Geographic Matryoshka with US States

A challenge resulted from the recent Outside of California article, in a comment by loyal reader Rhodent. However, I have to say that I spent way more time than I’d care to admit seeking an optimal solution. Readers who relish a good mental challenge will probably enjoy trying to improve upon my results. Others will hopefully at least appreciate the obsessive-compulsive effort I endured.

I’ll go ahead and let Rhodent explain the challenge. One can also view the full comment in its original context if this snippet doesn’t provide enough information. Per Rhodent;

“I made some GeoGuessr puzzles where the theme was places with this sort of name. I would string them together into what I called ‘Geographic Matryoshkas’, although I didn’t stick to the idea that the city name had to be the name of another state… Maryland (neighborhood), London; London, Ontario; Ontario, New York; New York, Texas; Texas, Queensland… I’d be curious to know what would the longest string one could come up with if one did restrict it to using cities whose names were that of American states.”

---

OK, Challenge Accepted

So I undertook a brute force method. I searched for every state name found in another state in the U.S. Geological Survey’s Geographic Names Information System that fell within feature classes “Civil” and “Populated Place”. This partitioned options to places where people likely lived, i.e., “A political division formed for administrative purposes (borough, county, incorporated place, municipio, parish, town, township)” or a “Place or area with clustered or scattered buildings and a permanent human population (city, settlement, town, village).”

Thus I eliminated topographic features like rivers, mountains, forests and such. Some of the resulting locations may have been tiny or insignificant, and that was fine, I simply wanted to make sure they all came from an authoritative source. If it was good enough for the USGS then it was good enough for me. I called the resulting places “towns” even though some were counties, villages, etc., just to keep things as simple as possible.

Additionally I eliminated all partial matches. As an example, the result had to be New York exactly and not simply York. And yes, six different states actually had towns named New York, which certainly surprised me.

Then I eliminated over-matches so the town named Kansas in Alabama was acceptable while Kansas City in Missouri and its half million residents fell to the cutting room floor. I know! How arbitrary! I wanted to keep the game as pure as possible. That still left 299 state-named towns even after strictly applying the rules, leaving a manageable number of options in my opinion.

---

Everything went into a Google Docs spreadsheet

I couldn’t figure out how to make the embedded spreadsheet larger so participants in this game will definitely want to open the spreadsheet in another window. It included three tabs.

Validating Towns

The first tab displayed all of the valid town names I discovered, sorted by state. I appended two-letter state postal abbreviations onto each name to try to reduce confusion. Trust me, it got very, very confusing — downright exasperating — trying to keep track of which were states and which were towns with state names.

Jumping In and Jumping Out

The second tab compared potential movement between states that could be used to create linkages between them, which I called jumping IN and jumping OUT. Let’s use Maine as an example. There was a town in the state of Arizona called Maine. Arizona could be used to jump IN to Maine. Similarly, there was a town in Maine called Virginia. One could jump OUT of Maine to get to Virginia (and conversely, from Virginia’s perspective, Maine could be used to jump IN to Virginia).

It was important to understand these relationships to avoid dead-ends. The best states were those with lots of opportunities to do both, to jump in and out, like Ohio, Kansas, Iowa, Oklahoma and Texas. Those would be key. I noticed they were all located physically in the continental interior.

Winners

I figured that historically they were well-positioned to adopt names from east coast states when the original pioneers began moving inland, and they were still growing and creating new towns as the west coast states first came aboard, putting them in a unique position to adopt names from both directions. Fascinating. I would have never noticed that if I hadn’t undertaken this laborious exercise.

Losers

Other states were less useful, offering little or no means to jump in or out. They would create immediate dead-ends. Hawaii was the only state completely eliminated from the game. No state had a town named Hawaii, and Hawaii had no town named for any other state.

Others were bizarrely unbalanced. I thought Washington would be a great option given George Washington’s prominence upon the geographic landscape. Forty-two states had towns named Washington! Then the state of Washington had to ruin everything by having only a single town named for another state. Wyoming performed similarly. Seventeen states had towns of Wyoming (thanks to Gertrude of Wyoming, no doubt). Yet Wyoming failed to return the love, creating an immediate dead-end.

Raw Data

Tab three held all of the raw data I compiled as I pulled information manually from GNIS. I suspected that it wasn’t useful to anyone else although I was too proud of my effort to hide it from view.

---

Chain of Fools

With apologies to Rhodent, I began to think of the progression as links in a chain instead of a series of Russian nesting dolls. I also thought that Aretha Franklin’s version of Chain of Fools would be a perfect theme song as I frittered away hour-after-hour on my complex task. Go ahead and turn it on. Perhaps it will offer some inspiration.

My Result

Maybe you’ve read this far. Most readers, I suspect, gave up awhile ago. Those who stuck around finally get to see where it led.

1. Wyoming, DE
2. Delaware, VA
3. Virginia, OH
4. Ohio, CO
5. Colorado, TX
6. Texas, KY
7. Kentucky, MI
8. Michigan, VT
9. Vermont, WI
10. Wisconsin, MN
11. Minnesota, GA
12. Georgia, LA
13. Louisiana, MO
14. Missouri, OK
15. Oklahoma, KY
16. Kentucky, KS
17. Kansas, IN
18. Indiana, IA
19. Iowa, NE
20. Nebraska, IA
21. Iowa, OK
22. Oklahoma, TX
23. Texas, GA
24. Georgia, IN
25. Indiana, KS
26. Kansas, TN
27. Tennessee, TX
28. Texas, OH
29. Ohio, NY
30. New York, NM
31. New Mexico, MD
32. Maryland, NY
33. New York, TX
34. Texas, OK
35. Oklahoma, MS
36. Mississippi, AR
37. Arkansas, KY
38. Kentucky, IL
39. Illinois, KS
40. Kansas, OH
41. Ohio, GA
42. Georgia, AL
43. Alabama, FL
44. Florida, CO
45. Colorado, AK
46. Alaska, MN
47. Minnesota, CA
48. California, VA
49. Virginia, WA
50. Washington, CT

I built a chain with 50 links referencing 31 states. Not bad. I think there’s still room to improve upon it. The task started to become compulsive and I had to stop.

Your Results May Vary

Readers should feel free to embed additional links and loops within the chain, or use portions of the chain, or start a whole new chain from scratch. The rules are simple. Make the chain as long as you possibly can without repeating any towns. It might be useful to print the first tab of the spreadsheet and cross-out towns as you use them.

I’d be interested to see who can claim the most links, the most states referenced, and the most combined (links + states). The initial bar has been set at 50 links, 31 states, and 81 in total. Good luck and have fun. I’m done!