June 15, 2021 • 33 mins
Four years before Afrikka's Death, reporter Thomas Hargrove took clues from the Atlanta child murders and the Green River Killer and created an algorithm to detect serial killers. His algorithm flagged Gary, Indiana as the site of an unusual number of strangulations. But how did this algorithm work, and what came of the findings?
Thomas Hargrove discusses the creation of the algorithm and Dr. Mike Aamodt talks serial killing stats. Finally, host Ben Kuebrich receives an unexpected phone call.
Subscribe to Algorithm now, so you don't miss an episode and leave a review! This is a brand new show that could really use your help. Also, follow host Ben Kuebrich on Twitter for more information about the case.
Learn more about your ad-choices at https://www.iheartpodcastnetwork.com
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:01):
The views and opinions expressed in this podcast are solely
those of the authors and participants and do not necessarily
represent those of iHeart Media, Tenderfoot TV, or their employees.
This series contains discussions of violence and sexual violence. Listener
discretion is advised. Last time an algorithm. Lorie Townsend more
(00:24):
in the death of her daughter, Africa Hardy, don't think
that you know everything about child, because there's something that
they're not telling you. After moving to Chicago when she
was nineteen, Africa started escorting. In October, detectives found her
strangled in the bathtub of an Indiana motel. I just
(00:46):
think a lot of it, and I think it could
have been prevented. Years earlier, journalist Thomas Hargrove had learned
about the concept of linkage blindness. Most connected murders go unrecognized,
and I kept that in the back of my mind.
And then when Hargrove discovered an FBI database that tracked homicides,
(01:09):
it gave him an idea, could we teach a computer
to identify connected cases to find serial killings? From my
Heart Media and Tenderfoot TV, this is algorithm I'm ben
Keebrick I mentioned in the last episode that Africa's case
(01:35):
flipped what I knew about crime on its head this
episode you'll see why. Let's jump back to my conversation
with Thomas Hargrove. So, at what point in my um
you were saying that we shouldn't ignore her. Grove told
me that back in two thousand four, when he was
working as an investigative reporter in Washington, d C. He
(01:57):
had come across a database with information about homicides all
across the US. He wondered whether an algorithm could find
patterns within that data and to text serial killers. Maybe
it could even be used to detect active serial killers
who had not yet been caught. I did not know
(02:19):
that such an algorithm was possible. I was going on faith.
I did believe that it was important, and I did
believe that it could save lives, but I didn't know
that it would work. I begged my editors, let me
try this. Let's do a project looking at murder, with
the understanding that the actual goal is to see if
we could create a computer program that would identify serial murder.
(02:42):
But remember this was back in two thousand four, years
before people were talking about things like machine learning or
the power of big data. Because Harder was on the
cutting edge of these ideas, it was sometimes difficult to
get other people to understand him or to take him serious.
Lee My editors recognized that this could be something very cool,
(03:05):
that this could make news, but it's hard to commit
to a project you don't know upfront whether it's possible. Ultimately,
his editors told him that trying to design an algorithm
sounded too risky. It's normal for story pitches, especially ambitious
ones like this, to get rejected, but what's unusual is
(03:27):
that Hargrove didn't give up For the next six years.
He kept pitching the story six years, and he says
that his editors always considered it but would end up
assigning him to another safer story. Then, in two thousand ten,
Hargrove got his big break. He had just published a
(03:48):
provocative piece called Saving Babies, exposing sudden infant death. Hargrove
had analyzed data from Corners offices all across the US
and found that many times when infant deaths were listed
as sid's sudden infant death syndrome, the evidence actually showed
that the infants had died of accidental cufifocations. Most of
(04:12):
the time, babies don't die mysteriously. They die from avoidable
accidents and unsafe sleeping conditions. They got covered up because
of what was intended to be a very kindly diagnosis
called SIDS. That was meant to be a kindness to parents.
There was a mistake. This project prompted a national conference
(04:35):
to re examine infant death. It caused the creation of
a sudden Infant Death monitoring system by the CDC. I mean,
it was tremendously successful. It had got rave attention. Finally,
after six years, my stock was high enough in the
newsroom that I was able to get them to look
at murder. My editors that, okay, Tom, you've got a year.
(05:04):
I could do a national reporting project looking at unsolved murder,
with the understanding that what we were really about was
to try to develop a statistical means to identify overlooked
serial murder. Could we teach a computer to find serial killings?
(05:25):
After years of dreaming about making this algorithm, now Hargrove
had this chance to try and make it work. Hargrove
starting to design a computer program, one that could comb
through the five hundred thousand supplemental homicide reports and find
patterns between victims, patterns that might suggest the work of
(05:46):
a serial killer. So what is an algorithm? These days?
Algorithms dictate much of our digital lives. They determine what
TV shows get suggested to us, post we see on Facebook,
what route our GPS takes us on. Often these algorithms
(06:07):
are black boxes. We're not sure what data is being
fed into them and how they're deciding what to spit
back out. But at their core, algorithms aren't mysterious. They're
not even really that complicated. Algorithms are just sets of
instructions used to accomplish a specific task. You could say
a cake recipe is an algorithm for baking a cake.
(06:30):
You wouldn't say that, but you could. But while following
an algorithm might be simple, designing a whole new algorithm
is trickier. It wasn't like Cargroove was just baking a
cake from scratch. It was more like he was trying
to invent a brand new dessert. What we were doing
with the algorithm was the process literally of trial and error.
(06:52):
The ability to experiment with data was critical. If you're
trying to come up with a new dessert, might experiment
with ingredients until you come up with a combination that
tastes good. Hargrove knew what ingredients he had to play with.
That was the data from the supplemental homicide report, which
listed the weapon that was used he was shot with
(07:14):
a handgun, the time and location of the murder January,
and the age, race, and sex of the victim. Victim
is a blackmail eighteen years old. But as he experimented
with how an algorithm might sort through that data, Hargrove
needed a way to check if he was on the
right track. So we decided to test each new prototype
(07:35):
of his algorithm to see if it could detect a
known serial killer. We used as our test bed the
forty eight victims of serial killer Gary Ridgeway, so called
Green River Killer. He was convicted convicted in a court
of law of murdering forty eight girls and women in
(07:56):
the Seattle area. The question was, we'd teach a computer
a process that would tell us that something god awful
happened in Seattle. At the time, Gary Ridgeway was thought
to be the most prolific serial killer in America. Her
group figured if he could create an algorithm that could
detect Ridgeway. Maybe it would also detect other serial killers
(08:21):
that had gone unrecognized. To motivate himself, Hargrove stuck up
a picture of Gary Ridgeway in his office. It was
one of his booking photographs. He was glowering and looked
like it was glowering at me. Under that picture, I
typed the headline, what do serial victims look like? Statistically?
(08:45):
For over two decades, Ridgeway had killed women and dumped
their bodies along the Green River outside Seattle. In most cases,
by the time anyone had discovered those corpses, they were
already skeletal. Ridgeway has snuffed out does sense of lives.
He'd killed women and girls with their own hopes and dreams.
(09:06):
But Hargrove needed to approach their deaths like the algorithm would,
looking at their cases by their numbers, using only the
data from the FBI's supplemental homicide reports. So what did
ridgeways victims look like? Statistically? And more broadly, what do
a serial killer's victims look like? To find out, I
(09:28):
reached out to one of Hargrove's collaborators, Dr Mike A. Mott.
A Mott is a forensic psychology professor at Radford University,
and he curates the world's largest database of serial killers
and their victims. If you look at how serial killers
are portrayed in movies or on TV, the stereotypes not
really very consistent with with the back Are these also
(09:52):
stereotypes that law enforcement might have when they're trying to
decide could this murder possibly be a serial murder? Well,
law enforcement certainly has the same stereotypes. When we've done
presentations to law enforcement groups and even clinical psychologists that
are police psychologists, they're very surprised at the results. I'm
(10:27):
talking to Professor Mike al Mott about what serial murders
look like statistically, because statistical differences between serial killings and
other homicides are the kind of signatures and algorithm could
use to detect a serial killer. Professor Amatt has built
a database with information on thousands of serial killers, and
(10:49):
he told me that there are many misconceptions about these killers,
even within law enforcement and the true crime community. The
stereotypes about the profile of a serial killer or pro
file of the victims is not really very consistent with
the facts. If we want to catch serial killers, we
need to know who they really are in the reality
(11:09):
of their crimes. I want you to stop for a
moment and try to conjure up a mental image of
a typical serial killers victim. How old are they, what
do they look like? Now, imagine all of the victims
of a single killer. What would you guess they have
in common with one another? If we look at the
(11:31):
victims of serial killers, of the victims or women or
forty nine or men, so there's not really much of
a difference there. One of the stereotypes about serial killers
is that they have a type, you know, kinde of
victim that they're going to kill. The most consistent profile
we can have is really in the age category of
who they're killing. About are going to kill people that
(11:54):
are in the same age categories, so it's children, or
its teens, or its adults or elderly. Serial Killers are
less consistent when it comes to other attributes of their victims.
For example, only six kill victims there are all male
or all female. And then if you look at race,
for example, of serial killers only kill somebody of the
(12:17):
same race. And while the stereotypical victim of a serial
killer is white, actually a third of victims or people
of color. For Gary Ridgeway, the serial killer that Hargrove
picked as his test case, all his known victims were
young women, but they were diverse in terms of race.
In an interview with an FBI agent, Ridgeway says he
(12:40):
targeted prostitutes out of convenience, and studies show that there's
actually been a dramatic rise in serial killers targeting sex workers.
In the seventies, prostitutes are thought to have made up
around six of the female victims of serial killers. By
the two thousands, more than two thirds of women murdered
(13:01):
by serial killers were sex workers. Many serial killings are
sexual in nature. About one third of serial killers rape
at least one of their victims. Ridgeway told detectives he'd
have sex with his victims before he strangled them, and
when asked why he chose to choke all his victims,
Ridgeway replied because that was more personal and more rewarding
(13:24):
than shooting them. And compared to typical murderers, serial killers
are more likely to use methods that are up close
and personal, like strangulation or bludgending. So if we're looking
at victims in the US about were shot, were strangled
(13:46):
were stabbed and ten percent were blugend and serial killers
tend to be consistent in the method they used to kill.
Amont says that two thirds of serial killers use only
a single means to kill their victims. So what are
the takeaways from a MOOTS data? Serial killers show some
but not perfect consistency in terms of their victims age, race,
(14:10):
and sex, and also in their method of killing. But
it turns out that there's one more thing that serial
homicides tend to have in common, a property that would
be crucial for hard Groves algorithm. We spent the summer
of two thousand and ten finding at least the hundred
procedures that crashed and burned. Does the presence of a
(14:33):
serial killer increase the rated which women are murdered? No?
Does it increase the rate of which women are murdered
through unusual means? No? And do you get any indications
that you're getting closer? Yeah, So as we were progressing
through the hundred and one things that don't work, we
(14:54):
were starting to get closer. So um, the last thing
was to look at what the clearance rate was for
particular types of weapons. That term clearance rate refers to
the percentage of cases that police end up arresting someone
for the crime. Hard Group realized that Ridgeway had gotten
(15:16):
away with his murders for so long that they've been
listed as unsolved in the database. So if you looked
at killings in Seattle that matched his method, the number
of cases police had cleared was much lower than expected.
The presence of an active serial killer often destroys the
(15:36):
batting average for the local police department. They're able to
solve most murders, but not that type of murder because
there's a serial killer who's avoiding arrest hert group was
making progress. He'd picked up the hint of a signal
from Bridgeway in Seattle, but it wasn't enough to make
the algorithm useful as a tool. He racked his brain
(15:58):
trying to come up with a way to improve it.
One day, near the end of the summer, he looked
up at Ridgeway's mug shot and asked himself again, what
do a serial killers victims look like? Statistically? As he
talked through the project one last time with his research assistant,
Liz Lucas, an idea struck him. As I was taking
(16:21):
Liz to the airport because she had to go home
to defend her master's thesis. I told her that what
might work, and we're gonna try this next is a
kind of cluster analysis. Instead of querying all of the data,
we tried to assemble the data into smaller clusters according
to the county where the murderer occurred, the gender of
(16:43):
the victim, the method of killing, and at that time
age group. Then we calculated what the clearance rate was
for each cluster, how many of those murders were solved.
Up until this point, he tried looking at the data
for all the orders in Seattle. His new idea was
to further break down the data. Have the algorithm split
(17:06):
up the homicides for each county into buckets of victims,
and then rank these different buckets by the percentage of
cases that were solved. And Hargrove wasn't just doing this
for Seattle. He was doing this for data all across
the US, and he was hoping that out of the
thousands of clusters, Ridgeways Seattle killings would show up somewhere
(17:27):
near the top of the list. Hargrove ran his new
algorithm and waited anxiously for the result. When we did that,
the Green River killings jumped out plane as Day came
in third place most of the time seventies seven percent
of the time, and arrest is made when a woman
(17:49):
is killed in the cluster where the Green River killings
were grouped. The solution rate was less than and that
was our key. We're looking for groups of similar murders
that have very low clearance rates. We had hundreds of
results all over the country, highly suspicious clusters, and we
(18:12):
started investigating them. Heart Grove was a static. It seemed
(18:35):
like his algorithm might be working. It detected Gary Ridgeway,
the Green River killer, but at the same time, he
designed the algorithm to detect Ridgeway, and he kept tweaking
it until he did. So maybe the algorithm had just
detected Ridgeway by luck and it wouldn't generalize to other killers.
In statistics, this is a problem called overfitting, and it's
(18:58):
a common problem when scientists try to make algorithms that
predict things. So to get around this problem of overfitting,
people often train their algorithms with one set of data
and then validate the algorithm by testing it with new data. So,
since Hargrove had been using homicides in Seattle. To train
his algorithm, he needed to look at other cities to
(19:19):
see if it was really working. There were two larger
suspicious clusters, and in first place was Los Angeles. They
were a large group of almost entirely African American women
who were killed by handguns, and those murders had a
very low solution rate. So I assembled a spreadsheet of
(19:43):
those murders and emailed them to the public relations department
of the l a p D. Got one of the
representatives on the phone, is there a chance that any
of these could be serial murders? And he spent a
minute or so looking through the files, and then he
can back and said, what are you kidding? They're all
serial killings? I said, what. In fact, in the es
(20:08):
l ap D established what they hoped would be a
secret task force. They didn't want to alarm the public,
but they were exploring the possibility that there could be
a serial killer active. They called it the South Side
Slayer Task Force. Well, it turned out that the task
force was misnamed. It should have been called the South
(20:29):
Side Slayers Task Force. Because they had five they were
all quite independent of each other. They didn't know each
other but they were all killing women over a period
of twenty years. You heard that right. During the eighties
and nineties, Los Angeles had at least five different serial
killers who were shooting, strangling, and sexually assaulting women in
(20:52):
the area. This was happening at the height of the
crack epidemic and at a time when homicides across the
US where peaking. The large number of overall homicides probably
helped cover up the fact that there were these serial
killers operating in l A. But it's also likely that
the police didn't give these murders enough attention due to
(21:13):
a phenomena that's sometimes called victim discounting, and this is
the tendency to ignore crimes targeting marginalized groups. But when
Hargrove found out about these l A serial killers, he
took this as another sign that the algorithms seemed to
be working. It had detected a group of confirmed serial
killers that he hadn't even been aware of. Hargrove continued
(21:36):
down his list of top clusters that the algorithm had identified.
He called up the police department in Youngstown, Ohio, and
left a long message where he tried to explain the
algorithm and how it had detected a cluster of murders
in Youngstown. Thirty minutes later, the phone rings the chief
of detectives and you gotta give him credit, answering a
(21:57):
voicemail like that, he's a young guy. He called back
and said, that was the damnedest message I ever had
on my phone. And so I went back and I
interviewed my senior detectives and they told me something I
did not know. We thought we had a serial killer
in the nineties. We definitely thought we had one, and
we never got him, and so we started a new investigation.
(22:19):
He was attempting to locate the rape kids from those cases,
to try to DNA type all of the rape kits
they could find. Unfortunately, and this was very embarrassing, the
rape kits had all been destroyed, the property rooms had
been cleaned out, and he was not able to get
any of the kids from his cases or from surrounding jurisdictions.
(22:45):
There were other similar murders in neighboring jurisdictions, but they
too did not retain the rape kids. So it was
it was very sad, but they gave it the old
college try again. The algorithm had identified a cluster of
murders that seemed like it was the work of a
serial killer. Police couldn't prove that the killings were connected,
(23:06):
but the algorithms findings lined up with what police suspected,
and the algorithm had inspired a reinvestigation of those cold cases.
Hargrove felt like he was on the right track and
that maybe the algorithm could be a useful tool for
law enforcement. We selected ten major cities that appeared to
have a suspicious number of algorithm identified murders. Gary was
(23:31):
one of those ten. Gary, Indiana, the city right next
door to where Africa Hardy would be strangled. Four years later,
the algorithm flagged fifteen all unsolved murders in the Gary,
Indiana area. They were all women who were strangled. Not
one of the cases were solved, which is unusual. Seventy
(23:53):
seven percent of female murders get cleared, but not one
of these fifteen strangulation murder and Gary were cleared. I
called the public relations officer for the Gary Police Department,
gave him my name and said what we had found,
and said, is there a chance that you're dealing with
a serial killer? The next day the phone rings, it's
(24:17):
uh as I recall His name was Captain Roberts, who said, UM,
I've checked with our detectives and I can tell you
definitively that there are no unsolved serial murders in Gary, Indiana,
which is by definition an impossible statement to make. Unless
you have no unsolved murders, you cannot claim you'd be
(24:39):
definitively certain that there are no unsolved serial murders. Hargrove
felt like the police were just blowing him off, so
he started investigating the fifteen murders himself. Hargrove wanted to
know whether the algorithm was working, and he thought that
if he found evidence suggesting the murders were connected, maybe
the police would start take came seriously. But looking up
(25:02):
the information about the cases wasn't easy. The FBI supplemental
homicide report didn't include the victim's names or even the
exact date of their death. It just listed a month
in a year. So define more details about the cases,
Hargrove had to meticulously dig through old issues of local papers,
and as he started to piece things together, he was
(25:24):
unsettled by what he found. He was disturbed not just
by how grotesque these murders were, but by the patterns
that seemed to link them, and he became convinced that
Northwest Indiana had a serial killer on the loose, a
strangler who was targeting young women, women just like Africa Hardy.
He tried to talk to Gary's chief of police, but
(25:47):
he couldn't get through. I continued to suggest, have you
really looked at these cases? Soon the police stopped returning
his calls altogether, and Hargrove needed a response. He wasn't
just playing arm chair detective. We were about to publish
a story saying that Gary, Indiana has a serial killer
and the police would not talk about it. We were
(26:10):
afraid that the reason they weren't talking to us was
because they were hot on the heels of solving it.
That they had a suspect that we're trying to reel
them in, and we were going to screw that up.
We needn't have worried, but um, that was our fear.
I even sent registered letters to the chief of police
and to the mayor saying what we were about to do,
(26:32):
and if there's any issue they have, or any conversation
they want to have, for heaven's sakes, call me. This
is the letter that I wrote to the chief of
Police and Mayor in Gary, Indiana, and it goes dear Chief.
Carter Scripts Hour news service based in Washington, d C
(26:52):
is conducting a national reporting project looking into the thousand
unsolved thomicides committed in the United States since night. As
part of this project, we are investigating whether it's possible
to spot victims of serial murder among these unsolved killings.
Using the FBI's Supplementary Homicide Report. Using multivariate analysis, we've
(27:17):
determined that Gary, Indiana has an elevated number of unsolved
murders of women who were strangled in recent years. The
data that your department reported to the FBI are consistent
with the possibility that multiple victim killers have operated in
northwestern Indiana. Broadly, we see two possible patterns. In recent years,
(27:40):
several women have been strangled in their homes. In at
least two cases, of fire was set after the women
were killed. Also, starting in the nineteen nineties, we've seen
several women who were found strangled in or near abandoned buildings.
We doubt these killings are the result of convicted serial
(28:00):
killer Eugene V. Brit who admitted to killing eight people.
Please note the attached list of homicide victims. We'd be
grateful if your detectives would review these cases to determine
if any might have a common perpetrator. The US Justice
Department defines a serial killer simply as anyone who kills
(28:22):
two or more people in separate incidents. Experience thomicide investigators
tell us it's extremely difficult to spot a serial murderer.
There have been enough unsolved killings of women in Gary,
Indiana that your metropolitan area made our top ten list.
We are contacting authorities in all ten areas. Police and
(28:45):
five cities have already confirmed that cases we've cited contained
proven or suspected victims of serial murder. We are also
making a similar request to the Lake County Coroner's office.
Thank you for your time and consideration. Please help us
explore whether national crime data can assist local law enforcement. Sincerely,
(29:05):
Thomas Hargrove, Scripts Hour News service, and we had total
radio silence from those people. The Gary police chief and
mayor didn't respond to Hargrove's letter or to his follow
up phone calls, but Hargrove had also sent the letter
to the county corners corners or the public officials who
(29:26):
oversee autopsies and determined causes of death. They work with
the police, but they're an entirely separate entity. When I
called the Lake County Coroner's Office, identified myself, I'm Tom
Hargrove calling from Washington, d C. Oh just a minute,
Mr Hargrove, the senior deputy corner, wants to talk to you.
He came on and said, Mr Hargrove, we got your
(29:47):
packet of information. Thank you very much for sending it.
I'm assigning it to one of our assistant corners, a
lady named Jackie. We're going to have her look into it.
An entirely different reception them when I didn't get it Gary,
they agreed with us that there were too many unsolved murders.
She added three more cases that she thought belonged on
(30:08):
that pile we had identified fifteen. She added three, making
eighteen that she thought were connected and was trying to
have a conversation with the Carry Police department. She's never
gone on Mike to talk about this case. For the
Coroner's office, it is very, very difficult to speak ill
of a police department. It's considered bad form, and so
(30:31):
she probably still feels a reluctance to do that. Although
she has passion about this case, you shouldn't use this
recording where I named her unless she agrees. I was
looking up other names of people in that department, you know.
I think one went down for some kind of corruption charge,
and then it's kind of hard to find someone who
is there that can talk about this stuff. Yeah, now
(30:55):
your your only hope is to get Jackie to talk.
This is really one of the most frustrating experiences of
my life. I think the Gary Police Department should be
looking at some of those old cases. They still may
have a killer out there. When I finished speaking with
hard Grove, I tried calling Jackie, but I couldn't get through.
(31:16):
After a couple of failed attempts, I left her a
voicemail explaining the podcast how I was trying to look
into the murders Hargrove had identified to see if they
were indeed connected to Africa's death. Weeks passed, and I've
worked on other stories and chased down other leads, and
after what Hargrove had told me about her reluctance to
(31:37):
talk about the case, I didn't think she would want
to speak to me. Then one morning I woke up
to a missed call from a number I didn't recognize. Hey, Don,
this is Jackie. You would try to contact me a
while back in regard to heart Grow story you're doing. Yes,
(31:57):
you want to give me a call later next week? Um,
that would be fine. Sorry, had gotten back to you sooner.
It's just everything is fine, a little different. All right, guys,
I'll talk to you soon. Maybe that's coming next episode.
They don't want to talk about it either, I assume
(32:18):
because they don't want to embarrass their neighboring police agency.
All right, and said here and they had nothing to do.
They led to the death your friend. You should try
to find out. You'd be the first to do that.
Algorithm is released weekly on Tuesday's Subscribe Now so you
don't miss the next episode on the I Heart Radio app,
(32:40):
Apple Podcasts, or wherever you get your favorite shows. This
episode was written and produced by me ben Key. Brick.
Algorithm is executive produced by Alex Williams, Donald Albright, and
Matt Frederick. Production assistance in mixing by Eric Quintana. The
music is by Makeup and Vanity Set and Blue Dot Sessions.
(33:03):
Thanks to Christina Dana Miranda Hawkins. Jamie Albright, rema El Kaili,
Trevor Young, and Josh Thane for their help and notes. Again,
thanks for listening as it heads up. I'm still working
on this podcast as we release it, so any feedback
is appreciated. I think Algorithm is going to address some
(33:24):
really important issues about policing and how crimes are investigated
that don't receive enough attention. So if you can, please
leave a review on Apple Podcasts or tell a friend
about Algorithm, where brand new show and could really use
your help.
The views and opinions expressed in this podcast are solely
those of the authors and participants and do not necessarily
represent those of iHeart Media, Tenderfoot TV, or their employees.
This series contains discussions of violence and sexual violence. Listener
discretion is advised. Last time an algorithm. Lorie Townsend more
(00:24):
in the death of her daughter, Africa Hardy, don't think
that you know everything about child, because there's something that
they're not telling you. After moving to Chicago when she
was nineteen, Africa started escorting. In October, detectives found her
strangled in the bathtub of an Indiana motel. I just
(00:46):
think a lot of it, and I think it could
have been prevented. Years earlier, journalist Thomas Hargrove had learned
about the concept of linkage blindness. Most connected murders go unrecognized,
and I kept that in the back of my mind.
And then when Hargrove discovered an FBI database that tracked homicides,
(01:09):
it gave him an idea, could we teach a computer
to identify connected cases to find serial killings? From my
Heart Media and Tenderfoot TV, this is algorithm I'm ben
Keebrick I mentioned in the last episode that Africa's case
(01:35):
flipped what I knew about crime on its head this
episode you'll see why. Let's jump back to my conversation
with Thomas Hargrove. So, at what point in my um
you were saying that we shouldn't ignore her. Grove told
me that back in two thousand four, when he was
working as an investigative reporter in Washington, d C. He
(01:57):
had come across a database with information about homicides all
across the US. He wondered whether an algorithm could find
patterns within that data and to text serial killers. Maybe
it could even be used to detect active serial killers
who had not yet been caught. I did not know
(02:19):
that such an algorithm was possible. I was going on faith.
I did believe that it was important, and I did
believe that it could save lives, but I didn't know
that it would work. I begged my editors, let me
try this. Let's do a project looking at murder, with
the understanding that the actual goal is to see if
we could create a computer program that would identify serial murder.
(02:42):
But remember this was back in two thousand four, years
before people were talking about things like machine learning or
the power of big data. Because Harder was on the
cutting edge of these ideas, it was sometimes difficult to
get other people to understand him or to take him serious.
Lee My editors recognized that this could be something very cool,
(03:05):
that this could make news, but it's hard to commit
to a project you don't know upfront whether it's possible. Ultimately,
his editors told him that trying to design an algorithm
sounded too risky. It's normal for story pitches, especially ambitious
ones like this, to get rejected, but what's unusual is
(03:27):
that Hargrove didn't give up For the next six years.
He kept pitching the story six years, and he says
that his editors always considered it but would end up
assigning him to another safer story. Then, in two thousand ten,
Hargrove got his big break. He had just published a
(03:48):
provocative piece called Saving Babies, exposing sudden infant death. Hargrove
had analyzed data from Corners offices all across the US
and found that many times when infant deaths were listed
as sid's sudden infant death syndrome, the evidence actually showed
that the infants had died of accidental cufifocations. Most of
(04:12):
the time, babies don't die mysteriously. They die from avoidable
accidents and unsafe sleeping conditions. They got covered up because
of what was intended to be a very kindly diagnosis
called SIDS. That was meant to be a kindness to parents.
There was a mistake. This project prompted a national conference
(04:35):
to re examine infant death. It caused the creation of
a sudden Infant Death monitoring system by the CDC. I mean,
it was tremendously successful. It had got rave attention. Finally,
after six years, my stock was high enough in the
newsroom that I was able to get them to look
at murder. My editors that, okay, Tom, you've got a year.
(05:04):
I could do a national reporting project looking at unsolved murder,
with the understanding that what we were really about was
to try to develop a statistical means to identify overlooked
serial murder. Could we teach a computer to find serial killings?
(05:25):
After years of dreaming about making this algorithm, now Hargrove
had this chance to try and make it work. Hargrove
starting to design a computer program, one that could comb
through the five hundred thousand supplemental homicide reports and find
patterns between victims, patterns that might suggest the work of
(05:46):
a serial killer. So what is an algorithm? These days?
Algorithms dictate much of our digital lives. They determine what
TV shows get suggested to us, post we see on Facebook,
what route our GPS takes us on. Often these algorithms
(06:07):
are black boxes. We're not sure what data is being
fed into them and how they're deciding what to spit
back out. But at their core, algorithms aren't mysterious. They're
not even really that complicated. Algorithms are just sets of
instructions used to accomplish a specific task. You could say
a cake recipe is an algorithm for baking a cake.
(06:30):
You wouldn't say that, but you could. But while following
an algorithm might be simple, designing a whole new algorithm
is trickier. It wasn't like Cargroove was just baking a
cake from scratch. It was more like he was trying
to invent a brand new dessert. What we were doing
with the algorithm was the process literally of trial and error.
(06:52):
The ability to experiment with data was critical. If you're
trying to come up with a new dessert, might experiment
with ingredients until you come up with a combination that
tastes good. Hargrove knew what ingredients he had to play with.
That was the data from the supplemental homicide report, which
listed the weapon that was used he was shot with
(07:14):
a handgun, the time and location of the murder January,
and the age, race, and sex of the victim. Victim
is a blackmail eighteen years old. But as he experimented
with how an algorithm might sort through that data, Hargrove
needed a way to check if he was on the
right track. So we decided to test each new prototype
(07:35):
of his algorithm to see if it could detect a
known serial killer. We used as our test bed the
forty eight victims of serial killer Gary Ridgeway, so called
Green River Killer. He was convicted convicted in a court
of law of murdering forty eight girls and women in
(07:56):
the Seattle area. The question was, we'd teach a computer
a process that would tell us that something god awful
happened in Seattle. At the time, Gary Ridgeway was thought
to be the most prolific serial killer in America. Her
group figured if he could create an algorithm that could
detect Ridgeway. Maybe it would also detect other serial killers
(08:21):
that had gone unrecognized. To motivate himself, Hargrove stuck up
a picture of Gary Ridgeway in his office. It was
one of his booking photographs. He was glowering and looked
like it was glowering at me. Under that picture, I
typed the headline, what do serial victims look like? Statistically?
(08:45):
For over two decades, Ridgeway had killed women and dumped
their bodies along the Green River outside Seattle. In most cases,
by the time anyone had discovered those corpses, they were
already skeletal. Ridgeway has snuffed out does sense of lives.
He'd killed women and girls with their own hopes and dreams.
(09:06):
But Hargrove needed to approach their deaths like the algorithm would,
looking at their cases by their numbers, using only the
data from the FBI's supplemental homicide reports. So what did
ridgeways victims look like? Statistically? And more broadly, what do
a serial killer's victims look like? To find out, I
(09:28):
reached out to one of Hargrove's collaborators, Dr Mike A. Mott.
A Mott is a forensic psychology professor at Radford University,
and he curates the world's largest database of serial killers
and their victims. If you look at how serial killers
are portrayed in movies or on TV, the stereotypes not
really very consistent with with the back Are these also
(09:52):
stereotypes that law enforcement might have when they're trying to
decide could this murder possibly be a serial murder? Well,
law enforcement certainly has the same stereotypes. When we've done
presentations to law enforcement groups and even clinical psychologists that
are police psychologists, they're very surprised at the results. I'm
(10:27):
talking to Professor Mike al Mott about what serial murders
look like statistically, because statistical differences between serial killings and
other homicides are the kind of signatures and algorithm could
use to detect a serial killer. Professor Amatt has built
a database with information on thousands of serial killers, and
(10:49):
he told me that there are many misconceptions about these killers,
even within law enforcement and the true crime community. The
stereotypes about the profile of a serial killer or pro
file of the victims is not really very consistent with
the facts. If we want to catch serial killers, we
need to know who they really are in the reality
(11:09):
of their crimes. I want you to stop for a
moment and try to conjure up a mental image of
a typical serial killers victim. How old are they, what
do they look like? Now, imagine all of the victims
of a single killer. What would you guess they have
in common with one another? If we look at the
(11:31):
victims of serial killers, of the victims or women or
forty nine or men, so there's not really much of
a difference there. One of the stereotypes about serial killers
is that they have a type, you know, kinde of
victim that they're going to kill. The most consistent profile
we can have is really in the age category of
who they're killing. About are going to kill people that
(11:54):
are in the same age categories, so it's children, or
its teens, or its adults or elderly. Serial Killers are
less consistent when it comes to other attributes of their victims.
For example, only six kill victims there are all male
or all female. And then if you look at race,
for example, of serial killers only kill somebody of the
(12:17):
same race. And while the stereotypical victim of a serial
killer is white, actually a third of victims or people
of color. For Gary Ridgeway, the serial killer that Hargrove
picked as his test case, all his known victims were
young women, but they were diverse in terms of race.
In an interview with an FBI agent, Ridgeway says he
(12:40):
targeted prostitutes out of convenience, and studies show that there's
actually been a dramatic rise in serial killers targeting sex workers.
In the seventies, prostitutes are thought to have made up
around six of the female victims of serial killers. By
the two thousands, more than two thirds of women murdered
(13:01):
by serial killers were sex workers. Many serial killings are
sexual in nature. About one third of serial killers rape
at least one of their victims. Ridgeway told detectives he'd
have sex with his victims before he strangled them, and
when asked why he chose to choke all his victims,
Ridgeway replied because that was more personal and more rewarding
(13:24):
than shooting them. And compared to typical murderers, serial killers
are more likely to use methods that are up close
and personal, like strangulation or bludgending. So if we're looking
at victims in the US about were shot, were strangled
(13:46):
were stabbed and ten percent were blugend and serial killers
tend to be consistent in the method they used to kill.
Amont says that two thirds of serial killers use only
a single means to kill their victims. So what are
the takeaways from a MOOTS data? Serial killers show some
but not perfect consistency in terms of their victims age, race,
(14:10):
and sex, and also in their method of killing. But
it turns out that there's one more thing that serial
homicides tend to have in common, a property that would
be crucial for hard Groves algorithm. We spent the summer
of two thousand and ten finding at least the hundred
procedures that crashed and burned. Does the presence of a
(14:33):
serial killer increase the rated which women are murdered? No?
Does it increase the rate of which women are murdered
through unusual means? No? And do you get any indications
that you're getting closer? Yeah, So as we were progressing
through the hundred and one things that don't work, we
(14:54):
were starting to get closer. So um, the last thing
was to look at what the clearance rate was for
particular types of weapons. That term clearance rate refers to
the percentage of cases that police end up arresting someone
for the crime. Hard Group realized that Ridgeway had gotten
(15:16):
away with his murders for so long that they've been
listed as unsolved in the database. So if you looked
at killings in Seattle that matched his method, the number
of cases police had cleared was much lower than expected.
The presence of an active serial killer often destroys the
(15:36):
batting average for the local police department. They're able to
solve most murders, but not that type of murder because
there's a serial killer who's avoiding arrest hert group was
making progress. He'd picked up the hint of a signal
from Bridgeway in Seattle, but it wasn't enough to make
the algorithm useful as a tool. He racked his brain
(15:58):
trying to come up with a way to improve it.
One day, near the end of the summer, he looked
up at Ridgeway's mug shot and asked himself again, what
do a serial killers victims look like? Statistically? As he
talked through the project one last time with his research assistant,
Liz Lucas, an idea struck him. As I was taking
(16:21):
Liz to the airport because she had to go home
to defend her master's thesis. I told her that what
might work, and we're gonna try this next is a
kind of cluster analysis. Instead of querying all of the data,
we tried to assemble the data into smaller clusters according
to the county where the murderer occurred, the gender of
(16:43):
the victim, the method of killing, and at that time
age group. Then we calculated what the clearance rate was
for each cluster, how many of those murders were solved.
Up until this point, he tried looking at the data
for all the orders in Seattle. His new idea was
to further break down the data. Have the algorithm split
(17:06):
up the homicides for each county into buckets of victims,
and then rank these different buckets by the percentage of
cases that were solved. And Hargrove wasn't just doing this
for Seattle. He was doing this for data all across
the US, and he was hoping that out of the
thousands of clusters, Ridgeways Seattle killings would show up somewhere
(17:27):
near the top of the list. Hargrove ran his new
algorithm and waited anxiously for the result. When we did that,
the Green River killings jumped out plane as Day came
in third place most of the time seventies seven percent
of the time, and arrest is made when a woman
(17:49):
is killed in the cluster where the Green River killings
were grouped. The solution rate was less than and that
was our key. We're looking for groups of similar murders
that have very low clearance rates. We had hundreds of
results all over the country, highly suspicious clusters, and we
(18:12):
started investigating them. Heart Grove was a static. It seemed
(18:35):
like his algorithm might be working. It detected Gary Ridgeway,
the Green River killer, but at the same time, he
designed the algorithm to detect Ridgeway, and he kept tweaking
it until he did. So maybe the algorithm had just
detected Ridgeway by luck and it wouldn't generalize to other killers.
In statistics, this is a problem called overfitting, and it's
(18:58):
a common problem when scientists try to make algorithms that
predict things. So to get around this problem of overfitting,
people often train their algorithms with one set of data
and then validate the algorithm by testing it with new data. So,
since Hargrove had been using homicides in Seattle. To train
his algorithm, he needed to look at other cities to
(19:19):
see if it was really working. There were two larger
suspicious clusters, and in first place was Los Angeles. They
were a large group of almost entirely African American women
who were killed by handguns, and those murders had a
very low solution rate. So I assembled a spreadsheet of
(19:43):
those murders and emailed them to the public relations department
of the l a p D. Got one of the
representatives on the phone, is there a chance that any
of these could be serial murders? And he spent a
minute or so looking through the files, and then he
can back and said, what are you kidding? They're all
serial killings? I said, what. In fact, in the es
(20:08):
l ap D established what they hoped would be a
secret task force. They didn't want to alarm the public,
but they were exploring the possibility that there could be
a serial killer active. They called it the South Side
Slayer Task Force. Well, it turned out that the task
force was misnamed. It should have been called the South
(20:29):
Side Slayers Task Force. Because they had five they were
all quite independent of each other. They didn't know each
other but they were all killing women over a period
of twenty years. You heard that right. During the eighties
and nineties, Los Angeles had at least five different serial
killers who were shooting, strangling, and sexually assaulting women in
(20:52):
the area. This was happening at the height of the
crack epidemic and at a time when homicides across the
US where peaking. The large number of overall homicides probably
helped cover up the fact that there were these serial
killers operating in l A. But it's also likely that
the police didn't give these murders enough attention due to
(21:13):
a phenomena that's sometimes called victim discounting, and this is
the tendency to ignore crimes targeting marginalized groups. But when
Hargrove found out about these l A serial killers, he
took this as another sign that the algorithms seemed to
be working. It had detected a group of confirmed serial
killers that he hadn't even been aware of. Hargrove continued
(21:36):
down his list of top clusters that the algorithm had identified.
He called up the police department in Youngstown, Ohio, and
left a long message where he tried to explain the
algorithm and how it had detected a cluster of murders
in Youngstown. Thirty minutes later, the phone rings the chief
of detectives and you gotta give him credit, answering a
(21:57):
voicemail like that, he's a young guy. He called back
and said, that was the damnedest message I ever had
on my phone. And so I went back and I
interviewed my senior detectives and they told me something I
did not know. We thought we had a serial killer
in the nineties. We definitely thought we had one, and
we never got him, and so we started a new investigation.
(22:19):
He was attempting to locate the rape kids from those cases,
to try to DNA type all of the rape kits
they could find. Unfortunately, and this was very embarrassing, the
rape kits had all been destroyed, the property rooms had
been cleaned out, and he was not able to get
any of the kids from his cases or from surrounding jurisdictions.
(22:45):
There were other similar murders in neighboring jurisdictions, but they
too did not retain the rape kids. So it was
it was very sad, but they gave it the old
college try again. The algorithm had identified a cluster of
murders that seemed like it was the work of a
serial killer. Police couldn't prove that the killings were connected,
(23:06):
but the algorithms findings lined up with what police suspected,
and the algorithm had inspired a reinvestigation of those cold cases.
Hargrove felt like he was on the right track and
that maybe the algorithm could be a useful tool for
law enforcement. We selected ten major cities that appeared to
have a suspicious number of algorithm identified murders. Gary was
(23:31):
one of those ten. Gary, Indiana, the city right next
door to where Africa Hardy would be strangled. Four years later,
the algorithm flagged fifteen all unsolved murders in the Gary,
Indiana area. They were all women who were strangled. Not
one of the cases were solved, which is unusual. Seventy
(23:53):
seven percent of female murders get cleared, but not one
of these fifteen strangulation murder and Gary were cleared. I
called the public relations officer for the Gary Police Department,
gave him my name and said what we had found,
and said, is there a chance that you're dealing with
a serial killer? The next day the phone rings, it's
(24:17):
uh as I recall His name was Captain Roberts, who said, UM,
I've checked with our detectives and I can tell you
definitively that there are no unsolved serial murders in Gary, Indiana,
which is by definition an impossible statement to make. Unless
you have no unsolved murders, you cannot claim you'd be
(24:39):
definitively certain that there are no unsolved serial murders. Hargrove
felt like the police were just blowing him off, so
he started investigating the fifteen murders himself. Hargrove wanted to
know whether the algorithm was working, and he thought that
if he found evidence suggesting the murders were connected, maybe
the police would start take came seriously. But looking up
(25:02):
the information about the cases wasn't easy. The FBI supplemental
homicide report didn't include the victim's names or even the
exact date of their death. It just listed a month
in a year. So define more details about the cases,
Hargrove had to meticulously dig through old issues of local papers,
and as he started to piece things together, he was
(25:24):
unsettled by what he found. He was disturbed not just
by how grotesque these murders were, but by the patterns
that seemed to link them, and he became convinced that
Northwest Indiana had a serial killer on the loose, a
strangler who was targeting young women, women just like Africa Hardy.
He tried to talk to Gary's chief of police, but
(25:47):
he couldn't get through. I continued to suggest, have you
really looked at these cases? Soon the police stopped returning
his calls altogether, and Hargrove needed a response. He wasn't
just playing arm chair detective. We were about to publish
a story saying that Gary, Indiana has a serial killer
and the police would not talk about it. We were
(26:10):
afraid that the reason they weren't talking to us was
because they were hot on the heels of solving it.
That they had a suspect that we're trying to reel
them in, and we were going to screw that up.
We needn't have worried, but um, that was our fear.
I even sent registered letters to the chief of police
and to the mayor saying what we were about to do,
(26:32):
and if there's any issue they have, or any conversation
they want to have, for heaven's sakes, call me. This
is the letter that I wrote to the chief of
Police and Mayor in Gary, Indiana, and it goes dear Chief.
Carter Scripts Hour news service based in Washington, d C
(26:52):
is conducting a national reporting project looking into the thousand
unsolved thomicides committed in the United States since night. As
part of this project, we are investigating whether it's possible
to spot victims of serial murder among these unsolved killings.
Using the FBI's Supplementary Homicide Report. Using multivariate analysis, we've
(27:17):
determined that Gary, Indiana has an elevated number of unsolved
murders of women who were strangled in recent years. The
data that your department reported to the FBI are consistent
with the possibility that multiple victim killers have operated in
northwestern Indiana. Broadly, we see two possible patterns. In recent years,
(27:40):
several women have been strangled in their homes. In at
least two cases, of fire was set after the women
were killed. Also, starting in the nineteen nineties, we've seen
several women who were found strangled in or near abandoned buildings.
We doubt these killings are the result of convicted serial
(28:00):
killer Eugene V. Brit who admitted to killing eight people.
Please note the attached list of homicide victims. We'd be
grateful if your detectives would review these cases to determine
if any might have a common perpetrator. The US Justice
Department defines a serial killer simply as anyone who kills
(28:22):
two or more people in separate incidents. Experience thomicide investigators
tell us it's extremely difficult to spot a serial murderer.
There have been enough unsolved killings of women in Gary,
Indiana that your metropolitan area made our top ten list.
We are contacting authorities in all ten areas. Police and
(28:45):
five cities have already confirmed that cases we've cited contained
proven or suspected victims of serial murder. We are also
making a similar request to the Lake County Coroner's office.
Thank you for your time and consideration. Please help us
explore whether national crime data can assist local law enforcement. Sincerely,
(29:05):
Thomas Hargrove, Scripts Hour News service, and we had total
radio silence from those people. The Gary police chief and
mayor didn't respond to Hargrove's letter or to his follow
up phone calls, but Hargrove had also sent the letter
to the county corners corners or the public officials who
(29:26):
oversee autopsies and determined causes of death. They work with
the police, but they're an entirely separate entity. When I
called the Lake County Coroner's Office, identified myself, I'm Tom
Hargrove calling from Washington, d C. Oh just a minute,
Mr Hargrove, the senior deputy corner, wants to talk to you.
He came on and said, Mr Hargrove, we got your
(29:47):
packet of information. Thank you very much for sending it.
I'm assigning it to one of our assistant corners, a
lady named Jackie. We're going to have her look into it.
An entirely different reception them when I didn't get it Gary,
they agreed with us that there were too many unsolved murders.
She added three more cases that she thought belonged on
(30:08):
that pile we had identified fifteen. She added three, making
eighteen that she thought were connected and was trying to
have a conversation with the Carry Police department. She's never
gone on Mike to talk about this case. For the
Coroner's office, it is very, very difficult to speak ill
of a police department. It's considered bad form, and so
(30:31):
she probably still feels a reluctance to do that. Although
she has passion about this case, you shouldn't use this
recording where I named her unless she agrees. I was
looking up other names of people in that department, you know.
I think one went down for some kind of corruption charge,
and then it's kind of hard to find someone who
is there that can talk about this stuff. Yeah, now
(30:55):
your your only hope is to get Jackie to talk.
This is really one of the most frustrating experiences of
my life. I think the Gary Police Department should be
looking at some of those old cases. They still may
have a killer out there. When I finished speaking with
hard Grove, I tried calling Jackie, but I couldn't get through.
(31:16):
After a couple of failed attempts, I left her a
voicemail explaining the podcast how I was trying to look
into the murders Hargrove had identified to see if they
were indeed connected to Africa's death. Weeks passed, and I've
worked on other stories and chased down other leads, and
after what Hargrove had told me about her reluctance to
(31:37):
talk about the case, I didn't think she would want
to speak to me. Then one morning I woke up
to a missed call from a number I didn't recognize. Hey, Don,
this is Jackie. You would try to contact me a
while back in regard to heart Grow story you're doing. Yes,
(31:57):
you want to give me a call later next week? Um,
that would be fine. Sorry, had gotten back to you sooner.
It's just everything is fine, a little different. All right, guys,
I'll talk to you soon. Maybe that's coming next episode.
They don't want to talk about it either, I assume
(32:18):
because they don't want to embarrass their neighboring police agency.
All right, and said here and they had nothing to do.
They led to the death your friend. You should try
to find out. You'd be the first to do that.
Algorithm is released weekly on Tuesday's Subscribe Now so you
don't miss the next episode on the I Heart Radio app,
(32:40):
Apple Podcasts, or wherever you get your favorite shows. This
episode was written and produced by me ben Key. Brick.
Algorithm is executive produced by Alex Williams, Donald Albright, and
Matt Frederick. Production assistance in mixing by Eric Quintana. The
music is by Makeup and Vanity Set and Blue Dot Sessions.
(33:03):
Thanks to Christina Dana Miranda Hawkins. Jamie Albright, rema El Kaili,
Trevor Young, and Josh Thane for their help and notes. Again,
thanks for listening as it heads up. I'm still working
on this podcast as we release it, so any feedback
is appreciated. I think Algorithm is going to address some
(33:24):
really important issues about policing and how crimes are investigated
that don't receive enough attention. So if you can, please
leave a review on Apple Podcasts or tell a friend
about Algorithm, where brand new show and could really use
your help.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.