This is by far not the biggest issue facing the nation. But it’s fun. You apply rational thought and come to a definitive conclusion without having to worry about too many subtleties. Our goal? To define a method of voting that cannot be “gamed” by manipulating the boundaries of voting districts. We provide a mechanical solution that has the following outcomes:

  • Voters chose from among a limited number of candidates vying to represent their voting district, using a ‘first past the post’ voting system
  • Voting districts represent geographic areas that are subsets of the state geography (except perhaps in the case of sparsely populated states, e,g, Alaska).

We gather up names and locations of voters within a state, input this information into a computer, turn the crank, and out pops voting districts and assigned voters. Simple, elegant and foolproof.


  1. Gerrymandering
  2. The Mechanics
  3. Q&A

The Mechanics

Registered voters get randomly assigned to artificial, non-geographic voting tranches, two years in advance of the election. Geographic districts then are mechanically calculated based on the results of this assignment (example). The district boundary is (machine) drawn to maximize the number of voters within a tranche who live within 200 miles of each other (‘the 200 mile rule’). Voters who fall outside the calculated district boundaries for their tranche are then randomly reassigned into one of the other calculated districts such that 1) the voter lives within the district, and 2) all districts have approximately the same number of voters. Districts significantly overlap, so reassignment (the second randomization) does not ‘spoil’ the purity of the first randomization. Individuals’ assignments become public knowledge and candidates compete to represent the calculated voting districts.


Q. Why 200 miles?

We pick 200 miles as the diameter for a voting district based on the current numbers of districts and state sizes in the United States. It is important for randomization that the combined geographical size of all voting districts be 2 to 3 times that of the state (except for states with only 1 district). We need heavily overlapping voting districts. Voters who fall outside their district as a result of the 200-mile limit must be able to be reassigned during the second randomization into one of several districts within which they live (hence the overlaps).

Smarter mathematicians may be able to come up with diameters that more adequately meet the needs of states of all sizes and population densities. I am not a fan of one-size-fits-all. But for purposes of illustration, the 200-mile limit suits our needs. We seek a diameter that allows “local geographic interests” to be represented (i.e., small diameters), but also results in voting districts that are heavily overlapping (i.e., larger diameters). The optimal diameter will be a function of the geographic size of the state and its number of allotted voting districts.

Q. What about geography within the state?

Voters are for the most part geographically located within 200 miles of their fellow voters. This provides for representation of local (geography-based) interests. This rule is only violated in sparsely populated geographies.

There are no predefined geographical boundaries, but over time, population densities will cause voting districts to coalesce into semi-predictable boundaries. Imagine an urban center surrounded by poorly populated areas. Voting districts will act like the circles in a Venn diagram, with each circle encompassing most of the urban center. Voting districts (the circles) will ‘oscillate’ around this center from election-to-election.

In large, populated states many voters will end up with the same co-voters as in the previous election. This is based on the 200 mile rule. However, no two specific voters will likely remain in the same district for consecutive elections.

Q. I like to vote with my wife.

The schema illustrated here is based on randomized voters. The purpose is equally achieved if instead we set it up for randomized households.

Q. Circular geography does not adequately group voters according to their interests. We have a linear manufacturing region across the center of the state, surrounded north and south by rural areas.

“Their interests” is often an euphemism for gaming the redistricting process.

If the linear manufacturing region has a single common interest then they must convince enough of their rural neighbors that this interest is compatible with rural interests. Voters from the high-density manufacturing region will be over-represented in all voting districts and can therefore potentially win more districts than would be the case with having the manufacturing region wholly segregated. There is a greater risk that rural voters will have their interests under-represented, which they can overcome only through solidarity to rural values.

Districts will only roughly be circular, depending on the population densities and geographies.

Q. How about the parties?

Each party will work hard to find candidates most suitable for each defined district, much like today. Though, the districts can potentially be much more diverse in their issues and interests.

Local party bosses will have to be more flexible. There is still a geography-bias to voting (below), but it’s not inclusive of everyone within that geography. And the geography evolves (shifts or oscillates) over time.

Campaigns will have to reorganize their internal machinery. An approach emphasizing strictly geographical boundaries will place them at a disadvantage.

Q. What about minority interests?

Minorities will be represented in proportion to their percentage of the population, on average. There is no way to deliberately marginalize a defined group of voters (though ‘rural voters’ will now be more marginalized)

We don’t set aside districts for minorities. Minority-drawn districts are synonymous with “gaming” the voting system.

The interests of manufacturing or urban center voters will dominate within most of the voting districts, since the 200-mile rule favors high population densities. This redounds to the negative for rural voters. On the other hand, urban center voters will be heavily fractionated across many voting districts that include rural voters. A politically-divided urban voting population can often be offset by solidarity in heavily rural areas. Citizens in the north of Denver will have significantly different concerns from those in the south of Denver.

Q. How does this keep ‘big money’ from influencing the outcome?

It doesn’t. It only eliminates ‘gaming’ the system though the use of redistricting. For example, in states that are heavily unionized it can be expected labor unions will find it worthwhile to ‘target’ voting districts that randomly end up with a sizable union presence. This is true for any voting group that ‘self-discloses’ their preferences before an election. It does becomes much more challenging to define toss-up districts, which is a good thing.

Q. How do we protect an incumbent?

We don’t. Candidates who learn to best articulate the interests of “urban centers” will, on average, have an advantage. On average, urban center voters will be over-represented in all voting districts. But each election brings a new population of voters for each candidate, and they must learn to have at least something for non-urban voters who may tip the balance in the next election.

There is nothing that prohibits a candidate from running within any voting district. So parties can field their incumbents into voting districts being most favorable for those incumbents. There should be no rule that states a candidate must live within their voting district (but even that’s not a problem for any well-financed candidate).

Q. But I love my incumbent.

Candidates will come to feel they represent all the voters in their state (or at least their third- to half- of the state), and will do favors for any voter who can potentially vote for them in the future. Work to get your favorite incumbent re-elected no matter the voting district in which he or she is running. Incumbents will be back in your district in future elections.

Q. Voters will be confused.

Not any more than today. The campaigns will take great pains to ensure voters know in which election they are voting. Voters and campaigns have a two year advanced notice.

Although in the examples for this proposal we use non-descriptive labels for each voting district, in practice labels will not be much different than today’s.

Voters will not be able to ask any ole’ next-door neighbor about the next election, since the intent is for addresses within a neighborhood to be randomized across several voting districts. However, in a neighborhood of ten households, for example, at least 3 or 4 will be voting in the same district, and this information is publicly available.

Q. What is the simple message about this proposal to voters?

Once every two years you will be assigned a voting district that best reflects the other voters surrounding your home address. You will receive notification of your voting district assignment ~two years in advance of the next election. Voting district lines change from election to election (every two years), but these lines will most often have you voting with many of the same neighbors from one election to the next.

Q. You ignore the fact that elections are all about powerful interests getting their way under the illusion of fairness. You ignore political reality.

I’m all for finding ways to keep stupid voters away from the polls, but gerrymandering often gets used to encourage just the opposite. There are many more effective ways to disenfranchise stupid voters. Gerrymandering has reached a stalemate. Both sides know all the tactics. This proposal is merely a means of breaking the deadlock in a way that is non-threatening to both parties.

Q. How do we campaign?

Mass media and many more town halls. We give labels to each neighborhood (e.g., A, B, C…), and voters and candidates must needs find each other. Voters will often be within 200 miles of each other. The parties will have all their demographics and computer printouts as usual, and can stroke voters who have self-identified themselves as favorable, and ignore those voters who are not, just like today. You know the geographic location of the undecideds within your district, and you target them with all the usual predictive tools.

Q. How is this different from a state-wide election, for example the vote for the Senate?

We ensure representation of the interests of smaller geographic areas within a state.

Q. How do we get this implemented?

Unfortunately, each side will accentuate their losses under this proposal and be suspicious of any potential gains. Fancy political consultants will keep both parties anxious over this proposal for as long as they can, to maximize their consulting fees. You can’t expect both parties to embrace this simultaneously, since at any moment one or the other party will have a majority in today’s redistricting process. It’s probably best to implement this on a state-by-state basis as the sway between parties becomes more or less equal.

Of course this approach would have been impractical before the advent of the computer. Doing the calculations and reaching the voters is a piece-of-cake. Supreme Court rulings on geographies get in the way. This proposal may only serves as a model for a newly formed government.

Q. Isn’t this just an exercise in pedantry?

Absolutely. But there are folks out there clamoring for reform who are proposing inferior alternatives. So here’s some more to chew on.

Most gerrymandering reformers take as their goal ‘determining the will of the people’ as though voters were able to make informed choices about which candidate would actually best represent their ‘will’. They argue for solutions that blindly increase voter turnout (i.e., enfranchising even more poorly informed voters into the contests). Voter turnout has not (and probably cannot) be linked to quality of government.

So the difference comes down to “voters feeling they’ve had their say”, and since we’ve had no revolution with gerrymandering as its slogan, it appears gerrymandering does not ‘affect’ election outcomes. Gerrymandering only really matters to a handful of purists – folks uncomfortable with messiness in the real world.

The ‘google-ization’ of elections is a much more worrisome issue (much more liable to swing elections). Read Bernays’ ‘Crystallization of Public Opinion’ to get a sense of how much we can really be manipulated. We can only ‘know’ what the senses allow and there are experts dedicated to ensuring what reaches our senses is slanted (in a very subtle and sophisticated way) toward their own interests.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s