Related Work: Who Else Has Looked at This

Author

Justin Pietsch

Published

July 4, 2026

Draft

This document is the shorter map of everything else: the academic papers and the journalism that came before this report, what each one found, and where this report agrees, disagrees, or fills a gap nobody else measured. It checks one independent blog number by number, because that blog’s pipeline was close enough to mine to compare directly; the article series folds the same check into its three-point chapter.

Academic research

Three papers, in the order they were published. I read the full text of all three (not just the abstracts), so the descriptions are verified rather than secondhand.

The earliest to track the decline: Ribeiro, Mukherjee & Zeng (2016)

Ribeiro, Mukherjee, and Zeng, “The Advantage of Playing Home in NBA: Microscopic, Team-Specific and Evolving Features,” PLOS ONE is the earliest paper here to actually follow home advantage across seasons rather than describe it at a single point in time. It covers 16,133 games across 13 regular seasons, 2001-02 through 2013-14; playoffs aren’t addressed. Instead of box-score categories, it works from play-by-play data: how many more points per minute the home team scores, and how much faster its scoring comes relative to the away team’s. It finds the standard result, home teams score about 0.13 points per minute more than the away team, with the gap largest in the first quarter, and it reports that gap “slightly decreased” across its 13 seasons, an early, much narrower-window version of the same weakening this report tracks across four decades.

It also finds real team-to-team variation (Cleveland’s home scoring edge came in at roughly four times New Jersey’s, in its data), but doesn’t test how much of that spread is a genuine team effect versus what a small sample would produce by chance alone. That’s the same distinction this report tests directly for the franchise, referee, and per-team-decline findings elsewhere; worth naming as a difference in rigor, not a disagreement in substance.

The shot-mix account: Harris & Roebber (2019)

Harris and Roebber, “NBA team home advantage: Identifying key factors using an artificial neural network,” PLOS ONE covers 32 regular seasons, 1983-84 through 2017-18, playoffs excluded. It fed a model twelve shot-count inputs (2-point, 3-point, and free-throw makes, both by each side and by its opponents, home and away) and let the model find which ones track home advantage, rather than assuming a straight-line relationship up front. It explicitly tested attendance, altitude, and market size as inputs and found none of them helped the model, an early piece of evidence against the crowd-size story this report also rules out.

The paper’s finding lines up with the popular three-point story: home teams making relatively more 2-point and free-throw shots, and relatively fewer 3-pointers, than their opponents corresponds to a bigger home edge, and that mix shifted over the same 32 seasons that the edge shrank. The authors state plainly that fouls, blocks, and steals were left out and flag them as a direction for future work. That is exactly the gap this report fills: fouls (the narrowing whistle), rebounding, and turnovers turn out to carry more of the 40-year decline than shooting does on its own.

One caution on how to read it. The paper’s own language calls the shot-mix shift “responsible for” the change in home advantage, but what it actually shows is that the two moved together across the same three decades, not that one measurably led the other. That is the same gap this report found in the three-point hypothesis wherever it appears (see the three-point article): a real within-season effect once you separate it from the shared trend, but smaller and later than a moving-together chart on its own implies.

A different question: López-García et al. (2024)

López-García, Alonso-Pérez-Chao, Navarro Barragán, and Jiménez-Sáiz, “Home-Court Advantage and Home Win Percentage in the NBA: An In-Depth Investigation by Conference and Team Ability,” Applied Sciences (MDPI) asks a different question than this report does. It pools the regular seasons from 1999-2000 through 2022-23, using only final win/loss counts by venue pulled from a public scores site, no box-score data of any kind, so it can’t speak to fouls, shooting, rebounding, or turnovers the way this report or the Harris & Roebber paper above do. It deliberately excludes the 2020-21 season, the one played without crowds, since restricted attendance breaks its study design; that’s the same season this report leans on as its cleanest natural experiment for testing whether crowds matter at all. It does not include a year variable, so it’s a snapshot, not a decline study, the pooled 24 seasons are one sample, not a trend line.

Two splits, tested with standard group-comparison statistics rather than a regression or trend model. By conference: no difference in how much of a team’s wins come at home, but Western Conference teams win more of their home games than Eastern teams (64.5% vs. 58.5%), which the authors tentatively pin on higher average attendance in the West. By team strength (four tiers by season win rate): the better a team is, the more of its home games it wins, from 36.4% for the weakest tier up to 78.0% for the top (“contender”) tier. That’s a clean, sizeable spread, and a useful data point next to this report’s own playoff finding: a strong home team still wins big at home, while a weak one increasingly doesn’t.

One number in the paper is worth flagging rather than repeating uncritically. The text states contenders have a “significantly higher” share of their season’s wins coming from home games than weaker teams do, but the paper’s own descriptive table shows the opposite ordering: the weakest tier has the highest share, the contenders the lowest. That direction actually makes sense on its own (a team with fewer total wins naturally has a higher share of them coming from its stronger venue), so this looks like a mislabeled direction in the text rather than a real finding either way; noted here rather than silently picked one way or the other.

This paper’s snapshot split prompted a direct test, run here. Its team-strength split can’t speak to the 40-year decline (no year variable), but the underlying question, has quality’s grip on who wins at home changed over time, translates into a testable one, and this report runs it. The result diverges from what this report’s playoff seeding finding would predict: in the regular season, the gap between a clearly stronger and a clearly weaker home team has narrowed slightly over 40 years (34.9 points in 1984-94 to 28.2 points in 2023-26, p < 0.001), the opposite direction from the playoffs, where the equivalent gap widened sharply. Quality matters more for who wins a single playoff game than it used to; it matters very slightly less for who wins an ordinary regular-season game.

Background reading on officiating

A broader research literature exists on referee bias and home advantage across sports (crowd pressure on subjective calls, stoppage time, ball-strike counts, and similar). This report doesn’t adopt that literature’s framing or vocabulary; it describes the change in foul calls plainly as officiating growing fairer and more uniform, without characterizing it in terms of bias. Flagged here only as further reading for anyone who wants the deeper mechanism behind the whistle, not as something this report’s officiating section draws on.

Journalism and commentary

A handful of pieces made the rounds before this report, mostly building the same case with a chart and a paragraph rather than a full pipeline.

  • ESPN: Haberstroh, “Home-court disadvantage in NBA” (2015) is one of the earliest mainstream pieces to name the decline, narrative rather than data-driven.
  • Harvard Sports Analysis Collective, “NBA Home-Court Advantage is in Decline, Are 3s to Blame?” (2017; archived link, since the site’s own certificate has lapsed) makes the same three-point case as Harris & Roebber and the Sparkle blog this report checks number by number, in a shorter, undergraduate-level treatment.
  • PBS NewsHour, “What really causes home field advantage, and why it’s on the decline” is cross-sport, not NBA-specific, and points to instant replay reducing referee bias and better-understood travel and jet lag as the causes. Both are places this report actually disagrees. Travel distance moves home win rate by only about 0.07 points per hundred miles in this data, essentially nothing, and the NBA’s replay-review expansion mostly dates to 2014 onward, well after the decline was already decades under way. Worth naming as a disagreement, not a debate worth relitigating here.
  • A few smaller pieces (Three-Man-Weave, a Medium exploratory write-up, a Marc Stein Substack post, assorted Bleacher Report articles) repeat the same three-point or officiating themes without a new data pipeline behind them, so they aren’t treated individually here.
  • Sparkle Technologies published the one blog post checked number by number against this report’s own pipeline; the article series runs that comparison in the three-point article.

What this report adds

Measured against all of the above, three things stand out as genuinely new here, not just a re-run of someone else’s question:

  • Rebounding and turnovers as decline channels. No paper or article above measures either one. Rebounding turns out to be the single largest channel behind the 40-year decline, and it’s independent of the three-point shift.
  • A causal test on the three-point story, not just a moving-together chart. Every piece above (academic and otherwise) that names three-point shooting as the cause relies on the two trends tracking each other over decades. This report tests whether one actually leads the other and finds most of that season-level tracking is two shared trends, not cause and effect, with a smaller, real effect underneath once that’s separated out.
  • Regular season and playoffs, tracked side by side throughout, and tested against each other. All three academic papers above are regular-season-only (two say so explicitly, one doesn’t mention playoffs at all), and none of the journalism pieces separate the two either. The playoff seeding story here (a weaker home team compensating for the quality gap in the 1980s and 90s, and no longer doing so today) doesn’t appear anywhere else surveyed, and neither does the further step: running the same quality test on the regular season shows it moving the opposite direction from the playoffs, a genuine divergence nobody else was positioned to find without both contexts measured the same way.

None of this makes the earlier work wrong. Most of it asked a narrower question, or asked the right question without quite the tool needed to settle it. This report had the game log and the pipeline to take the next step.

How this was made: the writing here is mostly AI/LLMs working from my analysis and direction; the numbers are all Python on public NBA data. The full note is on the series hub.