
i mean you also face disincentives where
00:23:43
like
00:23:44
the downside like we got some elections
00:23:46
like really right like 2018
00:23:48
midterm we like nail like perfectly
00:23:51
about as good as you could right and
00:23:53
people are like oh that's nice right
00:23:55
um if you get kind of like the 2020
00:23:57
primary like we were very bullish on joe
00:23:58
biden when people weren't and thought he
00:24:00
was going to lose and so you don't get
00:24:01
like a lot of
00:24:03
credit and you do face like a lot of
00:24:04
downside um
00:24:06
for being
00:24:07
actually wrong or perceived to be wrong
00:24:10
um
00:24:11
and so yeah i'm not going to spend the
00:24:12
rest of my life making
00:24:14
election forecast i hope that's part of
00:24:16
the portfolio of things that i do but
00:24:17
like i'm interested in like a lot of
00:24:19
other questions about the world nate
00:24:20
what do you think about like social
00:24:22
markets that are defined about social
00:24:24
behavior like an election people are
00:24:25
going to vote one way or another you
00:24:27
then make a prediction and then that
00:24:29
changes the way people vote and so like
00:24:32
and the same is true in stock markets i
00:24:33
was going to ask you about your
00:24:34
experience in stock markets because
00:24:36
people have an assumption about rate
00:24:37
hikes the the market then reacts to the
00:24:40
assumption about rate rate hikes then
00:24:42
the market tanks and then the fed is
00:24:44
like let's change the rate
00:24:46
system because the market's tanking how
00:24:48
do you think about building
00:24:50
probabilistic models that account for
00:24:52
social feedback loops where the social
00:24:54
decision that's made is driven by the
00:24:56
forecast you make or by the the market
00:24:58
making a forecast i think the evidence
00:25:00
is that in two-way elections so democrat
00:25:02
versus republican
00:25:03
that there's no clear effect from and by
00:25:05
the way like what we do is
00:25:07
is a tradition there's always been horse
00:25:08
race media coverage has always been
00:25:10
polling going back for many years right
00:25:12
um
00:25:12
you know we think our version of it is a
00:25:14
little bit more
00:25:15
grounded and objective um in two way
00:25:18
races it's not clear if there's any
00:25:19
impact either way where can an effect is
00:25:21
when you're having a multiple candidate
00:25:23
race right um
00:25:24
in 2020 when elizabeth warren started to
00:25:28
fall behind bernie sanders in the polls
00:25:30
right if you're a
00:25:31
left-wing voter who prefers warren to
00:25:34
sanders but sanders to biden right you
00:25:37
may kind of switch your vote as a result
00:25:38
of the poll that can create momentum
00:25:41
mathematically trying to not uh to model
00:25:43
like a multi-way election we do it it's
00:25:45
very hard it's way more complex in part
00:25:47
because of the feedback loops that it
00:25:49
introduces for sure
00:25:50
and stock markets are the same
00:25:52
again i don't know
00:25:53
any financial market behavior it is
00:25:55
funny by the way that like if i like
00:25:57
lose a thousand bucks at poker i'd be
00:25:59
like furious right but like i lose like
00:26:00
much more than that in the stock market
00:26:02
and like oh i don't care i just invest
00:26:04
in you know
00:26:05
uh low fee like mutual funds and stuff
00:26:07
like that i mean for sure i mean it's um
00:26:10
i don't know you guys would know more
00:26:11
than i would but i think there are there
00:26:13
are certainly complexities that involve
00:26:15
um
00:26:16
feedback loops for sure nate we live in
00:26:18
a very polarized world right now we all
00:26:20
spend too far too much time on twitter
00:26:23
and we're seeing this you know sort of
00:26:25
house of mirrors i'm curious how
00:26:28
you know the 19
00:26:30
of people
00:26:31
who believe in uh you know banning
00:26:34
abortion and their impact
00:26:37
and their outsized voice versus the far
00:26:39
left and their outside voice with you
00:26:41
know sort of democratic socialism
00:26:44
um affect your ability to predict what's
00:26:47
going to happen in the country because
00:26:48
the majority of the country's beliefs
00:26:50
sometimes are in stark contrast to
00:26:53
uh political things that happen yeah and
00:26:56
even on our podcast people want to pit
00:26:59
you know um saks and i as being
00:27:01
incredible rivals no no you you want to
00:27:04
pit yourself
00:27:06
just to be clear in fact we agree on
00:27:08
almost everything yet it feels like in
00:27:10
our society
00:27:12
you know they want to push us to to pick
00:27:15
very binary belief systems and i i'm
00:27:18
just curious how that impacts what you
00:27:20
do day to day in your job in predicting
00:27:22
what's going to actually happen in
00:27:23
reality and maybe you could tell us
00:27:26
based on what you've learned over time
00:27:28
are are people as polarized as it seems
00:27:31
or is america actually got a
00:27:34
majority
00:27:36
um consensus on most of the issues that
00:27:38
we struggle with
00:27:40
well okay i one reason that markets work
00:27:43
is because you get feedback in the
00:27:44
market right and you adjust your
00:27:46
strategy as a result of negative
00:27:48
feedback from consumers people voting
00:27:49
with their feed um
00:27:51
i think one problem with twitter and
00:27:53
with kind of the political atmosphere
00:27:54
more generally is that
00:27:56
both the left and the right are now
00:27:58
insulating themselves from
00:27:59
from feedback in a lot of ways and if
00:28:01
you're someone who kind of like
00:28:03
tries to call bs i mean when you know if
00:28:05
there's a 12th character just like
00:28:07
having a good [Â __Â ] detector is kind
00:28:08
of an implicit theme of people who are
00:28:09
successful gamblers right and if you're
00:28:11
like someone who's like i that's kind of
00:28:13
[Â __Â ] i can't let that sit there i
00:28:15
have to say something about it even
00:28:16
though it'll piss off my followers like
00:28:18
you really get boundary policed a lot
00:28:20
and totally yeah boundary police yes
00:28:23
yeah and when feedback accidents are
00:28:25
broken then you run into potentially
00:28:26
some really
00:28:28
dangerous outcomes right like i think i
00:28:29
think um
00:28:31
both the left and the right have become
00:28:33
more authoritarian in certain ways i
00:28:34
think the type of authoritarianism on
00:28:36
the right is quite a bit more dangerous
00:28:38
in the moment
00:28:39
but
00:28:40
uh they're both kind of moving away from
00:28:43
kind of liberalism in some kind of
00:28:44
classical sense in some ways and and i
00:28:46
think that has to do with
00:28:48
the feedback loops that social media
00:28:50
tend to tend to produce that they're
00:28:51
very the other thing about social media
00:28:52
is like it really really flattens things
00:28:54
right one thing that people
00:28:56
in silicon valley get right for all
00:28:59
their faults and self-servingness in
00:29:01
some ways right
00:29:03
they kind of get the orders of
00:29:04
magnitudes right they understand scale
00:29:06
um
00:29:07
and social media and the daily news
00:29:08
cycle really collapses that right so
00:29:11
some extremely important problem and
00:29:14
some trivial nonsense are kind of
00:29:15
treated on the same level when one is
00:29:18
one thousand times more important than
00:29:20
the other and that and that pretty much
00:29:21
right in a feed it's one right after the
00:29:23
other yeah and you know again i don't
00:29:24
know if there are ways to redesign
00:29:26
twitter to make that better but like the
00:29:27
fact that you're kind of
00:29:29
always on right and even before twitter
00:29:30
that kind of cable news cycle you have
00:29:32
to have like some controversy to talk
00:29:34
about and things are completely dropped
00:29:36
the next day there's another news event
00:29:38
i mean sometimes it should be maybe
00:29:39
actually this thing was terrible but
00:29:40
something is like way more important and
00:29:42
so we should continue to focus on that
00:29:44
and so that's damaging i think is there
00:29:46
a way out of this
00:29:47
will this pass in your estimation is
00:29:50
this
00:29:51
you know an intractable problem
00:29:54
i'm a little bit um
00:29:56
bearish and
00:29:58
concerned in the short term yeah uh to
00:30:00
be honest
00:30:02
can you say more what does that mean
00:30:04
i mean there are different types of
00:30:05
risks that we face right i mean you know
00:30:08
one problem in particular is that if we
00:30:10
have a close
00:30:11
election next time particularly one
00:30:13
that's close where the where the
00:30:14
democrat appears to narrowly win right i
00:30:17
mean you're having more and more
00:30:18
republican elected officials secretaries
00:30:19
of state some of the candidates in
00:30:21
pennsylvania for governor for example
00:30:22
who kind of believe the election was
00:30:24
was stolen in 2020 and like um which
00:30:27
wasn't to be clear and like that's a
00:30:29
very dangerous
00:30:30
argument to have it may not resolve
00:30:32
itself in the end of democracy but like
00:30:34
that's like a you know a risk that's you
00:30:37
know in the 10
00:30:38
range or maybe higher of a severe
00:30:40
constitutional crisis in 2024 and maybe
00:30:43
repeated that in 2028 and so forth and
00:30:45
so i mean that's the i think the most
00:30:48
immediate threat um but also kind of
00:30:50
government
00:30:51
capacity to deal with big problems right
00:30:54
um
00:30:55
you know i don't necessarily take the
00:30:56
standard line about codeword where you
00:30:58
say oh we should have done this and that
00:31:00
and managed it more top down more but
00:31:01
clearly like state capacity seemed to be
00:31:04
not so great right as we deal with other
00:31:06
existential risks from climate change to
00:31:09
ai potentially to nuclear threats um
00:31:12
you know a little bit bearish about
00:31:14
state capacity
00:31:15
to make good decisions
00:31:17
in complex situations
00:31:19
perfect information yeah you know like
00:31:21
being down in um
00:31:23
[Music]
00:31:24
miami you know or traveling other parts
00:31:26
of the world you see like there's more
00:31:27
reason to be optimistic right where um
00:31:30
where people at least are excited and
00:31:32
dynamic and there's something happening
00:31:33
it may not be exactly the right thing
00:31:34
but people are
00:31:36
are
00:31:37
are happy um but no i think i think i
00:31:40
don't know i'm usually an optimistic
00:31:41
person and i'm a little bit worried
00:31:42
about
00:31:43
what should we expect then going into
00:31:45
the 2020 midterms
00:31:47
202 midterms 122 midterms so the
00:31:49
baseline is pretty strong which is that
00:31:51
the president's party loses seats the
00:31:53
large majority of the time um
00:31:55
and given that democrats majorities are
00:31:57
so narrow
00:31:58
they're pretty big underdogs to keep
00:32:00
both the house and the senate um
00:32:03
with that said
00:32:04
in senate races in particular individual
00:32:06
candidates do matter to some extent and
00:32:08
you might have enough off-the-wall gop
00:32:11
candidates nominated that would cost
00:32:13
them two or three seats potentially
00:32:16
and that might be enough to save
00:32:17
democrats
00:32:19
in the senate the house for the
00:32:20
candidates are more anonymous
00:32:22
and their more seats in play is a little
00:32:23
bit tougher um
00:32:25
i do think something like roe v wade
00:32:27
is important um democrats talk a lot
00:32:30
about republican extremism but a lot of
00:32:31
it is theoretical right it's saying well
00:32:33
what will happen in 2024
00:32:35
will be this or that this is the kind of
00:32:36
rare example of big social change by the
00:32:39
opposition party
00:32:41
and that could i think motivate
00:32:42
democrats a little bit but but midterm
00:32:45
elections are more predictable
00:32:47
than presidential elections and the and
00:32:48
the base cases that you wind up with um
00:32:51
with republican majorities of some
00:32:53
magnitude sac tony thoughts on nate's
00:32:56
handicapping of the situation and
00:32:59
description of the political climate
00:33:00
we're in
00:33:01
well i mean pretty clearly the
00:33:04
republicans are expected to have big
00:33:06
pickups uh in this midterm right i mean
00:33:08
i agree with you the senate is tougher
00:33:10
for the gop because
00:33:12
there aren't that many
00:33:14
most of the at-risk seats are on the
00:33:16
republican side right not the democratic
00:33:18
side yeah i mean there are there also
00:33:20
seats like pennsylvania the gp
00:33:22
held and could lose potentially but you
00:33:23
have like i mean i don't know how
00:33:25
technical you want to get um
00:33:28
if you have a generic ballot that favors
00:33:30
republicans by like two points the
00:33:31
democrats can probably keep the senate
00:33:34
if it gets to like six points or seven
00:33:35
points then
00:33:37
then the wall topples and like there's
00:33:39
just too many forces going the other way
00:33:41
and so yeah so what so where do you
00:33:43
where is it right now are you looking at
00:33:45
that
00:33:46
right now it's two or three however
00:33:48
that's among registered voters and
00:33:50
doesn't reflect voter enthusiasm
00:33:52
um so typically the challenger party is
00:33:55
more
00:33:56
motivated to turn out you saw that in
00:33:58
states like virginia last year um
00:34:02
that gap seems to be closing a little
00:34:04
bit um
00:34:07
i do think some of the uh
00:34:09
anti-abortion legislation right some of
00:34:12
the facts the gop kind of is now
00:34:14
back on offense on culture wars against
00:34:17
lgbt people right um i think
00:34:21
potentially is is
00:34:23
going to help motivate democrats to turn
00:34:25
out right um it wasn't the stuff in
00:34:27
virginia where they were talking about
00:34:28
oh schools and kind of i mean you always
00:34:31
want to like play
00:34:33
attack
00:34:34
in politics right um and republicans
00:34:37
found some effective counter attacks
00:34:38
with respect to educational policy
00:34:40
certain attitudes where kind of very
00:34:42
left-wing ideas have crept into
00:34:44
discussions of education and race and
00:34:46
other things right um but now they're
00:34:48
like kind of calling like
00:34:50
um
00:34:51
gay and lesbian people and trans people
00:34:52
pedophiles and stuff like that right and
00:34:54
that's like surprisingly mainstream
00:34:55
discourse and like that's not i don't
00:34:57
think
00:34:58
politically wise it's also a little bit
00:34:59
dangerous in terms of potentially you
00:35:02
know triggering violence and things like
00:35:04
that um
00:35:05
so i don't know i mean neither party can
00:35:07
really like
00:35:08
help itself i don't think can kind of
00:35:10
being like the worst
00:35:12
version of itself
00:35:14
right yeah that's interesting but yunkan
00:35:17
is an interesting case because one year
00:35:19
after biden won that state by 10
00:35:22
yonkin wins that race by two or three so
00:35:25
it was
00:35:26
12 point swing yeah in one year
00:35:29
um
00:35:30
how
00:35:31
emblematic is that of how quickly the
00:35:35
mood of the country has changed is it
00:35:36
like a one-off because janka was a
00:35:38
strong candidate against a weak
00:35:39
candidate or does it represent
00:35:42
you know a much stronger generic
00:35:43
republican ballot
00:35:45
i think it's it's i mean that's what you
00:35:47
might get if you had
00:35:49
50 glen younkins running for for senate
00:35:51
right um he clearly
00:35:53
i mean there's usually a pretty big
00:35:54
midterm shift but that was a little
00:35:56
bigger
00:35:57
than normal in virginia um and that
00:35:59
reflects the fact that i mean you can
00:36:01
like model it out mathematically we have
00:36:02
lots of data points we have 35 center
00:36:04
races every two years right you can
00:36:06
actually see the effect of being more
00:36:08
moderate we'll gain you an extra
00:36:09
two or three or four points right being
00:36:12
less
00:36:13
moderate more radical in some cases will
00:36:15
lose you two or three or four points and
00:36:16
so and so
00:36:18
the fact that the gp didn't kind of take
00:36:19
that lesson
00:36:21
and say we can actually kind of occupy
00:36:23
uh
00:36:24
the mainstream and
00:36:27
have potentially really large majorities
00:36:28
at least in 2022 i mean i think that's
00:36:30
kind of
00:36:32
kind of disheartening in in some ways
00:36:34
and of course trump is still a big
00:36:35
figure
00:36:36
democrats learnt the lesson of that race
00:36:38
either
00:36:39
because of course they haven't yeah i
00:36:41
don't think anyone's really learning
00:36:42
very much
00:36:43
nobody says
00:36:44
because okay that's right the lesson for
00:36:46
democrats would be to tack towards the
00:36:48
center as well and
00:36:50
uh instead of embracing sort of the the
00:36:51
progressive left
00:36:53
i mean the democrats are still talking
00:36:55
about canceling student debt which is a
00:36:57
fringe issue that affects eight to ten
00:36:59
percent of the democratic party
00:37:01
without realizing that it actually is
00:37:03
because it represents 50 or 60 percent
00:37:05
of the working majority of the
00:37:07
democratic infrastructure and that's
00:37:09
really for sure i mean people are people
00:37:10
are
00:37:11
people are self-serving and if you don't
00:37:13
have good vs detectors then
00:37:15
uh you become more self-serving right
00:37:17
like clearly like the class of people
00:37:19
who are uh
00:37:20
high student loan debt
00:37:23
are people who are well educated and
00:37:25
aren't necessarily making as much money
00:37:28
and those described people that are tend
00:37:30
to be democrats right it describes
00:37:31
people like academics some people in
00:37:33
media like i'd be very
00:37:35
democrats used to also be the the
00:37:37
workers party for the welders the
00:37:39
truckers the construction workers well
00:37:41
you should say you should say let's
00:37:43
forgive up to ten thousand dollars in
00:37:44
debt
00:37:45
for
00:37:46
public
00:37:47
university students
00:37:49
undergraduate educations right that
00:37:51
really does capture a lot more people
00:37:53
who are who are you know truly middle
00:37:55
class right half of student loan debt is
00:37:58
people going to graduate school
00:38:00
um
00:38:00
you know i'm not that sympathetic if you
00:38:02
went to
00:38:03
dartmouth and got a philosophy
00:38:05
degree master's master's right um yeah
00:38:08
you're on your own
00:38:10
school and yeah i got a degree in
00:38:12
nursing and kind of yeah why should an
00:38:14
electrician bail out the person getting
00:38:15
a master's in philosophy but like this
00:38:17
is kind of a matter of like the
00:38:20
i mean that's we that's a reasonable
00:38:22
thing to say yeah it doesn't seem
00:38:24
partisan it doesn't seem unreasonable
00:38:26
super reasonable but but you know but uh
00:38:29
i don't know and again people are are
00:38:30
it's also different i think if you
00:38:31
needed more financial stimulus then it
00:38:33
might make more sense to find some way
00:38:35
to do that this is a thing that the
00:38:36
president can probably do
00:38:38
by executive authority it's a little bit
00:38:40
disputed um but no you kind of have i
00:38:42
mean
00:38:43
both parties are are quite self-serving
00:38:47
um and if you're kind of not listening
00:38:48
to critiques or feedback then you can
00:38:51
kind of wallow in in
00:38:53
in that self-serving i think a little
00:38:55
bit more easily all right everybody
00:38:57
please thank nate silver thank you
00:39:03
we'll let your winners ride
00:39:06
rain man david
00:39:10
and it said we opened sources to the
00:39:12
fans and they've just gone crazy
00:39:20
[Music]
00:39:34
we should all just get a room and just
00:39:35
have one big huge orgy because they're
00:39:37
all just
00:39:37
it's like this like sexual tension that
00:39:39
they just need to release
00:39:45
[Music]
00:39:47
we need to get mercies
00:39:52
[Music]
00:39:56
i'm going on
00:39:58
[Music]
