Congratulations, Chicago! Your city is a finalist for Amazon's 50,000-job second headquarters… along with 19 other cities. As shortlists go, it's a long list, culled down from 238 bids from North America. At first glance, it's more or less what you'd expect—all the cities that seemed like real candidates from the beginning are still candidates now, so it's not terribly big news.

Or is it? Since it's always good to check your instincts, I decided to take a look at what the conventional wisdom had highlighted as the most likely options for HQ2 and see how well the wisdom of the relatively well-informed crowd did.

Conveniently, Sperling's did a big metaranking of predictions from Moody's, the New York Times, famous urbanologist Richard Florida, talented economist and demographer Lyman Stone, CityLab, Quartz, Slate, and more—18 different sets of predictions of varying length.

The consensus was pretty close. Here's the top 20 and whether they made Amazon's first cut. I've highlighted the ones that didn't.

1. Atlanta: Y
2. Boston: Y
3. Chicago: Y
4. Philadelphia: Y
5. Washington, DC: Y
6. Austin: Y
7. Dallas: Y
8: Denver: Y
9. New York: Y
10. Raleigh: Y
11. Pittsburgh: Y
12. Toronto: Y
13. San Jose: N
14. Salt Lake City: N
15. San Francisco: N
16. Minneapolis: N
17. Portland: N

18. Nashville: Y
19. Miami: Y
20: Los Angeles: Y

Not bad. Here are the rest, with their Sperling rank:

Indianapolis (34)
Columbus (35)
Northern Virginia (44)
Newark (66)
Montgomery County, Maryland (Unranked)

You could throw Montgomery County in with the Northern Virginia suburbs at 44, but they're still a mild surprise. Or are they? This very interesting thread points out that Amazon Web Services—which appears to be the company's big profit center right now—chose suburban-DC Fairfax County for its East Coast corporate campus.

That makes a lot of sense for AWS, because it's heavily centralized in Northern Virginia, because a lot of cloud computing resources are centralized there, for historical reasons. And they're adding a lot more.

It's not the first time that someone has pointed this out. Late last year a Goldman Sachs analyst floated the idea that Amazon's second headquarters could be related to a spinoff of AWS. Whether or not they have any plans to do so, the importance of AWS to Amazon and the importance of Northern Virginia to AWS could go a long way towards NOVA and Montgomery County making the list (Takoma Park in Montgomery County is about 30 miles from the AWS headquarters in Herndon, Virginia).

Speaking of AWS, it may be partially the reason that Columbus made the finalist list. In 2016, the company set up an AWS "region," US East (Ohio), to complement and take pressure off the one in Northern Virginia. It set up three data centers at a cost of $1.1 billion dollars in Hilliard, Dublin, and New Albany—all suburbs of Columbus. Of course Columbus is also thriving generally, due in large part to the massive university at its core, which makes it an appealing location for tech companies. So AWS's location there and Columbus's presence on the finalist list may be more indirectly related than directly related, but either way it makes some sense to see it there.

Newark? It's convenient to New York, and New Jersey has offered Amazon an astonishing seven billion dollar tax break.

That leaves Indianapolis. Surprise! Or also not. I asked Twitter for its wisdom, and got the answer that they host a lot of tech-company offices. Salesforce, the customer relationship-management and cloud computing titan is probably the most famous one; they're headquartered in San Francisco, but last year they moved into the tallest building in Indiana in order to house over 2,000 employees. Infosys is adding another 2,000. And the city has been adding tech jobs at a considerable pace:

In all, the Indianapolis region added nearly 9,200 digital services jobs between 2010 and 2015, according to Brookings calculations. That’s more than the nearby metro areas of Cincinnati, Ohio; Columbus, Ohio; and Louisville, Ky.; combined. It’s also the most of any Midwestern metro area with the exceptions of Chicago and Detroit, both of which grew but did so at a slower pace.

It's still probably a longshot. Pete Saunders argues that they haven't made their "livability" cut yet, which is what the New York Times cut Indianapolis for.

What can we take from their decision? Amazon doesn't want to be in the Bay Area; that's about it. A couple potential sites that observers favored didn't make the cut, particularly Charlotte (good airport, tech-job growth) and Minneapolis (good transit, quality of life), but if Amazon had cut the list down to 10, their absence wouldn't be that big a surprise. A handful of surprises did make the cut, but there's some logic to them. The speculation industry, and the process of courting, will remain strong for a while.