# District Compactness Measures Compactness is the intuition that a legitimate district should be a reasonably tidy blob rather than a writhing tendril. Roughly half the states impose some compactness requirement, and courts have relied on shape as evidence of manipulation since the original 1812 salamander. The trouble is that "compact" has no canonical mathematical definition, and the competing definitions disagree. ## The principal measures **Polsby-Popper.** The ratio of the district's area to the area of a circle with the same perimeter: `4πA / P²`. A circle scores 1; anything jagged scores near 0. This is the most-cited measure and the most sensitive to boundary detail. **Reock.** The ratio of the district's area to the area of the smallest circle enclosing it. Penalizes elongation. A long thin rectangle scores badly even with a perfectly smooth boundary. **Convex hull.** The ratio of the district's area to the area of its convex hull — the smallest convex shape containing it. Measures how much the district "reaches around" things. **Population polygon and related population-weighted variants.** Same idea as convex hull, but weighted by where people actually live rather than by land area, which matters enormously in states with large empty regions. **Schwartzberg, Length-Width, Minimum Convex Polygon** and a dozen others exist. There are more than thirty published compactness measures. ## The three problems **They disagree.** The same district routinely ranks well under one measure and badly under another. A mapmaker defending a map picks the flattering measure; a challenger picks the damning one. Because no measure has legal primacy, this is not cheating — it is the predictable consequence of leaving the term undefined. **They are sensitive to irrelevant detail.** Polsby-Popper depends on perimeter, and perimeter depends on how finely you trace a coastline or a river. This is the coastline paradox — a measurement that changes with the resolution of the map rather than with anything about the district. Two people computing Polsby-Popper from different-resolution census shapefiles get materially different answers for the same district. Perimeter has no fixed value at any resolution — a fractal property of real boundaries — so this is not a fixable data-quality issue. **They measure the wrong thing.** This is the deepest objection. Compactness is a proxy for manipulation, and it is a poor one in both directions. Perfectly compact maps can be aggressively gerrymandered — if the underlying political geography cooperates, you can pack and crack with tidy rectangles. Meanwhile genuinely fair maps are often ugly, because they follow rivers, county lines, municipal boundaries, or the actual shape of a community of interest. **A compactness requirement can force a mapmaker to split a community in order to make a prettier polygon.** ## Why they persist anyway Compactness measures survive because they are *legible*. A judge can look at a district shaped like a hook and a number confirming it scores 0.06 and reach a conclusion. That legibility is legally valuable in a way that a distributional argument about ensembles is not. They also serve as a **constraint rather than a test** — which is their best use. In [[Markov Chain Ensemble Methods]], compactness scores are used to define the space of "reasonable" maps that the sampler explores. Used that way, the measure does not have to detect gerrymandering; it only has to describe what a plausible map looks like, and any of the standard measures is adequate for that. ## The RWKB angle Compactness litigation is a good demonstration of a general pattern this branch tracks: **when a legal standard requires a number and the underlying concept admits many numbers, the choice of number becomes the site of the political fight, and it is a fight conducted where almost no one is watching.** The party with more analytic capacity picks the measure, frames the baseline, and supplies the expert. That advantage compounds quietly across dozens of state proceedings. ## Related [[The Mathematics of Redistricting]] · [[Markov Chain Ensemble Methods]] · [[Packing and Cracking]] · [[Gerrymandering]] · [[Mathematics and Power Hub]]