Experts: Broncos parade estimate 3-5 times too high


Posted Thu, Mar 10, 2016

Most detailed analysis pegs attendance at 198,000

SEA OF ORANGE: Denver Broncos Super Bowl 50 victory parade through downtown Denver on Feb. 9, 2016. [Photo: Ben Hays]

A crowd gathers at Union Station, the starting point for the Denver Broncos Super Bowl 50 victory parade through downtown Denver on Feb. 9. (Denver Broncos team photography)

A beautiful day. Smiles. Laughter. Heroes. A super parade. An ocean of orange. There must have been a million fans.

And the city said there were.

But it wasn’t true.

Not even close.

Four experts from around the world have provided estimates for the Feb. 9 crowd that indicate the city’s estimate of the crowd was about three to five times higher than the actual attendance at the Denver Broncos Super Bowl 50 celebration in downtown Denver.

The most sophisticated analysis of the crowd found that 198,000 people came to the entertainment, parade and rally in downtown Denver that day. That’s a little less than one-fifth as high as the 1 million-person mark that the Denver Police Department said was surpassed.

The four experts — from Austria, the United Kingdom, Hong Kong and Arizona— came to their conclusions after having seen photographs and maps of the parade route and Denver Civic Center. The detailed analysis—by Austrian expert Alexandar Kollaritsch, who owns Vienna-based 4mation Event and Security Consulting gmbh — goes into great depth about the intricate methodology used in calculating the number. (See the accompanying “methodology” story here and the analysis itself, which is at the link immediately below.)

-Denver Broncos Parade Kollaritsch 15116643 – V1.4 copy

The other three experts involved in producing estimates for the Post-Telegraph were G. Keith Still, professor of crowd science at Manchester Metropolitan University (UK), Paul Yip, professor in social sciences at the University of Hong Kong, and Stephen Doig, Knight Chair professor of journalism at Arizona State University in Tempe.

Why so high?           

How is it possible that the city’s estimate could have come in so high?

Aerial Photo of Civic Center, Photo from The Denver Post

This aerial photo of Civic Center was used in Kollaritsch’s analysis of the crowd. (Photo by RJ Sangosti/Denver Post found here.)

The answer is not entirely clear. Denver Police Department spokesman Doug Schepman declined to provide a name or arrange an interview with the specific person or persons who could explain the methodology used by the DPD in producing a figure of 1 million-plus for the event.

“We have provided you with a statement regarding the process for estimating the crowd and have provided numerous additional details here in response to your emailed questions, and will have no additional comments,” said Schepman in a Feb. 29 email.

Schepman did not respond to written emailed questions on Feb. 24 and March 2 asking for comment about the Post-Telegraph’s having received much lower crowd estimates from professors. Messages concerning the lower estimates left with Amber Miller, communications director in the mayor’s office, and Jenny Schiavone, director of citywide marketing for Denver, were not returned.

Explaining the DPD’s methodology on crowd estimation, Schepman’s emailed statement on the DPD’s post-event estimate of “more than 1 million attendees” said: “That estimate was based on the size of the areas involved, the crowd density observed by officers (including thousands of people seen watching the parade from inside buildings along the route) and by extrapolating estimates from prior large scale events.”

Security expert Alexandar Kollaritsch estimates the crowd to be between 178,200 to 217,800.

Security expert Alexandar Kollaritsch produced an estimate of 198,000 for the Feb. 9 event, with a plus or minus 10 percent margin of error. (Courtesy photo)

Extrapolation of a wrong number

Kollaritsch, who has run his own event security company in Austria since 1998 and has been responsible for the security of about 4,000 events (including some with more than 50,000 people), is particularly critical of using extrapolation. “I think that extrapolation of prior large scale events is not the answer — if the number of visitors at the other events have been estimated with the same accuracy you are just extrapolating a wrong number,” he said in an email.

Was that what was going on? Previous estimates of 650,000 for the Broncos’ first Super Bowl parade in 1998 and 375,000 for the Broncos’ Super Bowl parade of 1999 were widely quoted by local media. (Click here and on other highlighted words for links to articles and other backup information.)

The population of metro Denver has grown about 40 percent since 1998, so there is a certain logic to taking the 1998 number of 650,000 and adding 40 percent, thus yielding 910,000. Add in nice weather on Feb. 9, and the extrapolation method might seem to easily yield a million. A sensible procedure? No. Because the new 2016 estimate is almost certainly — as Kollaritsch put it — “extrapolating a wrong number.”

Did the DPD make the estimates for parade attendance in 1998 and 1999, and if so, would anyone from DPD be able to talk about the methodology used in arriving at those estimates? Said Schepman: “It’s possible that DPD provided crowd estimates for those events, however too much time has passed to say definitively.”

Nate Currey, a spokesman for the Regional Transportation District, noted that on Feb. 8, the day before the 2016 event, a pre-event estimate of 500,000 to one million attendees was produced in a conference call among public information officers from various entities. He noted that the 650,000 attendance estimate from one of the previous Super Bowls came up, and “obviously Denver has grown a lot since then.”

Currey declined comment when asked about his reaction to professors having produced far lower estimates for the Post-Telegraph. DPD’s Schepman did not respond to a written question about it. However, he did note in his email, “Clearly, the Department successfully allocated resources to manage the large rally/parade crowds and provide a safe environment. There was only one arrest during the event.”

Pre-event estimates

He also noted that the DPD takes estimated attendance into account when developing security plans for events. So did it make pre-event estimates? Schepman initially said in his Feb. 29 email that “the Denver Police Department did not provide pre-event estimates of anticipated crowd sizes,” but he later clarified that statement, stressing that “DPD did not publicly release a crowd estimate pre-event, but was using an estimated 500,000-1 million range for internal planning purposes.”

Schepman cited a YouTube video of a Feb. 8 press conference in which he and Schiavone did indeed decline to produce an estimated number prior to the event. But on that day, Feb. 8, RTD’s Currey was quoted by 9News about the 500,000 to a million estimated attendees. And the possibility of 1 million attendees came up in a Feb. 8 CBS4 interview with Brea Olson, marketing communications manager for the Downtown Denver Partnership.

Pressure to produce a high estimate?

Since those estimates of up to a million were circulating prior to the event, was there pressure to come up with a post-event estimate of 1 million so that the turnout would not look disappointing?

“No,” said DPD’s Schepman in a one-word emailed response.

RTD’s Currey took the same position. He said in a phone interview that he never heard any mention of that sort of pressure — to produce a number to meet expectations.

But there is some evidence that civic pride can somehow become enmeshed with parade size.

For instance, after the Feb. 9 celebration in Denver, the website of the Kansas City Star newspaper posted comparative photos of the Feb. 9 Denver event and the Kansas City Nov. 3 event after the Royals won the 2015 World Series. The attendance for the Royals event was pegged at 800,000, lower than the Denver estimate of 1 million. “Royals fans weren’t buying that the Broncos’ parade was bigger,” said the website.

In other words, it’s not just about winning the Super Bowl. It’s also about producing the biggest parade.

Denver not alone

Denver is not alone in producing high estimates for its parades. “In our experience the police have a lot to manage during major events and their estimates (as well as politicians, local promoters and event organisers [British spelling])

"Estimating the size of much more about public relations than a quest for truth." -- Arizona State professor Steve Doig

“In reality, estimating the size of crowds at mass public events is much more about public relations than a quest for truth.” — Arizona State journalism professor Steve Doig (Courtesy photo)

are typically much higher than a crowd safety and risk analysis estimate,” said Manchester Metropolitan professor Still in an email.

Indeed, Arizona State professor Doig, in a 2009 article for, said, “In reality, estimating the size of crowds at mass public events is much more about public relations than a quest for truth” as “organizational reputations and personal egos are ballooned or deflated by public perceptions of whether the crowd is surprisingly large or disappointingly small.”

In the article he cites the Miami Calle Ocho street festival, which at one time claimed a “million-plus” crowd — at least until the Miami Herald determined that, based on square footage, the maximum capacity would be 175,000.

The huge range in estimated Super Bowl parade crowds also might be cause for some legitimate eyebrow-raising. (See the story about other Super Bowls below this one.) For instance, the Boston Patriots did not provide a parade estimate after the 2015 Super Bowl victory, but had previously estimated 1.5 million attendees in 2004 and just under a million attendees in 2005 for Super Bowl celebrations.

Kollaritsch, who produced the detailed analysis for the Post-Telegraph, is a top student in an online master’s course in crowd safety and risk analysis that is taught by Still in England and reaches students at home and abroad. It’s the first class of its kind in the world and focuses on crowd modeling, monitoring and management, said Still.

Kollaritsch tackled the problem of measuring the Feb. 9 Denver crowd, and Still concurred with the conclusion: 198,000. “The analysis looks OK to me,” Still said in a March 9 email, but he added that the estimate is based on “the information provided” and that the final count would have a margin of error of plus or minus 10 percent. That means the range could be from 178,200 to 217,800.

"Take into account people who might not stay for the whole time, or people who are moving in and out during the parade." -- Paul Yip, professor in social sciences at the University of Hong Kong

Paul Yip, professor in social sciences at the University of Hong Kong, estimated that 200,000 people attended the Broncos’ Feb. 9 victory celebration. (Courtesy photo)

Yip initially provided an estimate of 200,000. “The official estimate based on one of your enclosed is 200,000!” he said in a Feb. 20 email. Later, he saw a preliminary analysis of Kollaritsch that estimated the crowd to be 215,000.

Yip said of the Kollaritsch preliminary analysis, “It looks good to me.” He added that— to take into account people who might not stay for the whole time or people who are moving in and out during the parade — adding 10 percent might be reasonable. Ten percent added to the Kollaritsch preliminary estimate of 215,000 would yield an estimate of 236,500, and ten percent added to the Kollaritsch final estimate of 198,000 would yield 217,800.

In a Feb. 24 email, Doig provided an estimate of 350,000. In early March he was traveling and didn’t have time to take a look at the Kollaritsch analysis. “The estimate I gave you was a rough back-of-envelope one, and I’ll let it stand at that,” he said in a March 8 email.

Does anyone care?

So, do crowd estimates matter, and does anyone care? Prominent media attention in Denver to the million-person mark would indicate that reaching such a milestone makes for a good news story. And, as indicated by the Kansas City Star website, civic pride can be involved. But other issues—such as the popularity of a particular movement or safety—can make crowd estimation a very serious science indeed.

In his MSNBC article, Doig notes a couple of examples, including the 1995 estimate by the National Park Service of 400,000 people at Louis Farrakhan’s Million Man March. The estimate was well below the touted 1 million, and it was seen by some as disappointing. Boston University researchers, funded by ABC-TV, shortly thereafter estimated the crowd first at 870,000, with a 25 percent margin for error, and then at 837,000, with a 20 percent margin for error. But those estimates were controversial, the National Park Service never retracted its estimate, and other academics have supported the Park Service number. Farrakhan threatened to sue about the low estimate, and Congress then prohibited the National Park Service from making estimates.

G. Ketih Still - G. Keith Still, professor of crowd science at Manchester Metropolitan University

G. Keith Still is professor of crowd science at Manchester Metropolitan University (UK) and instructor of Kollaritsch in an online course on crowd safety and risk analysis. (Courtesy photo)

But it’s not just about the intangible aspect of perceived success or failure of an event or community pride. As Currey noted, safety is a big concern.

Indeed, Still has on his website a 59,000-word history of crowd-related tragedies dating back to a 1902 incident in Scotland. Many of those disasters were tied to sports or religious events. The latest listed was the October 2015 stampede at an evangelical crusade in Cameroon that killed one woman and injured several other people. The most recent large one was in September 2015 in Saudi Arabia, where at least 2,110 religious pilgrims were killed in a crush of people. The website also cites the November 2008 incident in which a Long Island Walmart employee was trampled to death by a crowd of shoppers on the day after Thanksgiving.

Professor Still — who has consulted on Olympics in Sydney, London and Beijing and on events ranging in location from Austin, Texas, to Saudi Arabia — attributes crowd-caused tragedies in part to the inappropriate utilization of space, and the failure of organizations to provide a safe environment for public assembly.

Those years of experience, and his backing of the Kollaritsch analysis, lend some weight to that analysis, as do the comments from Yip and Doig. They leave little doubt that the event estimate coming from the Denver Police Department and the city as a whole were about three to five times too high.

The 1 million-fan, perfect storybook exclamation point to the Broncos Super Bowl 50 win — and, as it turns out, quarterback Peyton Manning’s career — was probably never a realistic possibility in the first place (see the “transportation” accompanying article by clicking here). Metro Denver’s public transportation and parking infrastructure probably could never have handled it.

But that doesn’t mean the event or the event attendance need be considered a disappointment. Indeed, if the most recent crowd estimate is any indication of the accuracy of past crowd estimates, the Feb. 9 crowd may have been a record-breaking one at 198,000. After all, RTD said in a press release that it had a “record setting boost in ridership on Feb. 9.”

And what might be concluded about Feb. 9? Regardless of attendance, it might still be said: It was a magnificent day. It was an exclamation point — for the Broncos, for Manning. There was delight, joy, exhilaration. And there were many, many, many fans.

There just weren’t a million of them.


Corrected: Because of miscommunication, the original story posted early on March 10 was the wrong version. The correct story was posted later that morning. Also, the figure of 370,000 used for the Broncos parade estimate in 1999 was corrected to 375,000.


Ratio of other Super Bowl parades and metro populations show wide range

                   by Tom Locke 

A look at other recent Super Bowl parades and the relationship between estimates of Super Bowl parade attendance and metropolitan populations shows a large range of ratios of parade sizes to populations. But a large number of variables can affect the size of a parade, including the length of the parade and weather. Still, if the parade size estimate is huge compared with the metropolitan population, it may be more likely to raise questions of accuracy.

Here is a list of recent Super Bowl winners, their Super Bowl parade estimates, their metropolitan populations, and their parade estimates versus metro populations at the time in percentage terms.

2016, Denver Broncos, 1.1 million, 3.3 million, 33 percent (1)

2015, New England Patriots, no parade estimate. (2)

2014, Seattle Seahawks, 700,000, 3.1 million, 23 percent (3)

2013, Baltimore Ravens, 200,000, 2.8 million, 7 percent (4)

2012, New York Giants, 500,000 to 1 million, 20.1 million, 3 percent to 5 percent (5)

1) Actual attendance at the Denver event has been shown to be around 198,000, which is 6 percent of the total metropolitan population.

2) Perhaps the lack of a parade estimate for New England in 2015 was related to very high estimates made in 2005 and 2004. The 2005 Super Bowl parade was estimated at just under a million attendees. And the 2004 parade was estimated at 1.5 million attendees. The metro Boston population was about 4 million in 2004, so the 2004 event estimate of 1.5 million was 37.5 percent of the metro population.

3) The 2014 Seattle Seahawks Super Bowl parade was estimated at 700,000. Metro Seattle has a population of 3.61 million, which is 10 percent higher than metro Denver’s population.

4) The 2013 estimate for the Baltimore Ravens parade was 200,000. Metro Denver’s population is about 18 percent higher than that of Baltimore.

5) The 2012 New York Giants parade was estimated at 500,000 to 1 million. Metro New York had 19.8 million people in 2012. That figure is about six times greater than metro Denver’s population in 2015.



About Tom Locke

Tom Locke, is a Denver-area journalist and affiliate professor at MSU-Denver.

View all posts by Tom Locke


  1. MMU in the News (Denver Post) | Gkstill's Blog - March 11, 2016

    […] Experts: Broncos parade estimate 3-5 times too high […]

Leave a Reply