Transcript: Jeffrey Rosenthal on The Curious World of Probabilities | Dec 16, 2006

[Theme music plays]

The opening sequence rolls. The logo of “Big Ideas” featuring a lit lamp bulb appears against an animated blue slate.
Then, Andrew Moodie appears in the studio. The walls are decorated with screens featuring lit lamp bulbs, and two signs read “Big ideas.”
Andrew is in his early forties, clean-shaven, with short curly black hair. He’s wearing an orange shirt.

He says HELLO.
I'M ANDREW MOODIE.
THIS IS
BIG IDEAS.

He shows a scratch lottery ticket and says
AND THIS
IS A SCRATCH AND WIN
LOTTERY TICKET.
I FOUND IT WHILE I WAS CLEANING
UP MY OFFICE.
THE GRAND PRIZE - A COOL TEN
MILLION DOLLARS.
MY FRIENDS, IF I HAVE IN MY
GRUMMY LITTLE HAND A WINNING
TICKET, I'M GONNA CASH THIS BABY
IN AND GET T HELL OUTTA HERE!
HA-HA!
YOU HEARD ME.
I'M GONNA GO STRAIGHT TO THE
BAHAMAS AND LIVE THE HIGH LIFE!
OF COURSE, YOU KNOW, I'M GONNA
SHARE SOME OF MY GOOD FORTUNE
WITH MY FRIENDS HERE AT THE
STUDIO.
I'LL SLIDE A COOL 50 THOUSAND TO
TINA, THE FLOOR DIRECTOR,
BECAUSE SHE'S BEEN NICE TO ME.
I'LL SLIP ROGER AND DAVID 20
THOUSAND EACH, AND WODEK, MY
PRODUCER...


He shows a picture of a man in his forties sitting behind a desk.

He continues OH, I'LL MAKE HIM PAY.
OH, YES.
NO MORE PUSHING ME AROUND.
LET'S SEE.
MAYBE I'LL BUY OUT HIS CONTRACT
HERE AND I'LL MAKE HIM MY
PERSONAL MAN SERVANT!
YEAH...
HUH?
BUT YOU'D HAVE TO WEAR A PINK
TUTU AND YOU'D HAVE TO CALL ME
HE-MAN MASTER OF THE UNIVERSE.
OH, I'M LOVING IT.
HA-HA-HA!
NOW, I KNOW WHAT YOU'RE
THINKING, OKAY.
WHAT ARE THE CHANCES OF MY
ACTUALLY WINNING TEN MILLION
DOLLARS?
WELL, YOU HAVE TO ADMIT, IT IS
POSSIBLE.
BUT IS IT PROBABLE?
LET'S LISTEN TO JEFFREY
ROSENTHAL.
HE IS A PROFESSOR OF STATISTICS
AT THE UNIVERSITY OF TORONTO AND
HE IS THE AUTHOR OF THE BOOK
“STRUCK BY LIGHTNING: THE
CURIOUS WORLD OF PROBABILITIES.”

A clip plays in which Jeffrey Rosenthal stands behind a wooden podium in a university classroom, giving a lecture.
He’s in his late thirties, with long curly brown hair and a stubble. He’s wearing a red shirt.

A caption reads “Jeffrey Rosenthal. Professor of Statistics, University of Toronto. J.J.R. Macleod Auditorium, October 29, 2006. Royal Canadian Institute for the Advancement of Science Lecture Series.”

Jeffrey says I WROTE
THIS BOOK LAST YEAR AND WHAT
I'LL DO TODAY IN MY PRESENTATION
IS TRY TO LARGELY DEVELOP THE
THEME OF THAT BOOK WHICH IS
THAT, UM, FOR A PROBABILITY AND
UNCERTAINTY IN RANDOMNESS ARE
WITH US EVERYWHERE, SO WHENEVER
WE'RE NOT SURE WHAT'S GONNA
HAPPEN NEXT, THAT'S UNCERTAINTY,
RIGHT?
THAT'S RANDOMNESS.
SO, WHETHER IT'S, UM, WHETHER
WE'RE ON AN AIRPLANE AND WE'RE
NOT SURE IF THE AIRPLANE'S GONNA
CRASH; OR WE'RE WAITING FOR A
BUS AND WE DON'T KNOW IF THE BUS
IS GONNA BE LATE; OR WE HEAR
THAT THE TERRORISTS ATTACK AND
WE DON'T KNOW IF THEY'RE GONNA
ATTACK AGAIN OR WHO'S GONNA BE
NEXT; OR OTHER THINGS, LIKE WE
HEAR ABOUT AN ELECTION POLL THAT
SAYS IT'S ACCURATE 19 TIMES OUT
OF 20 AND YOU SAY “WELL, WHAT
DOES THAT MEAN, RIGHT? .”
OR, UM, LET'S SAY WE HEAR ABOUT
A MEDICAL STUDY THAT SAYS IT
PROVES THAT, YOU KNOW, SOME, UH,
SOME DRUG IS GOOD FOR YOU.
WELL, HOW CAN A MEDICAL STUDY
PROVE SOMETHING?
WHAT DOES THAT MEAN?
UM, OR MAYBE YOU GO TO THE
CASINO AND YOU GAMBLE, OR MAYBE
YOU, UM...
MAYBE YOU HEAR ABOUT A
SURPRISING COINCIDENCE; YOU RUN
INTO SOMEBODY THAT YOU HAVEN'T
SEEN IN YEARS WHEN YOU LEAST
EXPECTED TO SEE THEM...
WELL, ALL
THOSE THINGS ARE RELATED TO
PROBABILITY AND RANDOMNESS AND
I'LL TRY TO TALK ABOUT SOME OF
THEM TODAY, BUT, UM, I WILL ALSO
TRY - SINCE THIS IS A SCIENTIFIC
LECTURE SERIES - I'LL TRY TO
MAKE CONNECTIONS TO MY OWN AREA
OF RESEARCH WHICH IS MONTE CARLO
ALGORITHMS AND I'LL TRY TO SORT
GIVE YOU A HINT ABOUT WHAT
THAT'S ALL ABOUT TOO, SO
HOPEFULLY THAT'LL ALL WORK
TOGETHER.
THIS IS A LITTLE SIMPLE SURVEY
TO GET US WARMED UP.
THIS IS SOMETHING THAT WAS
CONDUCTED LAST YEAR AND IT ASKED
THE FOLLOWING QUESTION THAT YOU
CAN READ...

A slate appears on a screen. It reads “Example: a survey, July 2005. Which medium do you rely on most to keep abreast of the news?”

Jeffrey says OKAY, A
REASONABLE QUESTION, AND IT'S
MULTIPLE CHOICE AND YOU CAN SAY,
YOU KNOW, THE NEWSPAPER OR THE
RADIO OR THE TELEVISION.
WELL, HERE'S WHAT THE SURVEY
FOUND - AN ASTONISHING 62 percent OF
PEOPLE REPLIED THAT THEY USE THE
INTERNET.
AND YOUR FIRST-- YOU SAY “GEE,
WELL, WHAT DOES THAT PROVE?”
DOES THAT PROVE, YOU KNOW, JUST
HOW MUCH COMPUTERS HAVE BECOME
INFLUENTIAL IN OUR SOCIETY?
DOES IT SHOW HOW MUCH THINGS CAN
CHANGE IN A SHORT PERIOD OF
TIME?
WELL, THEN YOU LEARN A LITTLE
MORE, AND WHAT IF I TELL YOU
THAT THE SURVEY WAS CONDUCTED ON
THE WEBSITE GLOBEANDMAIL.COM?

[Audience laughter]

Jeffrey continues THEN YOU
SAY, “WELL, WAIT A MINUTE,”
RIGHT?
THIS IS WHAT WE CALL A BIAS
SAMPLE.
THEY WERE SAMPLING PEOPLE WHO
WERE ALREADY GETTING THEIR NEWS
ON THE INTERNET, RIGHT?
SO IF THERE'S ANYTHING
SURPRISING ABOUT THE FIGURE IS
THAT IT'S NOT EVEN MORE THAN
62 percent.
SO WHAT HAVE WE LEARNED FROM
SOMETHING LIKE THIS?
WELL, THE FIRST THING WE LEARN
IS THAT YOU CAN'T BELIEVE
EVERYTHING YOU READ, AND IF
SOMEONE SAYS 62 percent SAID THE
INTERNET, TRY TO GET A LITTLE
MORE INFORMATION ABOUT IT.
AND THE REAL POINT OF THIS TO ME
IS THAT SOMEHOW ANYTHING YOU
HEAR ABOUT, ALWAYS THINK IN
TERMS OF, YOU KNOW, WHAT ARE THE
PROBABILITIES OR THE
UNCERTAINTIES OR THE SAMPLING
INVOLVED?
YOU HAVE TO ALWAYS KEEP THAT IN
MIND, AND I'LL ACTUALLY COME
BACK TO THIS ISSUE OF HOW TO
SAMPLE LATER ON 'CAUSE IT'S
RELATED TO MONTE CARLO STUFF.
UH, BUT ANYWAY, LET ME MOVE ON
FOR NOW IF IT'S ALL RIGHT.
LET ME TALK A LITTLE ABOUT THE
LOTTERY, SO...
SOMEHOW LOTTERIES, THEY REALLY
FASCINATE PEOPLE AND, UH, SINCE
I HAVE BEEN, YOU KNOW, DOING A
LOT OF BOOK PUBLICITY AND THE
MEDIA, I CAN TELL YA, SO MANY
QUESTIONS THAT YOU GET, THEY'RE
ALWAYS, YOU KNOW, THINGS TO DO
WITH LOTTERIES, RIGHT.
AND, UH, TO GIVE YOU ONE IDEA,
LAST YEAR AT ONE POINT, THE
LOTTO 649 JACKPOT HAPPENED TO
ESPECIALLY HIGH - IT WENT UP TO
ABOUT 54 MILLION DOLLARS.
WELL, IN THE 24-HOUR PERIOD
BEFORE THAT DRAW I DID EIGHT
DIFFERENT INTERVIEWS WITH
TELEVISION AND RADIO STATIONS
'CAUSE EVERYBODY WAS ASKING
ABOUT PROBABILITIES.
SO, FIRST OF ALL, LET'S JUST
THINK ABOUT THAT A LITTLE BIT.
SO, YOU CAN SAY, YOU KNOW, YOU
BUY A LOTTERY TICKET, WILL YOU
WIN THE BIG JACKPOT, RIGHT?
WELL, I PROBABLY DON'T HAVE TO
TELL AN AUDIENCE LIKE THIS, THAT
YOUR CHANCES OF WINNING ARE NOT
TOO HIGH.
SO THAT PROBABLY WON'T CAUSE
MUCH SURPRISE, BUT LET ME JUST
SAY A LITTLE MORE ABOUT IT
BECAUSE I THINK IT GETS US
THINKING,
SO YOU CAN SORT OF SAY “WELL,
JUST HOW UNLIKELY IS IT?”
YOU KNOW, THAT IF YOU BUY THIS
TICKET YOU'RE GONNA WIN THE
JACKPOT.
WELL, I CAN TELL YOU THE
NUMBERS.
I CAN SAY IT WAS SOMETHING LIKE
THE LOTTO 649, YOU KNOW, THE
CHANCE THAT YOU'LL WIN MEANS
YOUR SIX NUMBERS HAVE TO MATCH
THE SIX NUMBERS THEY CHOOSE, AND
THE PROBABLY IS, OH, ON CHANCE
IN 13,983,816.
ABOUT ONE CHANCE IN 14 MILLION.
A LOT OF PEOPLE WHEN THEY HEAR
THIS THEY SAY “WELL, OKAY, BUT
YOU KNOW, I STILL THINK I'M
GONNA WIN.”
OR...

[Audience laughter]

Jeffrey continues “YOU KNOW,
I DON'T REALLY-- WHAT IS THAT?
YOU KNOW, WHAT IS THIS ONE IN 14
MILLION?”
SO, I THINK IT'S HELPFUL TO SORT
OF TRY PUTTING IN PERSPECTIVE BY
COMPARING IT TO THINGS.
SO, HERE'S ONE COMPARISON - THE
CHANCE THAT YOU'LL WIN THE
LOTTO 649 WITH A SINGLE TICKET
IS ABOUT THE SAME AS THE CHANCE
THAT IF YOU CHOOSE AN ADULT
WOMAN AT RANDOM THAT IN THE NEXT
MINUTE AND A HALF SHE WILL GIVE
BIRTH TO A BABY.

[Audience laughter]

Jeffrey continues PRETTY
UNLIKELY, RIGHT?
OR, IF YOU HAVE TO DRIVE ACROSS
TOWN TO THE STORE TO BUY YOUR
LOTTERY TICKET YOU'RE ABOUT
TWICE AS LIKELY TO DIE IN CAR
CRASH GOING TO THE STORE AS YOU
ARE TO WIN THE JACKPOT.
OR IF YOU BUY A TICKET ONCE A
WEEK, YOU'LL WIN THE BIG JACKPOT
ON AVERAGE ABOUT ONCE EVERY
270,000 YEARS.
SO WHEN YOU PUT IT INTO THOSE
PERSPECTIVES PEOPLE SAY “OKAY.
WELL, MAYBE IT IS REALLY
UNLIKELY.”
BUT THEN PEOPLE SAY, “WELL...”
YOU KNOW, IN KIND OF A KILL-JOY,
RIGHT?
[Chuckles]
'CAUSE WHEN WE'RE ROOTING
SOMETHING IT SOUNDED LIKE A GOOD
PROBABILITY.
SO LET ME FIRST SAY THAT, OF
COURSE, MANY PEOPLE ACTUALLY
ENJOY BUYING LOTTERY TICKETS
'CAUSE IT'S FUN AND EXCITING TO
DREAM ABOUT WHAT YOU MIGHT DO
AND I'VE GOT NOT PROBLEM WITH
THAT, ESPECIALLY SINCE LOTTERIES
ACTUALLY MAKE A LOT OF MONEY FOR
THE GOVERNMENT THAT HELPS TO
SUPPORT THE UNIVERSITIES THAT
HELPS TO PAY MY SALARY, SO...

[Audience laughter]

Jeffrey continues I DON'T
ACTUALLY HAVE A PROBLEM WITH
LOTTERIES BUT YOU SHOULD JUST
UNDERSTAND HOW UNLIKELY IT IS
THAT YOU'RE GONNA WIN THE
JACKPOT.
BUT THEN I SAY “OKAY, SO IT'S
UNLIKELY.”
WELL, BUT THAT SOUNDS LIKE SORT
OF A BAD THING, BUT THERE'S A
FLIPSIDE TO THINKING ABOUT
UNLIKELY THINGS AND PEOPLE DON'T
ALWAYS REALIZE THAT.
SO WHY DON'T WE START WITH THAT,
AND LET ME TALK JUST A LITTLE
BIT ABOUT AIRPLANE CRASHES
AGAIN, JUST KIND OF WARMING UP.
SO...
[Chuckles]
NOT THAT WE'RE WARMING UP WITH
AN AIRPLANE CRASH; WE'RE WARMING
UP BY TALKING ABOUT AIRPLANE
CRASHES, BUT, UM...
SO LET ME START WITH A PERSONAL
STORY JUST TO GIVE MY OWN SORT
OF PERSPECTIVE.
UM, BACK WHEN I WAS A GRADUATE
STUDENT I WAS SCHEDULED TO FLY
TO NEW YORK CITY TO VISIT MY
RELATIVES, AND EXACTLY ONE WEEK
BEFORE MY FLIGHT THERE WAS A BIG
AIRPLANE CRASH AT JOHN F.
KENNEDY AIRPORT IN NEW YORK CITY
WHICH WAS WHERE I WAS FLYING
INTO, AND HERE I HAD A TICKET TO
GO TO THAT SAME AIRPORT ONE WEEK
LATER.
WELL, MY FIRST REACTION, I THINK
LIKE A LOT OF PEOPLE IS “WELL,
GEE.
I CAN'T FLY INTO THE AIRPORT ONE
WEEK AFTER AN ACCIDENT.
YOU KNOW, MAYBE I HAVE TO CANCEL
MY VISIT OR MAYBE I'LL TAKE THE
TRAIN, OR MAYBE I CAN WALK TO
NEW YORK CITY, BUT I'M NOT
GONNA FLY INTO THIS AIRPORT.”
AND THEN I SAID “WELL, HOLD ON.
LET'S JUST THINK ABOUT THIS A
LITTLE BIT,” AND PART OF THE
POINT THAT I TRY TO MAKE IN THE
BOOK AND TODAY IS ABOUT, UM, HOW
OFTEN WHEN YOU'RE FACED WITH
UNCERTAINTY AND RANDOMNESS, YOU
DON'T NEED TO DO ANY REALLY DEEP
THINKING.
YOU JUST NEED TO THINK ABOUT
THINGS JUST A LITTLE BIT, SO...
SO THE FIRST THING I SAID WAS
“OKAY, HOLD ON.”
THERE'S THIS OLD
CLICHE
THAT COMMERCIAL AIRPLANE TRAVEL
IS ACTUALLY QUITE SAFE.
WELL, IT TURNS OUT IT'S ACTUALLY
TRUE, RIGHT, SO, IF YOU GET ON
THE PLANE, THERE'S SOMETHING ON
THE COMMERCIAL AIR FLIGHT,
THERE'S SOMETHING LIKE ONE
CHANCE IN FIVE MILLION THAT
YOU'LL DIE ON THAT FLIGHT.
SO THAT'S SOMETHING LIKE THE
CHANCE YOU'LL WIN THAT JACKPOT,
RIGHT?
SO, IN OTHER WORDS, IF YOU
REALLY START UNDERSTANDING HOW
UNLIKELY YOU'LL WIN THE LOTTERY
JACKPOT, YOU SHOULD ALSO START
UNLIKE-- START UNDERSTANDING
JUST HOW UNLIKELY IT IS THAT
YOU'RE AIRPLANE'S GONNA CRASH.
SO THAT'S THE FIRST THING TO
SAY, BUT THEN YOU SAY “WELL,
WAIT A MINUTE.
BUT, I WAS GOING INTO THE SAME
AIRPORT, RIGHT.
SO, WHAT ABOUT THAT?”
WELL, FIRST OF ALL, I LOOKED IT
UP AND JUST INTO THIS ONE
AIRPORT, JOHN F. KENNEDY AIRPORT
IN NEW YORK CITY - THERE'S OVER
FIVE THOUSAND FLIGHTS EVERY
WEEK.
SO WHAT DOES THAT MEAN?
IT MEANS EVEN IF I KNEW THAT IN
THE NEXT WEEK THERE WAS GONNA BE
ANOTHER CRASH AT THE SAME
AIRPORT, THERE'LL STILL ONLY BE
ABOUT ONE CHANCE IN FIVE
THOUSAND THAT IT WOULD BE MY
FLIGHT, SO STILL VERY UNLIKELY.
AND THE THIRD THING THAT I
REALIZED IS THAT, WELL, JUST
BECAUSE THERE WAS ONE CRASH AT
THIS AIRPORT, THAT DIDN'T MEAN
THAT IT WAS REALLY ANY MORE OR
LESS LIKELY THAT THERE WOULD BE
ANOTHER ONE.
SO THEY'RE WHAT WE CALL, UH,
INDEPENDENT EVENTS, THAT IS,
THERE WAS NO EVIDENCE THAT THE
AIRPORT ITSELF WAS AT FAULT, SO
REALLY THE FACT THAT THERE AS
ONE CRASH DIDN'T MEAN THERE
WOULD OR WOULD BE ANOTHER ONE.
AND THAT'S A BIT LIKE THAT OLD
JOKE ABOUT THE GUY WHO WAS
AFRAID THAT SOMEONE WOULD BRING
A BOMB ONTO HIS AIRPLANE, AND
HIS SOLUTION WAS TO BRING HIS
OWN BOMB ONTO THE PLANE...

[Audience laughter]

Jeffrey continues 'CAUSE HE
FIGURED IT WAS EXTREMELY
UNLIKELY THAT THERE WOULD BE TWO
DIFFERENT BOMBS ON THE SAME
FLIGHT.

[Audience laughter]

Jeffrey continues AND IT WAS
THE SAME THING AT THIS AIRPORT.
SO FOR ALL THOSE REASONS I SAID
“OKAY, ACTUALLY, I DON'T NEED TO
BE WORRIED.”
AND SURE ENOUGH, I FLEW TO NEW
YORK AND IT WAS FINE, AND I
DIDN'T CRASH AND I'M STILL HERE.
AND WHEN I TELL PEOPLE THE STORY
SOME PEOPLE CAN BE KIND OF
SCEPTICAL.
SO ONE RADIO INTERVIEWER WAS
KINDA TYPICAL WHEN HE SAID
“OKAY, THAT'S GREAT.
YOU'VE GOT ALL THESE FANCY
STATISTICS ABOUT HOW SAFE
AIRPLANE TRAVEL IS, BUT DOES
THAT REALLY HELP WHEN YOU'RE UP
IN THE SKY AND THERE'S SOME
TURBULENCE AND THE PLANE'S
JIGGLING AROUND A LITTLE?”
AND IT'S A FAIR QUESTION, BUT MY
ANSWER WAS ACTUALLY “YES,” THAT
IF YOU REALLY START TO BELIEVE
THE STATISTICS, THAT IS, YOU
REALLY START TO THINK JUST HOW
UNLIKELY IT IS THAT EITHER
YOU'LL WIN THE JACKPOT OR YOUR
AIRPLANE'S GONNA CRASH, THEN YOU
SAY ACTUALLY, IT'S SO UNLIKELY
THAT YOU DON'T NEED TO WORRY
ABOUT IT.
SO, YOU KNOW, I, UH...
I USED TO BE A LITTLE BIT OF A
NERVOUS TRAVELLER, YOU KNOW,
WHEN I'D BE UP IN THE SKY I'D BE
LOOKING OUT THE WINDOW TO SEE IF
THE WING WAS STILL ATTACHED
PROPERLY AND ALL THAT.

[Audience laughter]

Jeffrey continues AND THEN
AT SOME POINT I REALIZED “NO,
THESE NUMBERS MEAN-- I REALLY
JUST DON'T NEED TO WORRY,” AND
NOW I DON'T; NOW I JUST WORRY
THAT THE FLIGHT IS GOING TO BE
LATE, WHICH IT OFTEN IS.
ABOUT A QUARTER OF ALL FLIGHTS
ARE LATE, BY THE WAY, BUT, UH...
BUT THE CHANCE OF ACTUALLY BEING
IN TROUBLE ARE REALLY, REALLY
SMALL.
SO I CALL THAT A PROBABILITY
PERSPECTIVE, AND IT'S THIS IDEA
THAT YOU JUST START THINKING A
LITTLE BIT IN TERMS OF THE
PROBABILITIES AND SO ON, AND YOU
CAN HELP TO WORRY LESS ABOUT
THINGS.
SO, LET ME TALK QUICKLY ABOUT A
COUPLE OTHER THINGS THAT ARE
ALSO UNLIKELY.
AND THE NEXT ONE LET ME TALK
ABOUT THE TERRORIST ATTACKS.
SO, AFTER THE, UH, 9-11 ATTACKS
FOR EXAMPLE, YOU KNOW, A SALES
OF ANXIETY MEDICATION WENT WAY
UP AND SO MANY PEOPLE WERE SO
SCARED AND, YOU KNOW, I KNOW A
COUPLE WHO DECIDED THEY'D BETTER
GET MARRIED RIGHT AWAY 'CAUSE
WHO KNS WHAT WILL HAPPEN
TOMORROW.
WELL, WAS ALL THAT REALLY
JUSTIFIED?
WELL, ONE WAY TO THINK OF IT IS
TO SAY JUST IN THE MONTH OF
SEPTEMBER 2001, THE MONTH OF THE
9-11 ATTACKS, MORE PEOPLE, MORE
AMERICANS DIED IN ORDINARY
AUTOMOBILE ACCIDENTS THAT SAME
MONTH THAN DIED IN THE 9-11
ATTACKS.
NOW WHAT DOES THAT MEAN?
WELL, IT'S NOT TO IN ANY WAY
EQUATE THEM IN A MORAL SENSE,
BUT IT IS TO SAFE YOU'RE
WORRIED ABOUT SORT OF THE
PROBABILITIES THAT YOU OR YOUR
LOVED ONE WILL BE KILLED IN THE
ATTACK, WELL...
YOU ACTUALLY HAVE MORE TO WORRY
ABOUT FROM ORDINARY AUTOMOBILE
ACCIDENTS EVEN THAT SAME
MONTH THAN YOU DO FROM TERRORIST
ATTACKS SO THE RL POINT, I
THINK, IS THAT JUST LIKE WE
DON'T ALL GO AROUND TAKING
ANXIETY MEDICATION BECAUSE WE
MIGHT GET IN AN ORDINARY...
ORDINARY CAR CRASH.
I HOPE I DIDN'T SAY AIRPLANE
CRASH; I MN CAR CRASH.
AND, UH, JUST BECAUSE, UM, WE
DON'T GO AROUND WORRYING ABOUT
THAT, WELL, WE ALSO SHOULDN'T GO
AROUND WORRYING THAT WE'RE GONNA
BE A VICTIM OF A TERRORIST
ATTACK.
AGAIN, NOT TO SAY IT'S NOT A
SERIOUS ISSUE, BUT TO SAY IT CAN
GIVE US SOME PERSPECTIVE ON
THINGS.
UM...
AND LET ME, UH, LET ME MENTION
SOMETHING ELSE TO WHILE I'M ON
THE SAME THING-- SAME THEME, AND
THIS IS, UM, HOMICIDE.
SO, UM...
SO YOU MAY KNOW, ESPECIALLY IF
YOU LIVE IN THE CITY OF TORONTO,
THAT, YOU KNOW, HOMICIDES, WELL,
THEY REALLY MAKE THE NEWS,
RIGHT?
AND, UM...
UH, YOU KNOW, IN PARTICULAR, YOU
SEE ALL SORTS OF MEDIA HEADLINES
OR, UH, POLICE SPOKESPERSONS,
OR, UM, OR LET'S SAY POLITICIANS
RUNNING FOR RE-ELECTION.
THEY'LL TELL YOU ABOUT HOW CRIME
IS OUT OF CONTROL AND HOW IT'S
INCREASING AND HOW DANGEROUS
EVERYTHING IS.
WELL...
I MEAN THE FIRST THING TO SAY IS
THAT ACTUALLY CANADA AND TORONTO
ARE EXTREMELY SAFE, AND DESPITE
ALL THE HEADLINES, IN FACT, THE
RATE OF HOMICIDES AND OTHER
VIOLENT CRIMES HAVE GENERALLY
BEEN GOING DOWN A LITTLE BIT IN
THE LAST DECADE AND A HALF.
NOW LAST YEAR IN TORONTO AS YOU
MAY KNOW, THE NUMBER OF
HOMICIDES ACTUALLY DID GO UP; IT
WENT UP ABOUT 22 percent COMPARED TO
THE YEAR BEFORE.
SO THAT'S A SERIOUS THING AND
CERTAINLY SOMETHING WORTH
THINKING ABBOT.
BUT TO JUDGE BY THE MEDIA
HEADLINES AND THE STORIES, YOU
KNOW, THERE WAS ALL THIS TALK
ABOUT HOW TORONTO HAD LOST ITS
INNOCENCE AND OUR STREETS WERE
BATHED IN BLOOD AND IT WAS, YOU
KNOW, THE SUMMER OF THE GUN AND
ALL THESE THINGS.
WELL, YOU CAN GET A LITTLE
PERSPECTIVE AGAIN BY, AGAIN,
JUST THINKING A LITTLE BIT ABOUT
THE NUMBERS.
FOR EXAMPLE, WELL, IT MAY NOT
SURPRISE YOU TO KNOW THAT
TORONTO IS SAFER THAN MOST
AMERICAN CITIES, BUT IT MIGHT
POSSIBLY SURPRISE YOU, GIVEN THE
HEADLINES, TO KNOW THAT
TORONTO'S ALSO SAFER THAN MANY
OTHER CANADIAN CITIES INCLUDING,
FOR EXAMPLE, WINNIPEG AND
EDMONTON AND REGINA, AND ALSO,
THAT EVEN LAST YEAR AT THE
HEIGHT OF THE HEADLINES ABOUT
THE HOMICIDES, THERE WERE FEWER
HOMICIDES IN TORONTO LAST YEAR
THAN THERE WERE BACK IN 1991.
SO, AGAIN, WHAT DOES ALL THAT
MEAN?
WELL, IT DOESN'T MEAN THAT
HOMICIDES AREN'T IMPORTANT, BUT
IT DOES MEAN THAT WE SHOULD KEEP
PERSPECTIVE ON THINGS.
AND ONE POINT IN WANNA MAKE IS
THAT THIS YEAR...
WELL, GUESS WHAT, I'VE BEEN
TRACKING THE HOMICIDE NUMBERS
AND ACTUALLY THIS YEAR, GUESS
WHAT?
THE NUMBERS ARE ACTUALLY BACK
DOWN TO SOMETHING LIKE WHAT THEY
WERE TWO YEARS AGO.
IN OTHER WORDS, THAT LAST YEAR
THING WAS ACTUALLY JUST A
ONE-YEAR SPIKE, IT SEEMS.
NOW, AGAIN, IT'S NOT TO MINIMIZE
IT, BUT IT IS TO SAY THAT
THERE'S BEEN HARDLY ANY MEDIA
HEADLINES ABOUT THAT, RIGHT?
HARDLY ANY BIG HEADLINES ABOUT
HOW, UH, THE MURDER RATE'S GONE
DOWN AND SO ON.
SO, I THINK WE GET WHAT I CALL A
HEADLINE BIAS, THIS IDEA THAT
THE HEADLINES WILL MAKE THINGS
SOUND BAD, BUT IF THINGS ARE
GOOD YOU WON'T HEAR IT AND THEN
PEOPLE GET THE IMPRESSION THAT
THINGS ARE MUCH, MUCH WORSE THAN
THEY ACTUALLY ARE.
UM, ANOTHER WAY TO THINK ABOUT THAT
IS THAT IT SO HAPPENS THAT IN
CANADA MORE PEOPLE ARE ACTUALLY
KILLED BY THEIR SPOUSE THAN ARE
KILLED BY A STRANGER.
SO, YOU COULD THINK, WELL,
INSTEAD OF WORRYING ABOUT WHO'S
HIDING BEHIND THAT LAMPPOST IN
THE PARK...

[Audience laughter]

Jeffrey continues YOU KNOW,
THINK ABOUT WHO YOU CHOOSE TO
MARRY, RIGHT?
SO, AND AGAIN, THE REAL POINT IS
NOT THAT WE SHOULD BE AFRAID
THAT OUR SPOUSE IS GONNA KILL
US, BUT IT'S THAT WE SHOULDN'T E
SO WORRIED THAT WE'RE GONNA BE
THE VICTIM OF A RANDOM SHOOTING.
SO, UM, ANOTHER WAY TO THINK
ABOUT IT IS, UM...
WELL, THERE ARE SOME THINGS
WHICH ARE NOT SO RARE, AND FOR
EXAMPLE, OVER ONE THIRD OF
CANADIANS WILL EVENTUALLY DIE
FROM CARDIOVASCULAR DISEASE, A
HEART ATTACK AND STROKE AND SO
ON
NOW, ONE OF THE RISK FACTORS
ASSOCIATED WITH CARDIOVASCULAR
DISEASE IS STRESS AND WORRY.
AND, IN FACT, IT'SUITE
POSSIBLE IT SEEMS TO ME THAT
MORE PEOPLE ACTUALLY DIE FRO
HEART ATTACKS BECAUSE THEY'RE SO
WORRIED ABOUT THESE VERY RARE
THINGS, THAT ARE ANNUALLY
VICTIMS OF THESE VERY RARE
THINGS.
SO, AND IT'S JUST SOMETHING TO
KEEP IN MIND, BECAUSE IN THE
HEADLINE, WELL OFTEN IT'S IN THE
HEADLINES BECAUSE IT'S RARE, NOT
BECAUSE IT'S COMMON, AND KEEP
PERSPECTIVE ON WHAT'S GONNA
HAPPEN TO YOU.
SO THAT ONE POINT I WANNA MAKE.
UM, LET ME MOVE ON A LITTLE BIT
AND LET ME TALK ABOUT, UH, THE
CASINOS, OR GAMBLING AT CASINOS,
BECAUSE IT'S, AGAIN, SOMETHING
THAT, UM, SORT OF FASCINATES
PEOPLE AND I OFTEN GET A LOT OF
QUESTIONS ABOUT IT, AND I THINK
THERE'S SOME IMPORTANT POINTS
THAT WE CAN LEARN FROM IT, UH,
EVEN IF YOU DON'T GO TO THE
CASINO, BUT MAYBE EVEN MORE SO
IF YOU DO.
AND TO SET IT UP, WELL LET ME
MENTION ONE THING, THAT, UM,
WHENEVER A NEW CASINO IS BUILT
OR IS PROPOSED THERE'S OFTEN
SOME CONTROVERSY.
YOU KNOW, SHOULD WE BUILD THIS
NEW CASINO, OR SHOULDN'T WE?
WILL IT RUIN THE NEIGHBOURHOOD?
WILL IT LEAD TO GAMBLING
ADDICTION?
WILL IT CAUSE PROBLEMS?
BUT NOBODY EVER SAYS “I DON'T
THINK WE SHOULD BUILD THIS
CASINO BECAUSE I DON'T THINK
IT'S GONNA MAKE ANY MONEY.”

[Audience laughter]

Jeffrey continues AND IF YOU
THINK ABOUT IT THAT'S KIND OF AN
INTERESTING THING, RIGHT?
BECAUSE ANY INDIVIDUAL GAMBLER
WHO GOES TO A CASINO, WELL, YOU
MIGHT WIN AND YOU MIGHT LOSE AND
ANYTHING CAN HAPPEN.
HOW CAN THE CASINO BE SO SURE
THAT THEY'RE ALWAYS GONNA MAKE
MONEY?
SO I THINK IT'S WORTH THINKING A
LITTLE MORE ABOUT THAT IN SOME
DETAIL SO LET ME USE AS AN
EXAMPLE THE GAME, UH, CALLED
CRAPS THAT YOU MAY KNOW, THIS
FUNNY-NAMED GAME WHERE YOU
REPEATEDLY ROLL TWO DICE AND,
UH, THERE'S RULES FOR WHETHER
YOU WIN, OR WHETHER YOU LOSE, OR
WHETHER YOU GET TO ROLL AGAIN
AND SO ON, AND IT'S NOT TO HARD
TO WORK OUT THE PROBABILITIES
AND FIGURE OUT THE PROBABILITY
OF WINNING.
AND IT TURNS OUT IT'S EQUAL TO
THIS - IT'S EQUAL TO 49.29 percent
WHY IS THAT INTERESTING?
WELL, THE FIRST THING YOU NOTICE
IS, THIS IS JUST A TINY BIT LESS
THAN 50 percent, AND SO, AS MANY OF YOU
WILL KNOW, THE FACT THAT IT'S A
LITTLE BIT LESS THAN 50 percent, THAT'S
WHAT WE'D SAY IS THE GAME IS
WEIGHTED IN FAVOUR OF THE HOUSE,
RIGHT?
'CAUSE IT'S SORT OF THE HOUSE
EDGE IN THE GAME.
AND IT'S JUST THIS TINY, TINY
LITTLE EDGE.
SO WHAT IT MEANS IS IF YOU GOT
TO THE CASINO AND YOU JUST BET
ONCE IN CRAPS, YOU HAVE JUST
ABOUT THE SAME CHANCE OF WINNING
AS LOSING - SO CLOSE YOU
COULDN'T REALLY NOTICE THE
DIFFERENCES - SO THAT SOUNDS
PRETTY FAIR, RIGHT?
SO IF YOU GO TO THE CASINO AND
YOU PLAY CRAPS ONCE, THAT'S
ACTUALLY PRETTY FAIR.
BUT THEN WHEN IT GETS
INTERESTING IS WHEN YOU SAY
“WHAT HAPPENS IN THE LONG RUN?”
SO WHAT I'M GONNA TRY TO DO IS
ILLUSTRATE SOME GRAPHICAL
SIMULATIONS HERE.
WHAT I'M DOING HERE IS I'M
SHOWING-- THIS IS SUPPOSED TO
REPRESENT SOMEBODY WHO'S GOING
TO THE CASINO TO BET ON CRAPS
AND THEY'RE GONNA BET TEN
DOLLARS ON CRAPS EACH TIME.
AND THE BLACK DOT INDICATES HOW
MUCH MONEY THEY HAVE, AND YOU
CAN SEE IT'S AT A HUNDRED.
SO THEY GO TO THE CASINO WITH A
HUNDRED DOLLARS IN THEIR POCKET
AND, WELL, LET'S SAY, SO THEN
THEY BET.
SO THE FIRST TIME THEY'RE GONNA
EITHER LOSE AND GO DOWN TO
90 DOLLARS OR WIN AND GO UP TO
110 DOLLARS - AND THIS IS ALL
RANDOM, I HAVEN'T PRE-PROGRAMMED
THIS, SO ANYTHING CAN HAPPEN
ACCORDING TO THE PROBABILITIES
FOR CRAPS--
OH, LOOK AT THAT, YOU LOST.
SO IT MEANS NOW YOU GO DOWN TO
90 DOLLARS, RIGHT.
SO IS THAT CLEAR ENOUGH?
SO NOW YOU'RE AT 90 DOLLARS.
AND WE'RE GONNA KEEP GOING, SO
YOU KEEP MAKING BETS.
OH, YOU WON THAT TIME, YOU GO UP
TO A HUNDRED DOLLARS.
UM, OH, YOU WON AGAIN, YOU'RE UP
TO 110 DOLLARS AND SO ON.
AND WE'RE GONNA SAY, WHAT'S THE
CHANCE IF YOU START AT A HUNDRED
DOLLARS THAT YOU'LL MAKE IT UP
TO TWO HUNDRED DOLLARS?
THAT IS, YOU'LL DOUBLE YOUR
FORTUNE.
WHAT'S THE CHANCE YOU'LL DO THAT
BEFORE YOU LOSE ALL YOUR MONEY
AND GO DOWN TO ZERO DOLLARS?
OKAY, WELL, THIS IS SOMETHING WE
CALL THE GAMBLER'S ROOM PROBLEM,
AS FOR MAYBE OBVIOUS REASONS,
AND LET ME SPEED THIS UP A
LITTLE MORE IF IT'S CLEAR WHAT'S
GOING ON, AND THERE WE GO.

As he speaks, an animated chart on display on a giant screen behind him shows the progression of the bets.

Jeffrey continues SO YOU'RE WINNING--
OH, LOOK, YOU'RE UP TO 140, 120,
130...
AND THE FIRST THING YOU'LL
NOTICE IS IT ACTUALLY TAKES A
WHILE, RIGHT?
THERE'S A LOT OF UPS AND DOWNS
WHEN YOU GAMBLE, AND THAT'S PART
OF THE REASON GAMBLING'S
EXCITING--
OH, LOOK, YOU'RE DOING PRETTY
WELL.
NOW SINCE I ONLY HAVE A LIMITED
AMOUNT OF TIME, LET ME SPEED IT
UP A LITTLE MORE.
SO THERE, YOU'RE GAMBLING,
YOU'RE ALREADY UP TO 40 BETS...

[Audience laughter]

The graph slides right and left, representing the win and loss of money.

Jeffrey says 50 BETS, 60 BETS.
THERE'S THE NUMBER OF BETS THERE
IN CASE YOU WANNA FOLLOW ALONG.
WOW, THIS ONE'S TAKING A WHILE--
OH, YOU'RE NOT DOING VERY WELL
RIGHT NOW, BUT YOU'RE HANGIN' IN
THERE, RIGHT?

[Audience laughter]

Jeffrey continues YOU STILL
HAVEN'T LOST YOUR HUNDRED
DOLLARS.
[Chuckles]
AGAIN, I HAVE NO IDEA HOW LONG
THIS IS GONNA TAKE, ALTHOUGH I
CAN SPEED IT UP IF IT GETS TOO
BORING.
BUT, UH...
HERE WE GO, YOU'RE NOT DOING
WELL NOW.
SEE IF I NEED TO SPEED IT UP.
[Chuckles]

The chart finally hits the extreme left and a red square appears.

Jeffrey says OH, THERE YOU GO, YOU LOST.

[Audience sighs]

Jeffrey says SO WHAT
JUST HAPPENED HERE?
SO, YOU WENT TO THE CASINO WITH
A HUNDRED DOLLARS IN YOUR
POCKET, YOU REPEATEDLY BET TEN
DOLLARS AT CRAPS AND EVENTUALLY
YOU LOST YOUR HUNDRED DOLLARS
BEFORE YOU MADE IT UP TO TWO
HUNDRED DOLLARS.
NOW YOU MIGHT SAY, WELL MAYBE
THAT'S ALWAYS GONNA HAPPEN, BUT
IT TURNS OUT IT ISN'T.
SO I'M ALMOST TEMPTED TO RUN IT
AGAIN, BUT I THINK I WON'T PUSH
MY LUCK, BUT IF WE REPEATEDLY
RAN IT, YOU'D FIND OUT THAT
ABOUT 43 percent OF THE TIME, YOU WOULD
ACTUALLY WIN.
SO WHAT'S THE POINT OF THAT?
WELL, THE POINT IS IF YOU GO TO
THE CASINO AND YOU REPEATEDLY
BET TEN DOLLARS AT CRAPS, BUT
YOU ONLY START WITH A HUNDRED
DOLLARS, IT'S STILL PRETTY FAIR,
RIGHT?
IT'S NOT TOO BAD.
SO TO SPEED THINGS ALONG, LET ME
SAY, NOW WHAT IF YOU START WITH
A THOUSAND DOLLARS?
SO I RESET IT NOW, SO NOW YOU'RE
AT A THOUSAND DOLLARS AND YOU'RE
STILL GONNA BET TEN DOLLARS EACH
TIME, AND THE QUESTION IS ARE
YOU GONNA GET UP TO TWO THOUSAND
DOLLARS BEFORE YOU LOSE ALL YOUR
MONEY AND GET TO ZERO?
WELL NOW THEY'RE NOT AS HIGH AS
43 percent; IN FACT, NOW THEY'RE ABOUT
FIVE AND A HALF PERCENT, OR ONE
IN 18.
SO LET ME-- I'LL GO UP TO WARP
DRIVE AS THOSE OF US WHO WATCH
STAR TREK MIGHT SAY AND, UH, YOU
SEE, IT'S STILL TAKING A LONG TIME.

He shows an accelerated version of the animated graph.

He continues WE'RE UP TO FIVE THOUSAND BETS
NOW, SO WE'RE DOING SIX
THOUSAND, SEVEN THOUSAND BETS,
BUT THERE YOU GO, YOU EVENTUALLY
LOST.
AND NOW, AS I SAY, ONLY ABOUT
ONE TIME IN 18 WOULD YOU
ACTUALLY WIN.
LET ME DO ONE LAST THING BEFORE
I LEAVE THIS WHICH IS, WHAT IF
YOSTART WITH FIVE THOUSAND
DOLLARS?
STILL MAKE TEN DOLLAR BETS AT
CRAPS.
AS I SAY, YEAH, CRAPS UNTIL YOU
EITHER GET UP TO TEN THOUSAND
DOLLARS OR LOSE.
WELL, LET ME RUN THIS ONE AT
SUPER WARP SPEED - MAYBE EVEN
SUPER-DUPER WARP SPEED - AND NOT
LOOKIN' GOOD, ALTHOUGH WE'RE UP
TO 40 THOUSAND BETS ALREADY, 50
THOUSAND BETS, AND AT 52
THOUSAND BETS YOU LOST.
WELL, I'M NOT TOO SURPRISED YOU
LOST THAT TIME BECAUSE AT THIS
LEVEL, MAKING TEN DOLLAR BETS
STARTING WITH A THOUSAND DOLLARS
THE CHANCE THAT YOU'LL GET UP
TO, UM...
SORRY, IF YOU'RE STARTING WITH
FIVE THOUSAND, THE CHANCE YOU'LL
GET UP TO TEN THOUSAND BEFORE
YOU LOSE ALL YOUR MONEY IS ONE
CHAN IN 1.4 MILLION, OKAY.
OR IF YOU STARTED WITH TEN
THOUSAND AND TRIED TO GET UP TO
20 THOUSAND, IT'LL BE ONE CHANCE
IN TEN, TEN...
SORRY, IN TEN MILLION BILLION.
SO WHAT'S THE POINT OF ALL THAT?
WELL, THE FIRST POINT IS THAT AT
LEAST WE GOT SIMULATION WORKING,
SO THAT'S GOOD.
UM, THE SECOND POINT, IF I CAN
GET BACK TO OUR REGULARLY
SCHEDULED PROGRAMMING - THERE
WE GO - IS THAT SOMEHOW THIS
TINY LITTLE BIT EDGE IN FAVOUR
OF THE HOUSE, WHICH DIDN'T
MATTER SO MUCH IF YOU BET ONCE
OR EVEN IF YOU BET, SAY,
STARTING WITH A HUNDRED DOLLARS.
IT SURE STARTS TO MATTER AS YOU
BET MORE AND MORE AND MORE.
AND THIS IS RELATED TO WHAT WE
CALL THE LAW OF LARGE NUMBERS,
AND IT'S THIS IDEA THAT AS YOU
MAKE MORE AND MORE BETS, EVEN
THESE TINY LITTLE HOUSE EDGES
BECOME MORE AND MORE
SIGNIFICANT.
OR ANOTHER WAY OF THINKING ABOUT
IT IS THAT AS YOU DO THINGS MORE
AND MORE THEY TEND TO SETTLE
DOWN TO THEIR TRUE PROBABILITY,
SO IN THIS CASE, THE TRUE
PROBABILITY IS THAT THE HOUSE
HAS A SLIGHT EDGE.
AS YOU BET MORE AND MORE AND
MORE, IT'S GONNA BE MORE AND
MORE LIKELY THAT THE HOUSE IS
GONNA WIN.
SO THAT SHOULD BE OF NOTE IF YOU
GO TO THE CASINO, RIGHT?
'CAUSE IT SAYS, LOOK, IF YOU BET
A FEW TIMES, ANYTHING COULD
HAPPEN; BUT IF YOU KEEP ON
BETTING, FOR SURE YOU'RE GONNA
LOSE.
AND FROM THE CASINO'S POINT OF
VIEW, IF THERE ARE THOUSANDS AND
THOUSANDS OF PEOPLE MAKING
HUNDREDS AND HUNDREDS OF BETS
THE CASINO IS GUARANTEED TO WIN,
SO THAT'S THE POINT OF THIS AS
FAR AS THE CASINO'S CONCERNED.
SO THEN WE HAVE THIS LAW OF
LARGE NUMBERS, AND ONE POINT I
WANNA MAKE IS THAT THIS ISN'T
ONLY IMPORTANT WHEN YOU GO TO
THE CASINO, IT'S IMPORTANT IN A
LOT OF WAYS.
UM, LET ME FIRST SAY CASINOS
REALLY DO CARE ABOUT THIS STUFF
AND, IN FACT, THE LATEST THING--
OR NOT THE LATEST ANYMORE, BUT
IS, UH, ONLINE CASINOS AND EVERY
CASINO TRIES TO BE MORE EXCITING
THAN THE OTHER CASINO, BUT WELL,
THEY HIRE PEOPLE, IN FACT, I WAS
ONCE HIRED TO DO CONSULTING WORK
FOR AN ONLINE CASINO AND THEY
HAD ALL THESE NEW COMPLICATED
GAMES WITH DIFFERENT OPTIONS FOR
POKER AND SO ON, AND THEY WANTED
ME TO MAKE SURE THAT NO MATTER
WHICH OPTION YOU CHOSE AND NO
MATTER WHAT THE PLAYER DID, THE
ODDS WERE STILL ALWAYS IN THE
HOUSE'S FAVOUR.
I SHOULD SAY I FELT GUILTY
ENOUGH ABOUT THE WORK THAT I
DONATED PART OF MY FEE TO AN
ADDICTION RESEARCH FOUNDATION,
BUT, UH...

[Audience laughter]

Jeffrey continues BUT
NONETHELESS, THE POINT STANDS
THAT, YOU KNOW, THEY WANT THINGS
TO BE A LITTLE BIT IN THEIR
FAVOUR 'CAUSE THEY KNOW IN THE
LONG RUN THAT'S GONNA GUARANTEE
THEM A PROFIT.
AND ONCE YOU HAVE THIS IDEA
THAT, OKAY, OVER THE LONG RUN,
THINGS TEND TO SETTLE DOWN TO
THEIR TRUE PROBABILITIES.
WELL, SOMETIMES I THINK OF THAT
AS WHAT I CALL A PROBABILIST FORM
OF KARMA.
SO, PROBABLISTS DON'T TEND TO
BELIEVE IN THE USUAL FORM OF
KARMA, BUT YOU CAN THINK OF THE
LARGE, LARGE, NUMBERS AS KIND OF
BEING OUR OWN FORM AND IT SORTA
SAYS “WELL, OVER THE LONG RUN
THINGS ARE GONNA GET CLOSE TO
THEIR TRUE PROBABILITIES.”
LET ME MENTION A FEW OTHER
EXAMPLES OF THAT.
ONE IS THAT THERE'S THIS, UM,
THIS TRAFFIC LIGHT RIGHT NEAR MY
HOUSE, AND I LIKE TO THINK THAT
I'M A PRETTY RATIONAL PERSON,
BUT WHEN I DRIVE DOWNTOWN, IT'S
NOT SO CLEAR, AND I'M CONVINCED
THAT THIS TRAFFIC LIGHT, WHEN I
COME UP TO IT, I'M CONVINCED
IT'S ALWAYS RED.
I'M DRIVING ALONG, I'LL BE WITH
MY WIFE, I'LL SAY “LOOK, YOU SEE
THIS LIGHT, IT'S ALWAYS RED.”
SHE TURNS TO ME AND SAYS “WELL,
LOOK.
HAVEN'T YOU RED YOUR OWN BOOK?'

[Audience laughter]

Jeffrey continues 'CAUSE SHE
KNOWS THAT, YOU KNOW, THIS IS
ONE OF THOSE LIGHTS THAT'S RED
FOR ONE SIDE AS OFTEN AS IT'S
RED FOR THE OTHER SIDE, SO IT
MEANS IN THE LONG RUN IT SHOULD
BE RED FOR ME ABOUT HALF THE
TIME AND IT COULDN'T ALWAYS BE
RED FOR ME.
AND, IN FACT, WHAT'S REALLY
GOING ON THERE IS OBSERVATIONAL
BIAS, RIGHT?
WHEN THE LIGHT'S GREEN I JUST
DRIVE BY AND I DON'T GIVE IT A
SECOND THOUGHT; WHEN THE LIGHT'S
RED, BOY, DO I REMEMBER THAT,
RIGHT?
SO, MY RESEARCH HERE IS ABOUT
HOW TO USE RANDOMNESS IN ORDER
TO COMPUTE THINGS
SCIENTIFICALLY.
BUT LET ME FIRST SAY THAT
SOMEHOW RANDOMNESS IT'S VERY
USEFUL IN A LOT OF ENDEAVOURS
AND LET ME BEGIN AGAIN WITH A
PERSONAL STORY, IF I MAY, AND
THIS STORY GOES BACK TO WHEN I
WAS 25 YEARS OLD.

Text on the giant screen reads “Usefulness of randomness. Interpreting new evidence: lupus test.”

Jeffrey continues I WAS PRETTY HEALTHY, KIND OF
ATHLETIC, MUCH MORE SO THAN I AM
NOW.
AND I WENT FOR A ROUTINE MEDICAL
EXAM, AND THEN I WENT HOME, AND
A FEW DAYS LATER I GOT A LETTER
IN THE MAIL AND THE LETTER SAID
“DEAR MR. ROSENTHAL, THE TEST
FOR LUPUS INDICATES THAT YOU MAY
HAVE IT.”
AND IF YOU'RE A YOUNG HEALTHY 25
YEAR OLD AND YOU GET A LETTER
LIKE THAT, THAT'S PRETTY
DEVASTATING.
THEN IT SAID, OH, WELL “ABOUT
FIVE PERCENT OF THE NORMAL
POPULATION WILL TEST POSITIVE
FOR THIS TEST.”
IN OTHER WORDS, THEY HAD A FIVE
PERCENT FALSE-POSITIVE RATE.
SO YOU SHOULD COME IN FOR SOME
ADDITIONAL TESTS INVOLVING DNA
THAT WOULD GIVE A MORE ACCURATE
ANSWER.
SO, THE THING IS, IT WAS GONNA
TAKE ME A FEW DAYS TO GET TH
APPOINTMENT FOR THIS OTHER TEST,
AND THEN IT WAS GONNA TAKE ME A
FEW WEEKS BEFORE I'D GET THE
RESULTS BACK FROM THIS OTHER
TEST...
SO, FOR A FEW WEEKS THIS IS WHAT
I HAD, RIGHT.
I HAD THIS LETTER SAYING “THE
TEST FOR LUPUS INDICATES YOU MAY
HAVE IT.”
IT WAS PRETTY DEVASTATING, AND
YOU KNOW, I DIDN'T KNOW WHAT TO
DO REALLY, SO I SAID “WELL, WAIT
A MINUTE.
I WORK IN PROBABILITY THEORY.
MAYBE CAN THINK ABOUT THIS
AGAIN.”
AND AGAIN,
AS USUAL, JUST A LITTLE BIT OF
THINKING CAN GO A LONG WAY.
SO IN THIS CASE, FIRST OF ALL, I
LOOKED UP THE INCIDENCE OF LUPUS
IN THE, UH, GENERAL POPULATION,
AND IT'S A LITTLE LESS THAN ONE
PERCENT.
SO I SAY, OKAY, SO THAT MEANS
BEFORE THIS EXAM OR BEFORE THIS
TEST CAME BACK, I WOULD HAVE
SAID I HAD A BIT LESS THAN A ONE
PERCENT CHANCE OF HAVING LUPUS;
NOW THIS TEST HAS COME BACK.
WELL, NOW WHAT'S THE CHANCE I
HAVE IT?
WELL, THERE'S THIS FIVE PERCENT
OF THE NORMAL POPULATION THAT
WILL TEST POSITIVE FOR LUPUS, SO
AT FIRST IT SOUNDS LIKE MAYBE I
HAVE A 95 percent CHANCE OF HAVING
LUPUS.
WELL, THEN YOU SAY “WELL, WAIT
A MINUTE.”
SO IF FIVE PERCENT OF THE NORMAL
POPULATION WILL TEST POSITIVE,
PLUS ABOUT ONE PERCENT OF THE
POPULATION WILL HAVE LUPUS, THAT
MEANS MY CHANCES OF HAVING LUPUS
NOW ARE ACTUALLY ONE PERCENT
OVER SIX PERCENT, THAT IS, OF
THE SIX PERCENT YOU WILL GET A
POSITIVE TEST - FIVE PERCENT
PLUS ONE PERCENT.
WHAT'S THE CHANCE I'M PART OF
THAT ONE PERCENT WHO ACTUALLY
HAS LUPUS?
SO THE CHANCE WAS REALLY ONE OUT
OF SIX, OR ABOUT 17 percent.
NOW A 17 percent CHANCE OF HAVING
LUPUS, I CAN TELL YA, IT'S NOT
SUCH A GREAT THING, RIGHT?
BUT IT'S A LOT BETTER THAN A 95 percent
CHANCE OF HAVING LUPUS, AND IT
WAS ENOUGH TO CALM ME DOWN
ENOUGH THAT I COULD WAIT THOSE
FEW WEEKS, AND EVENTUALLY I GOT
THE TEST BACK SAYING THAT I DID
NOT ACTUALLY HAVE LUPUS, WHICH
WAS A GREAT RELIEF.
BUT THE POINT IS THAT, THAT
THINKING HAD HELPED ME QUITE A
BIT.
AND WHEN I TELL PEOPLE THIS
STORY, SOME PEOPLE SAY “WELL,
WAIT A MINUTE.
SO THAT LETTER FROM THE DOCTOR,
WHY DIDN'T
IT
EXPLAIN THAT
YOU REALLY ONLY HAD A 17 percent CHANCE
OF HAVING LUPUS?”
WELL, THAT'S A GOOD QUESTION,
ISN'T IT?
AND SO I HAD TO SORT OF THINK
ABOUT THAT, AND I THINK THERE'S
A TWO-FOLD ANSWER.
AND THE FIRST THING IS THAT
WELL, THE MEDICAL PROFESSION IS
NOT REALLY GEARED TOWARDS
ALLEVIATING ANXIETY.

[Audience laughter]

Jeffrey continues AND HERE
WE ARE IN THE MEDICAL SCIENCES
AUDITORIUM, BUT I STILL THINK I
CAN SAY THAT, RIGHT.
AND, UH, YOU KNOW, TO BE FAIR,
IT'S BECAUSE THEY HAVE OTHER
THINGS ON THEIR MIND OR OTHER
THINGS TO DO, BUT, AS WE ALREADY
SAID, ANXIETY COULD BE A BAD
THING TOO.
SO THAT'S THE FIRST ISSUE.
BUT THE SECOND ISSUE IS, I'M NOT
SURE IF THE DOCTOR KNEW I HAD A
17 percent CHANCE OF HAVING LUPUS,
'CAUSE PEOPLE DON'T REALLY THINK
IN TERMS OF PROBABILITIES LIKE
THAT.
SO, AGAIN, THE POINT IS THAT THE
WORLD IS NOT ALWAYS GONNA DO
THESE THINGS FOR US.
WE NEED TO JUST THINK A LITTLE
BIT AND THEN WE CAN, SO THAT'S
THE POINT I'M TRYING TO MAKE.
UM, LET ME QUICKLY MENTION SOME
OTHER USES OF RANDOMNESS
INCLUDING SOME THAT ARE IN THE
BOOK A LOT, BUT I WON'T TALK
ABOUT TODAY.
SO, IF YOU'RE DOING, LET'S SAY,
A SURVEY OR CONDUCTING A POLL
FOR AN ELECTION AND THE LAST
ELECTION WAS QUITE INTERESTING
FOR THE POLLS THAT WERE DONE,
UM...
IT'S VERY IMPORTANT TO HAVE, UM,
A RANDOM SAMPLE, RIGHT.
AND, IN A WAY, IT'S RELATED TO
THE GLOBE AND MAIL SURVEY, THE
ONLINE SURVEY THAT WE MENTIONED
AT THE BEGINNING.
AND, SO THE POINT IS, IF YOU
HAVE A NICE RANDOM SURVEY OF
CHOOSING PEOPLE EQUALLY AT
RANDOM THEN THAT'S FINE.
BUT IF YOU ONLY PICK PEOPLE WHO,
UH, YOU KNOW, GET THEIR NEWS
ONLINE ALREADY, THEN YOU'RE IN
TROUBLE.
UM, SO WITH POLLS, FOR EXAMPLE,
WELL, A GOOD POLLING COMPANY
WILL TRY TO EQUALLY LIKELY PHONE
ALL THE, YOU KNOW, THE ADULT
CITIZENS OF THE REGION, BUT,
WELL, THEY'LL HAVE SOME TROUBLE.
LIKE, FOR EXAMPLE, NOT EVERYONE
IS WILLING TO ANSWER THE PHONE
AND TALK TO POLLSTERS SO THEY'RE
REALLY JUST SAMPLING PEOPLE WHO
ARE WILLING TO TALK TO
POLLSTERS.
AND ALSO, NOW SOME PEOPLE JUST
HAVE A CELL PHONE AND TURNS OUT
THEY'RE NOT ALLOWED TO PHONE
PEOPLE WHO JUST HAVE A CELL
PHONE BECAUSE THE CELL PHONE
OWNER MIGHT HAVE TO PAY FOR THE
CALL, SO THOSE PEOPLE AREN'T
BEING SAMPLED EITHER.
SO, THERE'S A LOT OF ISSUES OF
USING A RANDOM SAMPLE, BUT
WITHOUT A RANDOM SAMPLE WE'D
HAVE NO WAY TO DO POLLS OR TO DO
SURVEYS.
ANYWAY, I WON'T SAY MORE ABOUT
THAT NOW - I'M JUST MAKING A BIT
OF A LIST.
HERE'S ONE THAT MAYBE NOT
EVERYONE THINKS ABOUT.
SO MOST OF US WILL BUY THINGS
OVER THE INTERNET, OR IN OTHER
WAYS WILL, UM, USE THE WORLD
WIDE WEB TO TRANSMIT SECURE
INFORMATION, AND THE WEBSITE IT
SAYS “OH, DON'T WORRY.
THIS IS OUR SECURE SERVER,
RIGHT, SO YOU CAN CHECK OUT ON
OUR SECURE SERVER AND GIVE US
YOUR CREDIT CARD NUMBER.”
WHAT DOES THAT MEAN, A “SECURE
SERVER?”
WELL, IT MEANS THAT THE
COMMUNICATION FROM YOUR COMPUTER
TO THE COMPUTER WHERE YOU'RE
BUYING THE THING IS SECURE, SO
IT'S, UH, ENCODED, AND THERE'S A
LOT TO SAY ABOUT HOW THE
ENCRYPTION WORKS AND THEIR
SO-CALLED PUBLIC KEY
CRYPTOGRAPHY AND SO ON, AND I
WON'T GET INTO IT EXCEPT TO SAY
THAT PART OF WHAT'S NEEDED ARE
RANDOM OR UNDETECTABLE STRINGS
OF USUALLY BITS, SO USUALLY, UH,
ZEROS AND ONES.
AND EVERY TIME YOU BUY SOMETHING
ON THE INTERNET, YOUR COMPUTER
DOES THE EQUIVALENT OF FLIPPING
A WHOLE BUNCH OF COINS, MAKING
SOME RANDOM ZEROS AND ONES IN
ORDER TO ALLOW IT TO ENCRYPT THE
INFORMATION SO IF SOME HACKER
INTERCEPTS THE, UH, STREAM OF
DATA, THEY STILL WON'T BE ABLE
SO, AGAIN, ION'T SAY MOREUMBER.
ABOUT IT NOW, BUT IT'S JUST AN
EXAMPLE OF HOW WE REALLY NEED
RANDOMNESS.
IT'S ALL OUT THERE.
UM, HERE'S ONE LITTLE FUN
EXAMPLE BEFORE I GET TO MY
RESEARCH STUFF, AND THIS ONE,
UM, IS MAYBE, AGAIN, APPROPRIATE
TO BE IN A LARGE CLASSROOM AT
THE UNIVERSITY OF TORONTO, AND
WE HAVE LOTS OF LARGE CLASSES
WITH LOTS OF STUDENTS IN THEM,
AND MOST OF THEM ARE WONDERFUL.
SOME OF THEM, A FEW, A VERY
SMALL NUMBER, HAVE BEEN KNOWN TO
CHEAT ON EXAMS.
I KNOW THIS COMES AS A SHOCK TO
ALL OF YOU, I'M SURE, BUT, UM...

[Audience laughter]

Jeffrey continues SO I KNOW
AN INSTRUCTOR WHO SAYS WELL, HE
WANTS TO DEAL WITH THIS BY
MAKING SURE THAT PEOPLE ARE
GONNA TRY TO COPY FROM EACH
OTHER THAT THEY DON'T SIT
TOGETHER, BUT HE DOESN'T KNOW
WHO'S FRIENDS WITH WHOM AND
WHO'S GONNA TRY TO COPY FROM
WHOM.
WELL, HE ASSIGNS A SEATING PLAN
RANDOMLY.
THAT MEANS THE PROBABILITY THAT
THOSE TWO PEOPLE WHO WANTED TO
SIT TOGETHER THAT THEY'RE
ACTUALLY TOGETHER, THE
PROBABILITY'S VERY SMALL.
SO, IT'S A NICE LITTLE EXAMPLE
OF USING RANDOMNESS.
SO WE SORT OF USE IT ALL THE
TIME.
ANOTHER USE OF RANDOMNESS, WHICH
I THINK IS KIND OF INTERESTING
IS RELATED TO THE GAME THEORY,
WHICH IS AS YOU'LL KNOW, IS THE
SCIENCE OF SORT OF HOW TO DEAL
WITH, WELL, WHETHER IT'S A GAME,
OR IT'S SORT OF INTERNATIONAL
RELATIONS, OR AWARD, OR BUSINESS
COMPETITIVE STRATEGIES, HOW TO
SORT OF DECIDE WHAT THE OTHER
GUY'S GONNA DO AND WHAT YOU'RE
GONNA DO.
WELL, IT TURNS OUT RANDOMNESS IS
A VERY IMPORTANT PART OF THAT,
AND ONE SIMPLE WAY TO THINK
ABOUT IT, AND LET ME JUST SAY
THIS QUICKLY, IS THAT, UM...
SO ROCK, PAPER, SCISSORS IS ONE
OF THOSE SILLY LITTLE CHILDREN'S
GAMES, RIGHT.
YOU KNOW THE ROCK SMASHES THE
SCISSORS AND THE SCISSORS CUTS
THE PAPER AND ALL THAT.
WELL, I USED TO THINK IT WAS
JUST A SILLY CHILDREN'S GAME.
IT TURNS OUT THERE'S AN ANNUAL
WORLD ROCK, PAPER, SCISSORS
CHAMPIONSHIP.

[Audience laughter]

Jeffrey continues AND SOME
OF YOU MAY KNOW, WITH THOUSANDS
OF DOLLARS IN PRIZES.
AND THE CHAMPION LAST YEAR WAS
QUOTE AS, HOW DID YOU DO IT?
HE SAYS “WELL, MY PERSONAL
STRATEGY WAS TO READ THE MINDS
OF MY COMPETITORS AND FIGURE OUT
WHAT THEY WERE THINKING.”

[Audience laughter]

Jeffrey continues SO,
DEPENDING ON WHAT MOVE I MAKE, I
EITHER THINK “OH, THAT'S JUST
CRAP, HE JUST GOT LUCKY.”
OR I THINK “WELL, IN THESE
COMPETITIONS THEY DO HAVE TO DO
IT OVER AND OVER AGAIN PRETTY
QUICKLY, AND YOU COULD ACTUALLY
START TO DETECT PATTERNS,
RIGHT?”
YOU SAY “OH, I SEE, YOU KNOW,
THIS GUY ALWAYS DOES THE SAME
THING I DID LAST TIME, OR IF
THIS GUY DID PAPER, HE ALWAYS
DOES ROCK THE NEXT TIME,” OR
SOMETHING LIKE THAT.
AND YOU MIGHT ACTUALLY BE ABLE
TO OUTSMART SOMEONE.
WELL, YOU CAN IF THEY HAPPEN TO
HAVE AN ORDINARY SIX-SIDED DIE
WITH THEM, AND IF THEY HAPPEN TO
SAY “LOOK, HERE'S WHAT I'M GONNA
DO.
I'M GONNA ROLL THE DIE EACH
TIME.
IF IT'S A ONOR A TWO I'LL SAY
ROCK, IF IT'S A THREE OR FOUR
I'LL SAY PAPER, IF IT'S A FIVE
OR SIX I'LL SAY SCISSORS.”
NOW [Unclear] MAY SAY, “WELL,
HOW'S THAT GONNA HELP?”
WELL, WHAT IT MEANS IS THAT
THERE'S NO WAY THE OTHER GUY CAN
OUTSMART THEM NO MATTER HOW MUCH
THEY LOOK INTO THEIR EYES OR
LOOK AT THEIR PAST STATISTICS OR
ANYTHING ELSE.
SO WE CALL THAT A RANDOMIZED
STRATEGY, AND IN THIS CASE, IT
GUARANTEES YOU A OVERALL DRAW.
THAT IS, YOU'LL WIN AS OFTEN IN
THE LONG RUN AS YOUR OPPONENT
WILL REGARDLESS OF HOW GOOD HE
IS AT THE GAME, RIGHT?
SO, WELL, YOU MIGHT THINK “WELL,
WHO CARES FOR ROCK, PAPER,
SCISSORS?”
BUT THE SAME THING APPLIES FOR
LOTS OF OTHER SORT OF
COMPETITIVE SITUATIONS OR HIDING
GAMES OR OTHER THINGS LIKE THAT,
WHERE WHEN YOU DON'T WANNA BE
OUTGUESSED OR OUTSMARTED, TURNS
OUT RANDOMNESS CAN HELP A LOT.
SO, ANYWAY, I COULD TALK MORE
ABOUT THAT, BUT I WANT TO HAVE
PLENTY OF TIME FOR A MONTE CARLO
ALGORITHMS WHICH ARE MY RESEARCH
AREA.
SO WHAT IS THE IDEA OF A MONTE
CARLO ALGORITHM?
WELL, IT WAS NAMED AFTER, OF
COURSE, THE CASINO IN MONACO,
AND IT'S THIS IDEA AGAIN OF
USING RANDOMNESS TO HELP US, BUT
IN THIS CASE, USING IT TO HELP
US WITH SERIOUS SCIENTIFIC, UH,
UH, SCIENTIFIC COMPUTATIONS, SO,
SO FORGET ABOUT ROCK, PAPER,
SCISSORS NOW; WE'RE ONTO
SOMETHING MORE IMPORTANT MAYBE.
BUT, UH...
SO, LET ME SAY SO FROM MY
PERSPECTIVE AS A STATISTICIAN,
WE OFTEN HAVE SORT OF BIG
COMPLICATED STATISTICAL MODELS -
THIS IS JUST FOR BACKGROUND -
SO, YOU KNOW, MAYBE WE WANNA
MEASURE THE AFFECT OF SOME
MEDICINE ON YOUR BLOOD PRESSURE
BUT IT DEPENDS ON YOUR, YOU
KNOW, YOUR AGE, AND IT DEPENDS
ON YOUR MEDICAL HISTORY, AND IT
DEPENDS ON YOUR UNKNOWN LITTLE
HEART CONDITION, AND DEPENDS
WHETHER YOU SMOKE, AND WE DON'T
JUST HAVE ONE PATIENT IN THE
STUDY, WE HAVE A THOUSAND
PATIENTS IN THE STUDY...
AND THERE'S A WHOLE BUNCH OF
VARIABLES, RIGHT?
EACH PATIENT AND EACH
CHARACTERISTIC AND SO ON.
SO, THE FIRST THING - THAT'S
JUST SORT OF FOR CONTEXT, IS
THAT WE APPLY THESE MONTE CARLO
ALGORITHMS TO BIG COMPLICATED
MODELS.
SO THINK OF JUST LOTS OF
VARIABLES AND LOTS OF UNKNOWNS
AND TRY TO MAKE SENSE OF IT ALL.
AND EVEN THOUGH COMPUTERS REALLY
FAST, THEY'RE STILL NOT FAST
ENOUGH - IF THERE'S ENOUGH
VARIABLES - TO MAKE SENSE OF IT
ALL.
SO HOW DOES THIS RANDOMNESS HELP
US TO COMPUTE?
THAT'S THE POINT OF MONTE CARLO
ALGORITHMS.
AND LET ME SET IT UP BY WAY OF
AN ANALOGY.
SO LET'S IMAGINE THAT YOU'RE,
UM, CLIMBING IN A MOUNTAIN RANGE
AND SOMEONE SAYS “SAY, WHAT'S
THE AVERAGE ALTITUDE OF THIS
MOUNTAIN RANGE?”
YOU MIGHT SAY “WELL, WHY DO YOU
CARE?”
BUT, SUPPOSE, YOU KNOW, SUPPOSE
YOU WANTED TO GIVE THEM A GOOD
ANSWER, YOU'D SAY “WELL, LET'S
SEE.
WELL, YOU KNOW, THAT MOUNTAIN UP
THERE, IT'S REALLY HIGH, BUT
WELL, THAT VALLEY DOWN THERE IS
PRETTY LOW, AND WELL, OVER THERE
THAT PLATEAU'S KIND OF IN THE
MIDDLE AND...”
HOW ARE YOU GONNA AVERAGE THIS
ALL UP?
WELL, SUPPOSE YOU HAD SOME
DEVICE LIKE A, LIKE A BAROMETER
THAT COULD ALLOW YOU TO MEASURE
THE ALTITUDE AT ANY GIVEN PLACE.
SO YOU SAY “OKAY, NOW I'M AT 382
METRES.”
WELL, BUT THEN YOU GO SOMEWHERE
ELSE, “OH, NOW I'M AT 512
METRES.”
BUT THEN YOU LOOK OUT AT THIS
BIG MOUNTAIN RANGE AND SAY HOW
ARE YOU GONNA AVERAGE THAT ALL
UP?
SO THAT'S A GOOD ANALOGY FOR
WHAT WE TRY TO DO WITH MONTE
CARLO ALGORITHMS.
AND WHAT YOU MIGHT THINK OF
DOING IS EXPLORING RANDOMLY,
SAY, “OKAY, HERE'S THE HEIGHT
OVER HERE, AND OVER HERE'S THE
HEIGHT OVER HERE, AND OUT HERE'S
THE HEIGHT UP THERE,” AND YOU
CAN THINK OF THAT AS KIND OF
TAKING A POLL, BUT IT'S NOT A
POLL OF PEOPLE, NOW IT'S A POLL
OF HEIGHTS OF DIFFERENT PLACES
ON THIS MOUNTAIN RANGE.
AND IF YOU COULD DO THAT AND GET
A NICE SAMPLE, WELL THEN YOU
COULD JUST AVERAGE IT UP, RIGHT?
YOU COULD SAY, WELL, WE'VE TAKEN
SOME RANDOM POINTS AND WE'VE
AVERAGED THEM.
THAT'S A PRETTY GOOD GUESS FOR
WHAT THE AVERAGE ALTITUDE OF
THIS MOUNTAIN RANGE IS.
SO IN PARTICULAR THE KIND OF,
UH, MONTE CARLO ALGORITHMS THAT
I WORK ON AND THAT HAVE BECOME
VERY POPULAR NOW ARE WHAT ARE
CALLED MARKOV CHAIN MONTE CARLO
ALGORITHMS, AND THEY'RE KIND OF
LIKE THE MOUNTAIN CLIMBER,
BECAUSE WHAT DOES A MOUNTAIN
CLIMBER DO?
WELL, WHERE HE GOES NEXT MIGHT
DEPEND ON WHERE HE WAS BEFORE,
RIGHT?
SO MAYBE HE'S GONNA TAKE
MEASUREMENTS PRETTY NEAR TO EACH
OTHER, BUT OVER TIME HE'S GONNA
TAKE MEASUREMENTS IN FARTHER AND
FARTHER PLACES.
AND THEN THE QUESTION BECOMES
“IS THIS REALLY GONNA BE A GOOD
WAY TO GET THIS AVERAGE?”
THAT IS, IS IT REALLY GONNA BE
TRUE THAT AS A MOUNTAIN CLIMBER
GOES AROUND TAKING THESE
MEASUREMENTS AND AVERAGING THEM
UP, HE'S GONNA GET A GOOD
ESTIMATE OF THE TRUE AVERAGE OF
THE ALTITUDE OF THIS MOUNTAIN
RANGE?
NOW, PROBABLY SOME OF YOU ARE
THINKING “BOY, WHAT AN OBSCURE
STRANGE TOPIC, THIS MARKOV CHAIN
MONTE CARLO.”
“I DON'T CARE IF IT'S HIS
RESEARCH AREA, DOES HE REALLY
HAVE TO TALK ABOUT IT TODAY?”
WELL...

[Audience laughter]

Jeffrey continues LET ME SAY IN MY DEFENCE THAT
ALTHOUGH THIS IS NOT WELL KNOWN
TO THE GENERAL PUBLIC, IT IS NOW
VERY COMMONLY USED IN A LOT OF
DIFFERENT SCIENCES.
SO IT'S USED IN STATISTICS ALL
THE TIME FOR COMPLICATED
MODELS...
IT'S USED
IN COMPUTER SCIENCE TO ENUMERATE
THE COMPLICATED, SORT OF,
STRUCTURE SPACES.
IT'S ALSO USED IN, FOR EXAMPLE,
STATISTICAL PHYSICS OR JUST, UH,
PHYSICAL CHEMISTRY TO FIGURE
OUT, YOU KNOW, DIFFERENT
POSSIBLE CONFIGURATIONS OF ATOMS
AND MOLECULES AND SO ON, SO,
WHENEVER THERE'S LOT OF
VARIABLES AROUND IT'S KIND OF,
UH, IT'S USED ACTUALLY QUITE A
BIT.
AND YOU MIGHT BE SCEPTICAL OF
THAT, BUT, YOU KNOW, I'M
PREPARED FOR DEFENCE BECAUSE,
YOU KNOW, I KNOW THE MODERN WAY
TO DECIDE HOW IMPORTANT
SOMETHING IS, AND THAT, TO DO IS
SEARCH IN GOOGLE AND SEE HOW
MANY HITS YOU GET.
AND IF YOU DO MARKOV CHAIN MONTE
CARLO IN QUOTES, YOU GET 872,000
HITS, SO, I'M NOT THE ONLY ONE
WHO THINKS THIS IS AN IMPORTANT
THING.
BUT ANYWAY, LET ME-- SO I'M
GONNA DO ANOTHER GRAPHICAL
SIMULATION HERE, AND THESE BLUE
BARS NOW REPRESENT-- YOU COULD
THINK OF IT AS REPRESENTING THE
HEIGHTS IN THE MOUNTAIN RANGE OR
YOU COULD THINK OF IT MORE
GENERALLY WITH SOME SORT OF A
BIG COMPLICATED MODEL.

A bar chart appears on the giant screen.

Jeffrey continues NOW HERE, WE ONLY HAVE SIX OF
THEM, BUT IMAGINE THERE'S, YOU
KNOW, THOUSANDS OR MILLIONS OR
EVEN AN INFINITE NUMBER OF THEM,
AND THEY REPRESENT THE
PROBABILITIES THAT WE WANT TO
SAMPLE FROM.
SO YOU CAN SEE, SAY-- AND I'M
GONNA BE A LITTLE VAGUE HERE.
I'M NOT GONNA GET INTO THE
ACTUAL EQUATIONS, BUT I'M GONNA
SAY WE WANT TO SAMPLE FROM
PROBABILITIES REPRESENTED BY
THESE BLUE BARS.
SO THE BLUE BARS, THE HEIGHT OF
THE BLUE BAR REPRESENTS SORT OF
HOW IMPORTANT THAT IS OR HOW
MUCH WE WANNA SAMPLE FROM IT.
NOW, OF COURSE, THIS IS JUST SIX
BARS SO WE COULD DO IT EASILY,
BUT IF WE IMAGINE THIS BIG AND
COMPLICATED, THEN WE SAY, WELL,
HERE'S WHAT WE'RE GONNA DO.
WE'RE GONNA TAKE SOME POINT TO
START WITH - AND HERE WE START
AT THE POINT FIVE - AND THEN
EACH TIME WE'RE GONNA EITHER ADD
ONE OR SUBTRACT ONE, SO WE'RE
GONNA JUST MOVE A LITTLE BIT,
AND THEN WE'RE EITHER GONNA
ACCEPT IT OR NOT, AND I WON'T
QUITE GET INTO DETAIL THERE, BUT
THIS TIME, AGAIN, IT'S RANDOM.
IT DECIDED TO TRY GOING TO THE
POINT FOUR.
IT ACCEPTED IT AND IT MOVED ON.
AND MAYBE I'LL JUST GET THIS
RUNNING.
SO IT'S GONNA REPEATEDLY CHOOSE
TO GO UP OR DOWN AND THEN
THERE'S A CERTAIN SIMPLE
EQUATION FOR-- THAT TIME IT
REJECTED IT, IT DIDN'T GO UP OR
DOWN.
OH, IT DID.
AGAIN, IT DIDN'T GO UP OR DOWN.
BUT IT'S GONNA HAVE SOME RULES
FOR WHERE IT GOES AROUND.
LET ME INTRODUCE ONE MORE THING
INTO THE PICTURE AND THAT'S
THESE BLACK BARS.

Black bars appear next to the blue bars.

Jeffrey continues THESE BLACK BARS REPRESENT WHAT
FRACTION OF TIME THIS ALGORITHM
HAS SPENT AT THAT LINE-- AT THAT
POINT SO FAR.
SO, SO FAR, WELL, THIS GUY HAS
SPENT A LOT OF TIME IN FOUR, SO
IT MEANS IT'S GOT A GREAT BIG
BLACK BAR SAYING MOST OF THE
TIME IT'S BEEN A FOUR.
A LITTLE BIT IT WAS A FIVE AND A
LITTLE BIT IT WAS THREE.
MOSTLY IT'S BEEN A FOUR AND IT
HASN'T BEEN AT SIX OR ONE OR TWO
AT ALL, YET.
WELL THAT DOESN'T SEEM SO
INTERESTING, BUT THEN AGAIN, THE
ISSUE IS WHAT HAPPENS IN THE
LONG RUN, SO LET ME, UM, SPEED
THINGS UP A BIT.
MAYBE I'LL SPEED IT UP A LITTLE
MORE TO SAVE TIME.
SO, WHAT'S GONNA HAPPEN OVER THE
LONG RUN IS THAT THE BLACK BARS
ARE GOING TO GET CLOSE TO THE
SAME HEIGHT AS THE BLUE BARS.
NOW THAT MIGHT NOT BE QUITE AS
THRILLING TO YOU AS IT IS TO ME...

[Audience laughter]

Jeffrey continues BUT LET ME
TRY TO EXPLAIN WHY THAT'S
IMPORTANT.
BECAUSE THE POINT IS THE BLUE
BARS WERE THE THING THAT WE
WANTED TO SAMPLE FROM.
THAT WAS THE RANDOMNESS THAT WE
WANTED, WHETHER IT WAS TO
ESTIMATE THE HEIGHT OF THAT
MOUNTAIN RANGE, OR WHETHER IT
WAS TO DO THIS BIG COMPLICATED
STUDY WITH LOTS OF VARIABLES,
BUT IT WAS TOO COMPLICATED.
WE JUST COULDN'T FIGURE OUT
THOSE BLUE BARS, BUT WE HAVE
SIMPLE WAYS OF RUNNING THESE
BLACK THINGS.
THAT IS, HAVING RULES FOR, YOU
KNOW, THE MOUNTAIN CLIMBER;
RULES FOR WHERE HE GOES NEXT AND
WHEN HE GOES UP AND WHEN HE GOES
DOWN AND SO ON.
THAT'S THE ESSENCE OF THESE
THINGS, IS THAT THERE'S THESE
SIMPLE LITTLE RULES FOR
RANDOMNESS THAT RUN OVER A LONG
PERIOD OF TIME LIKE THAT LAW OF
LARGE NUMBERS BEFORE.
OVER A LONG PERIOD OF TIME THEY
GET CLOSER AND CLOSER TO
COMPUTING THINGS THAT WERE JUST
TOO HARD TO COMPUTE.
SO THE BLUE BARS WERE JUST TOO
HARD TO COMPUTE BUT THE BLACK
BARS WE CAN RUN ON A COMPUTER
AND THEY CONVERGE TO THE SAME
THING.
SO, I'LL JUST LEAVE YOU WITH
THAT BECAUSE I KNOW IT'S NOT
GOING TO, UM, EXPLAIN THE WHOLE
STORY, BUT IT'S SOMETHING TO
THINK ABOUT AT LEAST.
I PRETTY MUCH GOTTA WRAP UP, BUT
LET ME JUST QUICKLY MENTION ONE
MORE THING.
UM, I WAS GONNA TALK ABOUT
COINCIDENCES.
I DON'T REALLY HAVE TIME, BUT
LET ME MENTION JUST ONE THING,
BECAUSE I'VE GOT A LITTLE
PICTURE HERE, SO...

A picture pops up on the screen of two sets of dots, with the title “Poisson clumping. Which dots are truly random, the left or the right?” The picture shows two boxes; the one on the left shows dots in an irregular pattern, with empty spaces and clumps. The one on the right shows a much more regular pattern, with no large empty spaces or clumps.

Jeffrey continues SO THIS HAS TO DO WITH HOW DOES
RANDOMNESS LOOK, IS ONE WAY TO
SAY IT.
OR, THE POINT IS THAT TRUE
RANDOMNESS IS THE VERSION WHERE
THEY'RE ALL CLUMPED UP.
SO WHAT DOES THAT MEAN?
THAT MEANS IF THINGS HAPPEN AT
RANDOM, AND WHETHER IT'S, UH,
WHEN IS THE NEXT HOMICIDE IN THE
CITY, OR WHEN YOU GET YOUR NEXT
WRONG NUMBER PHONE CALL, OR THE
NEXT TIME YOU SEE SOMEONE WHO'S
HAIR IS DYED PURPLE, OR ANYTHING
ELSE LIKE THAT, EVEN IF THEY'RE
HAPPENING RANDOMLY WITH NO
SPECIAL CAUSAL CONNECTION, IT'S
STILL GONNA BE THAT SOMETIMES
YOU'LL SEE, YOU KNOW, THREE
WRONG NUMBERS IN THE SAME NIGHT,
OR FIVE HOMICIDES IN THE SAME
WEEK, AND THEN THERE'LL BE OTHER
WEEKS WHEN YOU DON'T SEE ANY
HOMICIDES AT ALL, OR THERE'LL BE
MONTHS WHEN YOU GO BY WITHOUT
SEEING ANYONE WITH PURPLE HAIR,
AND THAT'S THE WAY THINGS
HAPPEN, SO THERE'S SORT OF A
SCIENTIFIC PRINCIPLE THERE, IT'S
CALLED “Poisson” CLUMPING AND
IT'S THIS IDEA THAT SOMEHOW
THINGS MAY APPEAR TO HAVE
PATTERNS WHEN THEY REALLY DON'T,
AND IF YOU'VE GOT LOTS OF
RANDOMNESS HAPPENING EVEN
WITHOUT ANY SPECIAL PATTERNS OR
INFLUENCE, YOU'LL HAPPEN TO FIND
CERTAIN PATTERNS IN IT, SO YOU
HAVE TO BE REALLY CAREFUL WHEN
YOU SEE PATTERNS WHETHER THEY
REALLY MEAN SOMETHING OR WHETHER
THEY DON'T.
THE POINT, REALLY, IF THERE'S A
SINGLE POINT OF THIS LECTURE AND
OF THE BOOK, IS THAT, UH, SORT
OF RANDOMNESS AND UNCERTAINTY.
WELL, THEY'RE WITH US
EVERYWHERE, RIGHT?
AND SOME PEOPLE MIGHT THINK “OH,
YOU KNOW, IF I PLAN THINGS
REALLY CAREFULLY, THEN MAYBE I
WON'T HAVE ANY RANDOMNESS.”
BUT YOU CAN'T PLAN THINGS THAT
WELL, RIGHT?
THERE'S ALWAYS GONNA BE PEOPLE
AND TRANSIT SYSTEMS AND WEATHER
SYSTEMS AND JOBS.
YOU'RE ALWAYS GONNA HAVE
RANDOMNESS.
RATHER THAN BEING AFRAID OF IT,
YOU CAN USE IT TO HELP YOU, TO
WORRY LESS ABOUT THINGS THAT ARE
UNLIKELY TO, UH, TO MAKE BETTER
DECISIONS AND UNDERSTAND THINGS
BETTER, AND IT'S ALSO BEING USED
MORE AND MORE, AS I SAY, FOR
SORT OF MODERN SCIENTIFIC
COMPUTATIONS.
SO, ANYWAY, I'LL STOP THERE.
THANK YOU VERY MUCH.

[Applause]

A black slate reads “Questions and answers.”

Jeffrey says SO THE
QUESTION IS IF YOU'RE, LET'S
SAY, A PHYSICIST OR A SCIENTIST
AND YOU'RE, UH, USING SOME MODEL
WITH VARIABLES AND YOU RUN MAYBE
SOME ALGORITHMS LIKE THE MONTE
CARLO ALGORITHM...
HOW DO YOU KNOW YOU'VE GOT ALL THE
VARIABLES, RIGHT?
OR HOW DO YOU KNOW YOU'VE GOT A
GOOD MODEL?
AND IT'S A GOOD QUESTION, AND I
MEAN, REALLY, ANY MATHEMATICAL
ANALYSIS OF, YOU KNOW, THE “REAL
WORLD” IN ANY SENSE ALWAYS
REALLY INVOLVES TWO DISTINCT
STEPS WHICH AREN'T ALWAYS
THOUGHT OF QUITE AS CLEARLY AS
THEY SHOULD BE IS, THE FIRST
STEP IS TO TRANSLATE THAT “REAL
WORLD” INTO SOMETHING WITH A
MATHEMATICAL MODEL WITH
EQUATIONS, AND THAT'S WHERE YOU
SAY, YOU KNOW, “WHAT VARIABLES
DO WE NEED, AND WHAT ARE THE
RELATIONS BETWEEN THEM, AND WHAT
CAN WE ASSUME, AND WHAT CAN WE
NOT ASSUME?”
AND THEN THE SECOND PART IS TO
SAY, “ONCE WE'VE GOT THIS
MATHEMATICAL MODEL, NOW HOW DO
WE DO THE COMPUTATIONS AND
FIGURE THINGS OUT?”
AND THE MONTE CARLO ALGORITHMS
ARE ALL FOR THE SECOND PART;
THEY'RE FOR FIGURING THINGS OUT
ONCE WE HAVE SOME SORT OF A
MATHEMATICAL MODEL.
THE FIRST PART, IT'S QUITE
SPECIFIC TO THE SUBJECT YOU'RE IN.
SO IF YOU'RE MODELLING PHYSICS
YOU BETTER UNDERSTAND A LOT OF
PHYSICS, OR IF YOU'RE MODELLING
ECONOMICS YOU BETTER UNDERSTAND
A LOT OF ECONOMICS, AND YOU CAN
ALWAYS DEBATE, AND DISCUSS, AND
REVISE THE MODEL, SO YOU'RE
ABSOLUTELY RIGHT.
THERE'S WHETHER THE MODEL'S
RIGHT, AND THEN THERE'S WHETHER
YOU'VE DONE THE RIGHT, UH,
COMPUTATIONS WITHIN THE MODEL.

A black slate reads “Probability and prices.”

A Man in the audience stands up. He’s in his late sixties and wears a white sweater and glasses.

He says WAS IT, UH, FEASIBLE
TO GIVE US SOME
UNDERSTANDING, UH, WHETHER IT'S
ECONOMICALLY VIABLE FOR THE CAR
OWNER...
TO DRIVE AROUND AT
NIGHT, FINDING THE LOWEST PRICE
FOR GASOLINE?

[Audience laughter]

Jeffrey says WELL, I
MEAN, I'VE NEVER DONE THAT
COMPUTATION, BUT, ACTUALLY, YOU
COULD, AND I MEAN, I'LL JUST
TAKE BIGGER POINT THAT I THINK
IT'D BE NICE IF WHEN PEOPLE DO
DECIDE WHETHER TO DRIVE AROUND
LOOKING FOR LOWER-PRICED GAS
THAT THEY START BY THINKING
ABOUT THE CALCULATION, RIGHT,
AND SAY “HOW MUCH AM I LIKELY TO
SAVE, AND HOW MUCH GAS AM I
LIKELY TO USE UP, AND HOW
VALUABLE IS MY TIME TO ME?”
AND YOU CAN PUT THESE THINGS ALL
TOGETHER, AND “WHAT'S THE CHANCE
I'LL FIND A PRICE THAT'S EVEN
LOWER?”
SO, ONE THING I DO ARGUE IN THE
BOOK IS THAT WHENEVER YOU'RE
MAKING A DECISION LIKE THAT, IF
YOU'VE GOT THE TIME AND THE
INCLINATION, STOP AND JUST MAKE
A LITTLE BIT OF AN ITEMIZATION.
WHAT ARE THE PROBABILITIES FOR
WHAT'S GONNA HAPPEN, AND WHAT
ARE YOUR, SORT OF, UH, UTILITIES
OR YOUR FEELINGS OF HOW MUCH YOU
CARE ABOUT THE RESULTS.
ABSOLUTELY, I HAVEN'T DONE THAT
ONE, AND IT MIGHT DEPEND HOW
VOLATILE THE GAS PRICES ARE.

A black slate reads “Lightning and lotteries.”

Another man from the audience, in his fifties, says
WE ALWAYS HEAR, UH, “YOU
HAVE A GREATER CHANCE OF BEING
HIT BY LIGHTNING THAN WINNING
THE LOTTERY.”
YET, EMPIRICALLY, I
WOULD QUEST-- UM, SOMEBODY BEING
HIT BY LIGHTNING, USUALLY IT'S A
NEWSWORTHY EVENT, SO YOU WOULD
SEE IT IN THE NEWSPAPER.
I DON'T SEE WEEKLY STORIES OF
PEOPLE BEING HIT BY LIGHTNING,
YET, GIVEN THAT THE LOTTERY'S
RESET USUALLY ALMOST EVERY WEEK,
SOMEBODY'S PROBABLY WON THE
LOTTERY EVERY WEEK.
SO, ALREADY HERE'S EMPIRICALLY,
THE RESULTS STILL SEEM TO
CONFORM TO THE STATISTICAL MODEL.

Jeffrey says REALLY,
THE ISSUE IS A BIT RELATED TO AN
EARLIER QUESTION BECAUSE THE
QUESTION IS, UM, THE PROBABILITY
OF BEING STRUCK BY LIGHTNING PER
WHAT, VERSUS WINNING THE LOTTERY
PER WHAT.
SO, IF YOU TALK ABOUT WINNING
THE LOTTERY WITH A SINGLE
TICKET, IT'S ABOUT ONE CHANCE IN
14 MILLION.
OF COURSE, THE MORE TICKETS YOU
BUY THE MORE IT GOES UP.
AND COMPARE THAT IT'S USUALLY
COMPARED TO THE CHANCE OF BEING
STRUCK BY LIGHTNING, LET'S SAY,
WITHIN THE NEXT YEAR...
NOW THE CHANCE OF ACTUALLY BEING
KILLED
BY A BOLT OF
LIGHTNING WITH I A ONE-YEAR
PERIOD FOR THE AVERAGE CANADIAN
IS ABOUT ONE CHANCE IN FIVE
MILLION, WHICH IS, ACTUALLY,
ALMOST THREE TIMES AS LIKELY AS
A CHANCE OF WINNING THE LOTTERY
JACKPOT WITH A SINGLE TICKET.
SO, THE SAYING IS ACTUALLY TRUE.
THE DIFFERENCES, THOUGH, IS
THAT, WELL, EACH OF US JUST GO
THROUGH ONE YEAR PER YEAR, BUT
ON AVERAGE PEOPLE BUY A LOT MORE
THAN ONE LOTTERY TICKET PER
YEAR, SO THAT'S THE DIFFERENCES.
IF YOU TAKE ONE TICKET VERSUS
ONE YEAR, THEN IT'S STILL TRUE.

A black slate reads “Exit polls.”

Jeffrey says IN THE
2004 AMERICAN ELECTION FOR
PRESIDENT, UM, IT'S QUITE
INTERESTING, 'CAUSE, YOU KNOW,
THE POLLS SHOWED QUITE A CLOSE
RACE, AND THAT IT WAS QUITE A
CLOSE RACE...
AND IN ADDITION, AS WAS MENTIONED,
THERE WERE, UM, SOME EXIT POLLS
AND, SO EXIT POLLS, YOU'LL KNOW,
ARE WHEN THE PEOPLE ARE VOTING,
AS THEY WALK OUT, THEY SAY “HEY,
EXCUSE ME, WHO DID YOU VOTE FOR?”
SOME PEOPLE AREN'T WILLING TO
SAY IN THE CONTEXT LIKE THAT WHO
THEY VOTED FOR, SOME ARE.”
THE EXIT POLLS SEEMED TO SHOW
THAT CAREY WAS DOING A LITTLE
BETTER THAN THE PRE-ELECTION
POLLS HAD PREDICTED, BUT THEN
THE FINAL CLAIM RESULT WAS VERY
CLOSE TO THE PRE-ELECTION POLLS
AND SHOWED A NARROW VICTORY FOR
BUSH, SO SOME PEOPLE HAVE SAID
“WELL, THOSE EXIT POLLS SHOW
THAT MAYBE THE FINAL COUNT
WASN'T ACCURATE AND REALLY CAREY
WON THE ELECTION.
NOW, IT'S CERTAINLY AN
INTERESTING DEBATE, AND IT IS
TRUE THAT IN THAT ELECTION SOME
OF THE, UM, POLLS WERE DONE BY,
UH, ELECTRONIC-- THEY HAD SORT
OF COMPUTER ELECTRONIC POLLS
WHICH DID NOT-- WERE NOT OPEN
SOURCE AND DIDN'T KEEP ALL THE
RECORDS, SO I THINK THERE'S
REASON TO SAY THAT IT SHOULD BE
A LITTLE MORE OF AN OPEN
PROCESS, BUT NONETHELESS, MY
READING OF THE NUMBERS, WHICH I
DID LOOK AT A FAIR BIT, IS THAT,
UM, THE FINAL RESULT WAS SO
CLOSE TO THE PRE-ELECTION POLLS
THAT THAT SEEMS PRETTY
CONVINCING TO ME, AND I THINK
THAT WHAT HAPPENED WAS THAT THE
EXIT POLLS WERE SOMETHING OF AN
ANOMALY BECAUSE, UH, THE
QUESTION OF WHO WAS WILLING TO
TALK TO THE POLLSTERS AND WHO
WASN'T, SO, ANYWAYS, IT'S AN
INTERESTING QUESTION THAT
WHENEVER AN ELECTION'S REALLY
CLOSE, IF YOU'RE A PROBABLIST
YOU'LL LOVE IT 'CAUSE ALL THE
POLLS GET REALLY INTERESTING,
BUT, I'LL LEAVE IT AT THAT.
THANK YOU VERY MUCH.

[Applause]

The clip ends and Andrew Moodie reappears in the studio.

He says THE LAW OF
LARGE NUMBERS IS ONE OF THE
CORNERSTONES OF PROBABILITY
ANALYSIS...
A STUDY THAT'S
RELATIVELY MODERN, STARTING WITH
THE WORK OF PIERRE FERMAT AND
BLAISE PASCAL IN 1654.
HOWEVER, EACH AND EVERY HUMAN
BEING HAS BEEN CONCERNED WITH
PROBABILITIES SINCE THE DAWN OF
TIME.
WHAT ARE THE CHANCES OF FINDING
A JOB IN A NEW CITY?
WHAT ARE THE CHANCES I CAN MAKE
IT TO WORK ON TIME IF I TAKE THE
GARDINER AT 3PM?
WE'VE ALL HEARD THE STATISTICS
ABOUT HOW TORONTO'S ACTUALLY
GETTING SAFER.
WHAT ROSENTHAL FAILED TO
MENTION, HOWEVER, IS THAT THE
ODDS OF DYING FROM GUN VIOLENCE
HAS ACTUALLY RISEN CONSISTENTLY
IF YOU'RE A YOUNG BLACK MALE.
CERTAIN PROBABILITIES CAN AND
MUST BE ALTERED BY OUR ACTIONS.
SPEAKING OF WHICH...

He shows the scratch lottery ticket again and says
HERE GOES NOTHING.
OKAY, COULD I GET A DRUM ROLL, PLEASE?

[Drum roll]

Andrew continues WITH ONE SCRATCH, MY LIFE COULD
BE CHANGED
FOREVER.
OR...
AND ACCORDING TO ALL ODDS, I
WILL INTRODUCE THE NEXT SEGMENT.

[Rim shot]

He scratches the ticket and says
SO, UH, WHAT IS THE PROBABILITY
OF AN EDGY Q and A SESSION FOLLOWING
THIS YEAR'S LAST MASSEY LECTURES
AT CONVOCATION HALL?
THE NAME OF THE LECTURER IS
MARGARET SOMERVILLE AND SHE IS
AGAINST GAY MARRIAGE.
IF YOU WERE TO SAY THAT THE ODDS
WERE OVERWHELMING, A SURE
THING...
[Chuckles]
YOU'D BE IN THE RIGHT.
SO HERE'S THE RENOWNED McGILL
ETHICIST, MARGARET SOMERVILLE,
THE 2007 MASSEY LECTURER IN A
Q and A SESSION.

In a full auditorium, a man in his twenties with long wavy black hair says
WITH THE RECENT
SOCIOLOGICAL CONSENSUS THAT
GENDER IS A CONSTRUCT OF OUR
SOCIETY, HOW DO YOU JUSTIFY YOUR
VIEWS ON SAME SEX MARRIAGE BEING
DETRIMENTAL TO THE RAISING OF
CHILDREN WHEN THERE'S NO
BIOLOGICAL CONNECTION BETWEEN
GENDER AND BIOLOGY, AND
THEREFORE NO REAL CONNECTION
BETWEEN THE ABILITY OF TWO MEN
OR TWO WOMEN TO RAISE A CHILD IN
RELATION TO A MAN AND A WOMAN TO
RAISING THAT CHILD?

[Applause]

Margaret Somerville stands behind an elaborate wooden podium on a stage. She’s in her fifties, with straight brown hair in a bob and bangs. She’s wearing glasses, a black cardigan and a red shirt.

A caption reads “Margaret Somerville. Founding Director, Centre for Medicine, Ethics and Law, McGill University. Convocation Hall, Toronto, October 27, 2006.”

She says UH, WITH
ALL RESPECT I DISAGREE WITH YOUR
PREMISE THERE.
I THINK THAT GENDER DOES MATTER,
AND COMPLIMENTARILY IN PARENTING
MATTERS AND INDEED SOME OF OUR
NEW RESEARCH, PARTICULARLY IN
GENETICS, IS SHOWING US THAT
THAT'S PROBABLY GOING TO BE EVEN
MORE TRUE THAN WE CURRENTLY
BELIEVE THAT IT IS.
AND THERE ISN'T A CONSENSUS THAT,
UH, GENDER DOES NOT MATTER, AND
I DON'T--
SEE, MY VIEW ON MARRIAGE IS THAT
IT'S NOT JUST A SOCIAL
CONSTRUCT.
I KNOW THAT IS THE VIEW THAT
SUPPORTS SAME SEX MARRIAGE.
MY VIEW ON MARRIAGE IS THAT
MARRIAGE CONSISTS OF A CORE
REALITY, A BIOLOGICAL REALITY,
NAMELY THE ABILITY TO TRANSMIT
LIFE TO THE NEXT GENERATION, AND
THAT WE HAVE-- WE HAVE A
SOCIETAL INSTITUTION AND WE
SURROUND THAT BIOLOGICAL REALITY
WITH CULTURE IN ORDER TO PROTECT
IT, AND MOST PARTICULARLY, TO
PROTECT THE CHILDREN THAT ARE
BORN FROM THAT BIOLOGICAL
REALITY, AND THAT'S MY--
MY PROBLEM WITH SAME SEX
MARRIAGE HAS GOT NOTHING TO DO
WITH BEING GAY OR HOMOSEXUAL,
AND AS IT HAPPENS, ALTHOUGH IT'S
NOT RELEVANT, I SUPPORT GAY
RIGHTS, BUT I THINK MARRIAGE
IS-- I ALSO SUPPORT CHILDREN'S
RIGHTS, AND THE REASON I OPPOSE
SAME SEX MARRIAGE IS IN THE
CONFLICT BETWEEN UPHOLDING SAME
SEX MARRIAGE PARTLY BECAUSE THE
GAY COMMUNITY WANTS IT, AND
SECONDLY BECAUSE THE STRONG CASE
FOR SAME SEX MARRIAGE IS THAT IT
WOULD BE A VERY POWERFUL
STATEMENT THAT DISCRIMINATION
AGAINST GAY PEOPLE IS HORRIBLE
AND WRONG WITH WHICH I AGREE
ENTIRELY.
THAT'S THE CASE FOR IT, BUT WHEN
I LOOK AT IT FROM AN ETHICAL
PERSPECTIVE, AND I DO A GENERAL
ETHICAL ANALYSIS, FOR EXAMPLE,
THE PRINCIPLE OF PROTECTING THE
MOST VULNERABLE PEOPLE INVOLVED
IN A SITUATION, THEN I THINK WE
HAVE TO GIVE CHILDREN PRIORITY,
AND I THINK THAT REQUIRES GIVING
THEM, IF AT ALL POSSIBLE, A
MOTHER AND A FATHER, PREFERABLY
THEIR OWN BIOLOGICAL PARENTS AND
THEIR WIDER BIOLOGICAL FAMILY,
AND BEING RAISED WITHIN THAT
CONTEXT, AND AS WELL, IT GIVES
THEM A RIGHT TO COME FROM
NATURAL BIOLOGICAL ORIGINS.
AND ONE OF THE PROBLEMS WITH
SAME SEX MARRIAGE IS THAT
MARRIAGE IS NOT A SINGLE RIGHT.
IT DOESN'T JUST RELATE TO TWO
ADULTS.
MARRIAGE IS A COMPOUND RIGHT
UNDER INTERNATIONAL LAW, FOR
EXAMPLE, THE UNIVERSAL
DECLARATION ON HUMAN RIGHTS.
IT'S A RIGHT TO MARRY AND FOUND
A FAMILY.
THEREFORE, IF YOU HAVE SAME SEX
MARRIAGE AND IT'S
UNCONSTITUTIONAL NOT TO HAVE IT,
YOU COULD-- YOU CAN ARGUE, AND
IT WILL BE ARGUED, THAT IT'S
UNCONSTITUTIONAL FOR ALLOWING
TWO SAME SEX PEOPLE TO HAVE
THEIR OWN SHARED GENETIC BABY,
WHICH WILL BECOME POSSIBLE.
IT'S RECENTLY BEEN DONE IN MICE
AND IT WILL BECOME POSSIBLE WITH
HUMANS.
WE'RE GOING TO BE ABLE TO MAKE
GAMETE SPERM FROM ADULT STEM
CELLS, FOR EXAMPLE, AND THERE
ARE OTHER WAYS IN WHICH IT COULD
OCCUR AS WELL.
SO THAT'S MY OPPOSITION.
THAT'S THE REASON FOR MY
OPPOSITION.

Another man in the audience rises. He’s in his late thirties, with short black hair.

He says UH, I RECENTLY CAME BACK
TO CANADA AND WHAT I FOUND WHEN
I REALIZED THAT CANADA HAD
LEGALIZED MARRIAGE FOR, UH,
PEOPLE LIKE ME, IS EXACTLY ONE
OF THE MOST BEAUTIFUL QUOTES YOU
HAD IN YOUR LECTURE TODAY.
UH, “A FEW VOICES CRY IN THE
ETHICAL WILDERNESS DO MATTER AND
CAN PREVAIL EVEN IN THE FACE OF
OVERWHELMING ODDS.”
THAT'S WHAT I BELIEVE.
[Applause]
SAME SEX MARRIAGE HAS
ACCOMPLISHED IN CANADA.
SO FINALLY I WAS ABLE TO GET
MARRIED RECENTLY, AND IT IS
AMAZING HOW I CAN PRESENT MYSELF
TO SOCIETY NOW, FEELING LIKE A
WHOLE-- A PART OF THE WHOLE, NOT
JUST...
COME ON, CIVIL UNION?
WHAT IS THAT?
YOU KNOW, LIKE, EVERYBODY KNOWS
WHAT MARRIAGE IS.
AND MY MOTHER WHO LIVES IN
BRAZIL UNDERSTANDS MARRIAGE AND
RESPECTS ME THAT MUCH MORE NOW
AND SPEAKS TO MY PARTNER BECAUSE
I'M MARRIED.
SO MY CONCERN ABOUT YOUR VOICE
WHEN YOU SAY YOU'RE DEFENDING
THE MOST VULNERABLE, IS THAT
YOU'RE NOT THINKING ABOUT THE
TEN PERCENT WHO ARE THE CHILDREN
OF EVERYBODY IN THIS ROOM, OKAY.
YOU'RE THINKING OF A MINUSCULE
PORTION OF THE POPULATION WHO
WILL BE BORN DUE TO THE EFFORTS
OF THIS VERY FANCY MEDICINE
YOU'RE TALKING ABOUT.
AND SO, WHAT I WOULD LIKE TO
SUGGEST TO YOU IS, AND TO POSE
TO YOU IS THIS QUESTION, YOUR
QUESTION.
MIGHT YOU HAVE PLACED YOUR
PSYCHE IN DENIAL ABOUT
COMPLEXITY AND IN DOING SO
CREATED A SOCIAL CLIMATE OF
ANXIETY, FEAR AND DEPRESSION.
WHICH IS WHAT I FIND IN MY
COMMUNITY THOSE OF US WHO FEEL
OPPRESSED WHEN YOU SO PUBLICLY
STAND AGAINST THOSE RIGHTS WE
HAVE FOUGHT SO HARD TO OBTAIN.

Margaret says I
UNDERSTAND YOUR POSITION IN THAT
RESPECT, AND YOU'VE JUST
PRESENTED THE STRONGEST CASE FOR
GAY MARRIAGE.
BUT DESPITE THAT, I STILL CAN'T
AGREE WITH IT BECAUSE, I MEAN, I
GENUINELY BELIEVE THAT KIDS NEED
A MOTHER AND A FATHER,
PREFERABLY THEIR OWN BIOLOGICAL
PARENTS.
AND I WOULD JUST POINT OUT TO
YOU, I'M NOT JUST PROTECTING
CHILDREN WHO WILL LATER ON BE
HETEROSEXUAL ADULTS.
ALL CHILDREN, INCLUDING
CHILDREN WHO GROW UP TO BE GAY
ADULTS, NEED A MOTHER AND A
FATHER.
SO I'M-- I ARGUE THAT FOR ALL
CHILDREN.
SO WHAT I'M REALLY DOING IS
GIVING CHILDREN THE PREFERENCE
AND NOT THE ADULTS, AND I THINK
WE CAN DO-- I KNOW THAT THE GAY
COMMUNITY DOESN'T LIKE THE CIVIL
UNION IDEA, BUT I THINK WE CAN
GO A LONG, LONG WAY TOWARDS
GIVING THE RESPECT, THE
PROTECTION, THE RIGHTS, THAT OF
CROUSE YOU DESERVE IN DOING
THAT, AND I DON'T KNOW WHETHER
YOU'VE SEEN THE LATEST REPORT
OUT OF THE FRENCH GOVERNMENT,
BUT THERE'S ABOUT A 200-PAGE
REPORT WHERE FRANCE HAD A BIG
GOVERNMENT COMMISSION TO LOOK
INTO WHETHER FRANCE SHOULD OR
SHOULD NOT LEGALIZE SAME SEX
MARRIAGE, AND THEY CAME TO THE
CONCLUSION THEY SHOULD HAVE WHAT
THEY CALL [French name,” WHICH
IS CIVIL UNION, AND NOT MARRIAGE
PRECISELY BECAUSE OF THE NEEDS
OF CHILDREN, AND ENGLAND HAS
JUST COME TO THE SAME
CONCLUSION, AND LIKEWISE,
ENGLAND'S GOT CIVIL UNIONS.
SO CANADA IS AN OUTLIER IN
HAVING SAME SEX MARRIAGE.
IT SHOULD HAVE CIVIL UNIONS, IS
MY OPINION.

Now another man from the audience rises. He’s in his twenties, with short blond hair.

He says YOU MENTION THAT, UM,
THE SEARCH FOR MEANING, THE
BY-PRODUCT WILL BE HAPPINESS.
YOU DON'T SEARCH FOR HAPPINESS.
IT'S ALMOST LIKE WE'VE TURNED A
PROCESS INTO A GOAL, BUT WHAT I
WANNA TALK ABOUT NOW IS, UM,
HOMOPHOBIA TO ME IS A FEAR OF
HOMOSEXUALS.
YOU ARE ARGUING AGAINST MARRIAGE
FOR FEAR OF CHILDREN'S RIGHTS
BEING DISPLACED BY GAY RIGHTS.
MY QUESTION IS THIS, AND I'LL
CONTINUE AFTER JUST VERY
BRIEFLY, IS HAVE YOU HEARD OF
SINGLE PARENT FAMILIES?
HAVE YOU HEARD OF HETEROSEXUAL
MARRIAGES WHICH DON'T EVEN WANNA
HAVE CHILDREN?
HAVE YOU HEARD OF ADOPTED
FAMILIES OR BLENDED SLASH
DIVORCED FAMILIES?
MARRIAGE - THE RIGHT TO MARRY
WHETHER IT'S A RIGHT OR NOT -
THE ACT HAS NOTHING TO DO WITH
THE ABILITY TO RAISE CHILDREN,
AND I THINK AS AN ETHE--
ETHICIST, YOU SHOULD KNOW
BETTER.

Margaret says WHAT I--
THE PROBLEM WITH SAME SEX
MARRIAGE IS, SAME SEX MARRIAGE
RADICALLY ALTERS THE BASIS OF
PARENTHOOD.
THE SECOND HALF OF THE CIVIL
MARRIAGE ACT WHICH IMPLEMENTED
SAME SEX MARRIAGE IN CANADA, UH,
A POINT ABOUT IT WHICH IS NOT
USUALLY DISCUSSED IS THAT IT
CHANGED THE DEFINITION OF
PARENTHOOD IN ALL FEDERAL
CANADIAN LEGISLATION FROM THE
TERM THAT HAD BEEN USED
PREVIOUSLY IN ALL THAT
LEGISLATION WAS THE PARENT IS
THE NATURAL PARENT OF THE CHILD,
AND IT CHANGED THAT TERM TO THE
LEGAL PARENT.
IN OTHER WORDS, A PARENT IS NOW
JUST SIMPLY WHO THE LAW SAYS
THEY ARE, NOT NECESSARILY THE
NATURAL PARENT, AND I THINK KIDS
HAVE GOT A RIGHT TO THEIR
NATURAL PARENTS, AND THAT'S WHY
I DISAGREE.

The man counters
NATURAL PARENTS ARE NOT
NECESSARILY BETTER PARENTS.

Margaret says BUT
WE'RE NOT TALKING-- WE'RE
TALKING ABOUT BASIC PRINCIPLES,
AND THEN THE-- ALL OF THE OTHER
THINGS YOU MENTIONED, YOU DON'T
GO INTO A MARRIAGE INTENDING TO
GET DIVORCED.
IF YOU DO, YOU SHOULDN'T BE
GOING INTO IT.
WHEREAS, HERE, YOU'RE GOING INTO
A RELATIONSHIP WHERE YOU KNOW IN
ADVANCE THAT THE CHILD WILL NOT
HAVE IT'S TWO NATURAL PARENTS
AND YOU WANT SOCIETY TO APPROVE
THAT.
WELL, I THINK THEY SHOULD
APPROVE THE UNION, BUT NOT
APPROVE THAT DEPRIVATION FOR
CHILDREN.
THAT'S MY ARGUMENT.

[Applause]

The clip ends.

Andrew reappears in the studio and says
TOMORROW, COME
BACK FOR THIS YEAR'S FINAL
MASSEY LECTURE.
A PLEA FOR AN ETHICS THAT
DOESN'T DO AWAY WITH NATURE.
ALL THE LECTURES ARE NOW
AVAILABLE AS A BOOK, THE ETHICAL
IMAGINATION.
I'M ANDREW MOODIE, AND I HOPE TO
SEE YOU TOMORROW.

[Theme music plays]

The end credits roll.

bigideas@tvo.org

416-484-2746

Big Ideas. Producer, Wodek Szemberg.

Producers, Lara Hindle, Mike Miner, Gregg Thurlbeck.

Logos: Unifor, Canadian Media Guild.

A production of TVOntario. Copyright 2006, The Ontario Educational Communications Authority.

Watch: Jeffrey Rosenthal on The Curious World of Probabilities