# Packing and Cracking Every partisan gerrymander, however baroque the resulting shapes, reduces to two operations applied in combination. **Packing** concentrates opposition voters into as few districts as possible, where they win by enormous margins. Those districts are conceded. The purpose is to remove opposition voters from every other district in the state — a packed 85–15 district wastes roughly 35,000 surplus opposition votes that would have been competitive somewhere else. **Cracking** disperses the remaining opposition voters across many districts in shares too small to win — 45%, 44%, 46%. Every one of those votes is a losing vote and therefore wasted entirely. ## Why they are complements, not alternatives The operations are mathematically linked by a conservation constraint: **every voter you do not pack, you must place somewhere.** Packing is what makes cracking possible, because concentrating opposition strength in a few sacrificial districts is what leaves the rest of the state thin enough to crack safely. Run either alone and it fails. Pack without cracking and you have simply given the opposition a few very safe seats while leaving the rest competitive. Crack without packing and you spread opposition voters at levels that risk flipping districts in a bad year. The combination is what produces the characteristic gerrymander signature: a small number of grotesquely lopsided opposition districts, and a large number of your own districts won by comfortable-but-not-wasteful margins. ## The risk the mapmaker accepts An aggressive gerrymander is a **leveraged position**. Winning many districts at 55% is more seat-efficient than winning fewer at 70%, but it also means a uniform swing of six points flips the entire portfolio at once. This is the "dummymander" — a map so aggressive it collapses when the political environment moves against its architects. Several 2010-cycle maps did exactly this. Modern mapmaking manages this risk explicitly. With precinct-level partisan data and enough computing, you can evaluate a candidate map against a range of simulated environments and select for maps that hold up across a swing band rather than maximizing seats in a single scenario. **The optimization objective changed from "most seats now" to "most seats robustly," and that is a strictly more sophisticated mathematical problem than the one 1990s mapmakers were solving.** ## The racial dimension Packing and cracking are also the mechanics of racial vote dilution, and the legal treatment of the two motives is now sharply asymmetric. Under the [[Thornburg v. Gingles]] framework, cracking a geographically compact and politically cohesive minority community across multiple districts was the paradigm Section 2 violation. [[Allen v. Milligan]] (2023) affirmed that framework. [[Louisiana v. Callais]] (2026) then rewrote it to require proof of discriminatory intent rather than effect. The result is an inversion with real operational consequences: **packing and cracking undertaken to dilute minority voting strength is now easier to defend — by attributing it to partisanship — than drawing a district to remedy dilution.** Because minority voting blocs in the American South are also reliably Democratic blocs, "we did it for partisan reasons" functions as a general-purpose safe harbor, and [[Rucho v. Common Cause]] made partisan motive federally nonjusticiable. The two doctrines interlock. ## Detection Neither operation is visible from a single map. Both are visible from the *distribution* of district margins: a gerrymandered map shows a characteristic bimodal signature — a cluster of extremely lopsided opposition districts and a dense cluster of moderately safe mapmaker districts, with the middle hollowed out. Comparing that distribution against [[Markov Chain Ensemble Methods]] output is the standard forensic approach. ## Related [[Wasted Votes]] · [[The Efficiency Gap]] · [[Markov Chain Ensemble Methods]] · [[The Mathematics of Redistricting]] · [[Gerrymandering]] · [[Louisiana v. Callais]] · [[Mathematics and Power Hub]]