THE BOOK--Playing The Percentages In Baseball

According to my notes (pre 2010 census data, so using estimates from the 2000 census extrapolated), the smallest 13 MLB MSA markets, in reverse order, are:
- Milwaukee
- Kansas City
- Cincinnati
- Cleveland
- Denver
- Pittsburgh
- Tampa
- Baltimore
- St. Louis
- San Diego
- Minneapolis
- Seattle
- Phoenix

The only deviation from that list of 13 cities and the 13 teams in the competitive draw (http://mlb.mlb.com/news/article.jsp?ymd=20111130&content_id=26059392) is the replacement of The Twins & Mariners with the Athletics and Marlins.

Its difficult to split the SF/Oakland MSA, but an Oakland only market would easily be among the smallest.  Miami, if you split away Ft. Lauderdale, also qualifies as a much smaller market.

So to me the list almost identically mirrors MSA values for market sizes.

I’ve always thought that it was important to differentiate small markets from small budgets.

Teams like St. Lou and Minnesota (i.e., their owners), should be heralded, rewarded and emulated in terms of taking a small metropolitan region and generating huge sums of revenue and then spending it on the team.

Teams like the Yankees and Philly shouldn’t necessarily be given credit for the gobs of money they make, but at least they should be given credit for putting that money on the field.

Teams like the Marlins shouldn’t necessarily be penalized for having a crappy market (I don’t know how big their market technicaly is), but they should be for pocketing any money they do get.

And teams like Houston should be penalized big time for not making more of the 4th biggest metropolitan region.

It seems like incentivizing St. Lou behavior and disincentivizing Marlins/Houston behavior will benefit everyone in the long run.

(Note, too, that Boston is the 10th biggest metro region in the US, smaller than the Astros, Rangers, Braves and so on. That should be a credit to management for generating the revenue for MLB from a medium market and for spending that money on the players)

mettle:

Toronto has 5.5 million people in the metro area.
St Louis metro is half that.

In terms of baseball, the Toronto market is somewhat smaller than the St Louis market.

In terms of hockey, the Toronto market is at least quadruple the St Louis market, if not 10 times.

You need to include the affinity.

Can you sell more snow tires in all of Texas, or just in Montreal?

when i rank city’s hypothetical ability to support major sports franchises i like to use GMP (gross metropolitan product).

http://en.wikipedia.org/wiki/List_of_metropolitan_areas_in_the_United_States_by_GMP

Milwaukee is still last, about 15X smaller than NYC, or about equal to Bridgeport, CT.

if NYC were a country is would be about the 13th or 14th richest, depending on who’s counting (non-PPP GDP) just ahead of Mexico.

Again, guys, the entire point of this thread is the affinity.

Unless you can propose something that shows Montreal to be a large market hockey team and small market baseball team, then you are going to be missing a major variable here.

Expand your thoughts beyond what you are used to doing.

maybe high school/little league participation rates?

revenue and attendance can show you fan affinity for those clubs but doesn’t that just correlate to winning? this is anecdotal but i thought places like miami and tampa have great “affinity” for baseball but a terrible following for their pro team, for reasons beyond local predispositions.

You are so cherry picking your data points and I can easily do the same: Minnesota was always as much a hockey “town” as any place else, yet baseball is hugely successful.

But, this gets to the bigger question: Where does this magical affinity come from?

The reason STL, MIN and BOS etc are great baseball cities is because they have had great baseball teams, which is a credit to ownership, and not to weather (MIN, BOS) or geography (STL).

Yes, there probably is a Canada factor due to extremes of weather and hockey, but consider:

- Why is SD small market and Detroit big market? Their metro regions are the same and So Cal certainly has enough of a baseball tradition and good enough weather as evidenced by the success of the Dodgers and Anaheim. It’s because, as good as the GMS may be, the Padres have done a poor job of making money off baseball.

- Why does Boston, 10th biggest metro area, have the 2nd biggest payroll? It’s because of a committed fan-base that grew. And Miami? The exact opposite has happened, with ownership alienating fans year after year.

So “affinity” is essentially a function of stadium (as determined by ownership), marketing (as determined by ownership), talent acquisition (as determined by ownership), winning history (as determined by ownership) and maybe team age.

Unless you can come up with a better definition of affinity in a non-teleological way (i.e., Boston is a good baseball city because it is), I’m not buying it.

I did a regression of MSA against my “BASE” numbers from my study.  I came up with this equation of estimated attendance:

23,340 plus .0012 * MSA

That gives us a range of 25,318 for Milwaukee to 35,458 for NY.

I divided my BASE numbers by this estimated to get an “affinity” number.

These 4 teams had an affinity number of 1.2 or higher:
Rox    
Dodgers   
DBacks    
Cards

Three teams had an affinity number of 0.8 or less:
Tampa
Florida
Oakland

So, as one suggestion, you increase the MSA by 20% for the 4 teams and decrease by 20% for the other 3 teams.

Obviously, we can come up more strict, but this is a starting point.

Use merchandise sales as an affinity proxy, I’m sure MLB knows how many hats, jersey’s, and T-shirts they have sold with each logo on them.  Combine that with a population of people that are within a 20-30 minute drive or able to take rapid transit (L, T, subway, metro, ...) to the park.

I’m not cherry picking anything.

Detroit has twice the metro area of SD.  According to my calculations, their attendance is a direct proportion to their MSA.

Huh?
http://en.wikipedia.org/wiki/Table_of_United_States_Metropolitan_Statistical_Areas
Pop ratio: 4:3
Payroll ratio: 7:3

But what’s your point there? We all agree that there is a huge amount of variance in revenue not accounted for by population size that we can call “affinity”.
You seem to suggest that affinity is baked into the earth or something.
I’m suggesting affinity is a function of ownership and team age, which seems to constitute “[not] expanding your thoughts beyond what you are used to doing” in your view.
I would counter with the same - you’re simply labeling, I’m trying to explain.

Affinity thoughts; some case-by-case team items:

Wertheim and his co-author in the book Scorecasting showed statistically that the Cubs attendance as actually in reverse-relation to the team’s W/L record, yet was directly in correlation with the price of Beer.

Clearly Boston’s attendance and market size don’t directly correlate to their payroll or revenues because, in reality, their “market” is basically the whole of New England.  Not to mention the millions of other fans around the country since they’re a national brand. 

The Yankees not only have a massive revenue stream but are also a National brand (along with Red Sox, Cubs).  So these national teams can’t possibly be included in any conventional attendance, revenue or market size/population studies because of it.  It doesn’t matter though, since all three teams are in massive enough markets not to qualify.

St. Louis is a smaller market, but has been relatively successful as of late.  When Cleveland was successful in the mid 90s they had one of the highest payrolls in the game and a consecutive sell-out streak that guaranteed mass revenues despite their small market size.  Same with Baltimore. 

Pittsburgh and Kansas City have both been plagued with poor management and haven’t had winning seasons in years, AND happen to be in really small markets. 

Tampa has been successful yet proved that attendance isn’t correlated with success by drawing fewer fans the season after a world series. 

Washington (my team) was a team “given” to a very wealthy friend of Seligs over other more deserving and more active baseball groups in this town, who then promptly ran the team as a profit center (like his own successful businesses) and really drove a lot of fans away.  He’s kept payroll in the bottom 1/3rd of the league and has paid the price for it.  He’s clearly not in a small market but his revenues have to be in the bottom half, and he’s taken revenue sharing dollars (ridiculous).  According to 2009’s forbes numbers the team profited something like $33M while putting a 100 loss team on the field.

Loria has pocketed millions in revenue dollars over the years while barely trying to field a decent team, finally being basically “forced” last off-season to spend money on his team to avoid a union grievance.  Great guy there.

I think the “affinity” discussion is entirely on a case-by-case basis.  Every team has an explanation for why they’re doing well or poorly, and its not always correlated with any one statistical measure.

for affinity you’re talking about a demographic’s predisposition to being a fan of a particular sport in their local metro area, right?

so if all things are equal, what is the likelihood of an average person in city X will become a fan of a pro sports franchise in city X vs an average fan in all over North America?

i do think such a hypothetical variable exists for baseball. however, at least with baseball, i dont think there is a major metro area where “affinity” makes much of a difference when judging what is a “small” or “large” market. the attendance differences from the base for the major league teams listed in #8 i dont think is captured by innate pro baseball “affinity” of the local population, but by the things mettle suggests in #7. the previous success of the established franchise and the ability/luck of the ownership.

so if we go back to 80’s and you wanted to pick a city for expansion, I wouldnt give Denver a plus or Tampa a minus in the “affinity” category.

I would use the local TV contract size in $$$. Maybe a combo of households and $$$.  It could the people under the MLB blackout area.

ken/7:

“for affinity you’re talking about a demographic’s predisposition to being a fan of a particular sport in their local metro area, right? “

Yes.

And I’ve already shown in my link in the main blog post how to figure out what the average attendance is for each team if they were all .500 teams.

It has nothing to do with ownership or anything, since I’ve got it baselined such that all teams are .500.

So, I’ve already figured out MSA x Affinity.

It may be hard to quantify, but St. Louis has a long history and tradition, as for the first half of the 20th century it was the western-most MLB team, and during the golden age of radio, its broadcasts were widely heard via KMOX. Sure, the Red Sox are New England’s team, but the Cardinals were middle America’s team. So the branding and history are quite strong.

#15 - how can you assume that just controlling for winning percentage that everything else is MSA x Affinity? thats not even taking into account the cost of the ticket.

for example, if we started two new franchises in new MSAs, Nashville and Norfolk maybe, but really wherever, and then counted their ticket stubs for a few years and adjusted their baseline for winning percentage, anything off the baseline would be due to the locals’ “affinity” to MLB baseball, even though one team could have 25 cent tickets and free parking and the other has $2000 tickets and the stadium is surrounded by a moat all fans have to swim across to get to their seats?

what am i missing?

I said it was a starting point.  If I had ticket prices, I would use that as well.  I mean, duh, right.  If a minor league park sells 30K seats at 5$ a pop, that’s not the same as selling 30K seats at 50$ a pop.

Please presume that if I didn’t specify something doesn’t mean that I am dismissing it.

ok but i’m still confused about this: “It has nothing to do with ownership or anything, since I’ve got it baselined such that all teams are .500.

So, I’ve already figured out MSA x Affinity.”

how do you know you’ve figured out MSA x Affinity? how do we know it has nothing to do with ownership or anything?

i would think that in just about all major north american cities, once you control for winning percentage and population, you still have a whole mess of other things to control for before you can you’ve figured out “affinity”.

If you have one market that has 5 million people and that draws 30,000 fans per game, and another market that has 2 million people and draws 30,000 fans per game, and both parks charge 20$ a ticket, and both teams are .500 for five years, what reasonable opinion are you going to make?

yeah there is still a whole host of things i think that will end up mattering more. the team’s history will be huge, whether they’ve got a storied past or not and how long theyve been around. and then all the factors related to the ballpark itself. transportation, parking prices, concessions, the price of beer, ample toilets, etc etc. then there’s the competition in the MSA for other sports and entertainment in general. weather. disposable income. there are tons of things i would want to look at before i would assume people in denver are significantly and naturally predisposed to going to baseball games over people than tampa.

controlling for price would clear up a whole lot though.

i think ideally you’d pick two expansion teams with similar GMP and then look at tv ratings after controlling for winning percentage.

we learned a similar kind of example in school. art house movies box offices are disproportionally high in NYC and LA. so people might assume that NYC and LA residents are more cultured and enjoy more high brow movies disproportionally than the rest of america. but when netflix released their stats, the art house movies did better in other places than in those cities (after taking into account the fact that more LA and NYC people already saw the movies in the theatres).

so thats why i said affinity is real, i just dont think it matters all that much for attendance figure or for how MLB should decided who is small and large market for the CBA.

I’m not disagreeing with you.  I am saying that what I’ve got is better than MSA.  And there’s no way St Louis is a small market baseball town.

Even if you don’t think my solution is as good as it should be, it’s better than MSA.

ok yeah, sorry i was going off track there with the affinity. i see what you mean about the cardinals and i agree. the large markets teams’ lawyers should have noticed this and raised some sort of fuss. like the cubs. who wants to have your lunch eaten by a competitor and then have to help subsidize them as well?

I agree with mette’s points and have little to add.  Here are three problems with using different metrics than the size and perhaps the wealth (domestic product) of MSAs:

1.  You start getting obvious issues with cherrypicking data, as mettle pointed out.

2.  Again as mettle pointed out, you start penalizing teams for doing any of making the most of their market, investing in putting a good team on the field that attracts fans vs a team in a bigger MSA where the owner pockets the money and people don’t attend games, or teams that have successfully transcended their market and created a national brand.  This is all behavior by owners and management that should be encouraged.  Any revenue sharing scheme that gives an incentive for a “small market” owner to pocket the money is a fraud.

3.  Mettle hasn’t pointed this out, but alot of the metrics I see to show that teams are “small market” are circular.  I’ve seen people actually use the size of payrolls to claim that teams are large market or small market.  You shouldn’t be able to artificially create a small market team, for example like running it like the St. Louis Browns in the first half of the century.  Anyone can take a team in a mid-sized city, not spend money on signing good players, and watch payroll and attendance shrink.  The metrics have to show potential.

Also, how much more “affinity” did Blue Jays fans have for hockey over baseball in 1992 and 1993.  How does this compare with the “affinity” today, or during a year where the Maple Leafs are actually winning?  For that matter, how does the “affinity” in Montreal for hockey vs. baseball compare in 1993 as opposed to today?

Also, when using the MSA metrics, make sure to cut New York and Chicago in half, and to separate Washington and Baltimore, Oakland and San Francisco, and Los Angeles and Anaheim.  I’m not sure why you would want to separate Miami and Fort Lauderdale.

Ed/24:

You are missing the point I am bringing up with your examples.  Specifically:

1. I did not cherry pick anything.  Don’t comment unless you confirm you read this:
http://www.insidethebook.com/ee/index.php/site/article/attendance_base_for_each_team/

2. This is the only part that makes sense:
“or teams that have successfully transcended their market and created a national brand.  “

Otherwise, in my link, I have already controlled for everything else, in that I am focusing on how many people attend games if the team is a .500 team.

So, if you want to argue that the Dodgers and Cardinals score as high as they do because their brand transcends whatever, then fine.  But, with my system, you will note that the Redsox and Yankees are NOT similarly advantaged.  They did not score above and beyond whatever their MSA suggested.

Therefore, I’m not buying even that part of the argument.

3. See, this is how I know you didn’t read my article, because I am not arguing about using payrolls or doing circular logic.

Furthermore, you bringing up the Jays in 92-93 ignores the fact from my link that I am controlling for the winning percentage effect to attendance.

And the cutting MSA in half: you obviously ignored Zumsteg’s article, because he did precisely that.  And in my testing, cutting MSA in half was exactly what needs to be done to get the best correlation to attendance.

***

So, please, stop with all the red herrings here.  Make an effort to read the three links in the main blog entry.

i hope this isnt nitpicking but i still get confused when you say “I have already controlled for everything else, in that I am focusing on how many people attend games if the team is a .500 team.” you still havent controlled for price (and i know you havent willfully ignored it) as well as all the other things that go into whether or not people go to a game.

so while i think what you have is better than MSA but i think just looking at TV ratings or club revenue is better than anything related to attendance. i know that gets back to the in disincentivzing clubs from making money so you look for more objective numbers and then you circle back to where we are. so maybe thats how the clubs went and they decided population was still the best method, even if you get someone like the cardinals in the “small” group.

24/25/
I read the linked pieces and Keith seems to get it best in actually trying to explain where affinity comes from as opposed to just labeling a variable.

Tango - The major thing you’re not getting is the way ownership contributes to “affinity” *outside* of just winning. You’ve worked for teams, so you know that there are hundreds of employees down the hall from you who do things that contribute to the team beyond just calculating future WAR.
The factors outside of winning % that contributing to affinity are (in order of importance):
- marketing
- stadium
- years team has been around
- history (sort of based on winning, but of the team 10 years ago)
- free agent strategy (keeping loved players, not doing firesales, etc)
- other sports teams in the area (but I doubt this since I think there’s more of a multiplier effect here)
- weather (maybe)

My point is that the ownership has control over a lot of these variable and should be *awarded* for maximizing affinity in their city.
To do otherwise is to award incompetence (e.g., Florida, Houston) over success (e.g., Boston, St. Lou).

mettle: I’ve already said that with my method the Redsox are NOT showing extra affinity.  Why do you bring them up?

ken: yes, I’d like to have ticket prices.  And, it’s not that much effort to get it, seeing that TMI does it every year.  So, once we have cost per ticket, and we have tickets sold, then we can figure out everything we need.

To simplify: MSA alone gives the teams 100% “credit” for creating affinity.  Cardinals, you’re in an area that loves Cardinal baseball?  Great work! You’ve done a great job building up your fanbase over the years, and we’re not going to penalize you for it.

Tango, on the other hand, gives the teams 0% credit for creating affinity. Cardinals, you are lucky dogs to be playing in a baseball-crazy area. You had nothing to do with your local fans’ relative greater willingness to spend money on baseball (or if you did, we don’t care).

Neither approach gets us where we’d really like to be.  I’m not sure which approach is better.

Given that we’d expect the Redsox and Yanks to be part of the “teams that created their market”, and my process does not put those two teams in there, then I question the idea that a team can create some market, over and above winning.

I would think you start with the most objective base possible, which would probably be MSA Population and TV Markets.

Doing my own rough calculations I get something like:

LARGE
NY Yankees
NY Mets
Chicago Cubs
Chicago White Sox
Los Angeles Dodgers
Los Angeles Angels of A.
Texas Rangers
Philadelphia Phillies
Boston Red Sox
Houston Astros

MEDIUM
Miami Marlins
San Franciso Giants
Detroit Tigers
Toronto Blue Jays
Washington Nationals
Atlanta Braves
Arizona Diamondbacks
Seattle Mariners
St. Louis Cardinals
Minnesota Twins

SMALL
San Diego Padres
Tampa Bay Rays
Colorado Rockies
Oakland A’s
Baltimore Orioles
Pittsburgh Pirates
Cleveland Indians
Cincinnati Reds
Kansas City Royals
Milwaukee Brewers

You might quibble at the edges, like moving Houston to Medium or Miami to Large, Minnesota to small and San Deigo to Medium.

Tom I think wants to add affinity, but if the data already shows that Yankees, Red Sox, don’t seem to have affinity, wouldn’t that argue somewhat AGAINST considering it?  Certainly certain teams may not be MAXIMIZING their market, or are limited in the short term by stadium locations, stadium age, other major sports in the same city, etc. but I would think you’d want to look more long-term at the Potential Long term of the market if the team ownership properly invested in the product, etc.

I find it hard to believe you keep saying “team 1 doesn’t fit your point, so it must be wrong” right after saying you’re not cherry-picking, but so be it. To highlight my point, and the point Hizouse is making (hopefully more clearly than me) consider:

I used the revenue/capita data from Zumsteg for the following reasons:
- Tango didn’t publish his current affinity data
- Woolner’s link to the Scarborough reports was dead
- Woolner’s data were opaque and I couldn’t find affinity by itself
- Tango’s old data at the link are problematic because he used buckets for years which skews the data big time—I’ll explain at the end of the post.

But, using the Z data, here are some of the teams with high affinity & possible reasons why:
- Cleveland (great ownership, commitment to winning, marketable players)
- Rockies (stadium, novelty)
- St Louis (see above)
- Dbacks (hardware)
- Giants (ballpark, marketing. city wealth)
- Atlanta (a decade of winning has generated a fanbase)

On the low end are:
- Mets (terrible front office, terrible stadium)
- Orioles (terrible owner)
- both Canadian teams (it’s Canada)
- Marlins (terrible owner)

So, already, there is evidence of something going on - I think we’d all agree that these 11 teams fit with our expectation of what’s a well run team vs. not.

There is a problem with the Z data, though: There is a ceiling that punishes the big market teams in revenue/capita. You can only get so many people into a baseball stadium at a time without degrading the experience—so even though NY is 15 times the size of KC, you can’t get 300,000 people in a stadium for a game. This accounts for the absence of BOS, NY, CHN from the list—they’re hovering in the middle. Plus, I think teams are regressed to the mean because of away-game revenue.

As for Tango’s attendance base data, I was going to use it to calculate affinity by dividing it into Z’s MSA, but the numbers were looking sketchy when I took a look at NYA 87-09 with an attendance of 30k/year. That certainly doesn’t pass the sniff test, and here’s why it’s skewed:
- NYY had only a few (4) losing seasons
- the losing seasons occurred at the beginning of the bucket
- NYY was winning a lot when MLB attendance went up from 24k in ‘87 at the beginning of the bucket to 28k-29k when the yankees were winning.

So, to normalize to .500, you were forced to use a very small sample size from the early 90s when attendance was lower and then discounted the many years of 40+k/yearin the late 2ks. Until a more sophisticated winning-neutral attendance measure is used, I don’t think we can use this.

(p.s. Boston and Chicago’s stadiums are limited to around 35k/game so that artificially caps anything you try to calculate there)

mettle: So, adjust it on a more annual basis if you don’t like how I did it.  You make it seem that because I did it a certain way, and if you find an issue with it, then just chuck the whole thing.

kjok: interesting results.  At first glance it looks good.  Can you post your MSA and TV market data?

Perhaps Redsox have average affinity because I didn’t include price per ticket, where presumably, they are highest in the league.

And, if you compare Yanks adn Mets attendance, and controlling for win%, you will see that Yanks don’t have more affinity than Mets maybe?

I am not coming to conclusions.  I am giving a STARTING point and various parameters to consider.

So, imagine you are at school, and you are charged with coming up with what a small market is for your thesis.  Think outside whatever myopic view you already have.  Think outside the box.

the more i think about it, the more i think MLB is better off just using population (or GMP if it were my call). if you could get ticket prices and include them into the attendance numbers you could start to get an idea of local affinity, but that still doesn’t take into account the location of the stadium.

MLB is just looking to see who’s got inherent advantages or disadvantages do to the territorial monoplies they’ve set up for their franchises. i think population size is the only way to objectively do that without punishing success and rewarding ineptitude.

maybe if you had TV ratings and controlled that for winning pct you could see if certain cities tune in at greater percentages than others. i still think the affinity factor for baseball is too small for MLB to bother with for things pertaining to competitive balance. if that means st louis is labeled a “small” market, then thats life. good for them i guess.

Mettle, Kendynamo, and Hizouse are making a lot of sense as I, an interested blog reader, follow through this discussion.

I’d just use franchise value as a proxy for large/small market and be done with it.

Tom asked for the MSA Data - probably won’t format well - MSA plus Neilson TV markets, Non-US city pops from other sources:

Rank Metropolitan Area 2010 TV Rank
1 New York-Northern New Jersey-Long Island, NY-NJ-PA 18,897,109 1
2 Los Angeles-Long Beach-Santa Ana, CA 12,828,837 2
3 Chicago-Joliet-Naperville, IL-IN-WI 9,461,105 3
4 Dallas-Fort Worth-Arlington, TX 6,371,773 5
5 Philadelphia-Camden-Wilmington, PA-NJ-DE-MD 5,965,343 4
6 Houston-Sugar Land-Baytown, TX 5,946,800 10
7 Washington-Arlington-Alexandria, DC-VA-MD-WV 5,582,170 9
8 Miami-Fort Lauderdale-Pompano Beach, FL 5,564,635 16
9 Atlanta-Sandy Springs-Marietta, GA 5,268,860 8
Toronto, ON 5,200,000
10 Boston-Cambridge-Quincy, MA-NH 4,552,402 7
11 San Francisco-Oakland-Fremont, CA 4,335,391 6
12 Detroit-Warren-Livonia, MI 4,296,250 11
13 Riverside-San Bernardino-Ontario, CA 4,224,851
14 Phoenix-Mesa-Glendale, AZ 4,192,887 12
Monterry, MX 4,080,000
Montreal 3,900,000
15 Seattle-Tacoma-Bellevue, WA 3,439,809 14
16 Minneapolis-St. Paul-Bloomington, MN-WI 3,279,833 15
17 San Diego-Carlsbad-San Marcos, CA 3,095,313 28
18 St. Louis, MO-IL 2,812,896 21
19 Tampa-St. Petersburg-Clearwater, FL 2,783,243 13
20 Baltimore-Towson, MD 2,710,489 26
San Juan, PR 2,700,000
21 Denver-Aurora-Broomfield, CO 2,543,482 18
22 Pittsburgh, PA 2,356,285 23
23 Portland-Vancouver-Hillsboro, OR-WA 2,226,009 22
24 Sacramento--Arden-Arcade--Roseville, CA 2,149,127 20
25 San Antonio-New Braunfels, TX 2,142,508
26 Orlando-Kissimmee-Sanford, FL 2,134,411 19
27 Cincinnati-Middletown, OH-KY-IN 2,130,151 34
28 Cleveland-Elyria-Mentor, OH 2,077,240 17
Vancouver, BC 2,200,000
29 Kansas City, MO-KS 2,035,334 31

oops, and Milwaukee:
39 Milwaukee-Waukesha-West Allis, WI 1,555,908 35