v2.5.0

What's Changed

• Add Stacked DiD estimator (Wing, Freedman & Hollingsworth 2024) by @igerber in #172
• Bump version to 2.5.0 by @igerber in #173

Full Changelog: v2.4.3...v2.5.0

v2.4.3

What's Changed

• Rewrite TripleDifference to match R's triplediff::ddd() by @igerber in #169
• Update TripleDifference summary table to reflect completed review by @igerber in #170
• Bump version to 2.4.3 by @igerber in #171

Full Changelog: v2.4.2...v2.4.3

v2.4.2

What's Changed

• Address tech debt from code reviews (PRs #115-#159) by @igerber in #165
• Clean up completed items from TODO.md and ROADMAP.md by @igerber in #166
• Add conditional platform BLAS linking for Rust backend by @igerber in #167
• Bump version to 2.4.2 by @igerber in #168

Full Changelog: v2.4.1...v2.4.2

v2.4.1

What's Changed

• Add tutorial notebook for Two-Stage DiD (Gardner 2022) by @igerber in #159
• Update TODO.md and ROADMAP.md for accuracy post-v2.4.0 by @igerber in #160
• Add matplotlib import guards to tutorials 11 and 12 by @igerber in #161
• Address TODO items: safe_inference migration, module splits, np.dot, and bug fixes by @igerber in #163
• Bump version to 2.4.1 by @igerber in #164

Full Changelog: v2.4.0...v2.4.1

v2.3.1

What's Changed

• SDID methodology review: rewrite to match R synthdid + Rust parallel variance by @igerber in #145
• Remove Rust outer-loop SDID variance to fix SE mismatch and perf regression by @igerber in #147
• Fix docs/PyPI version mismatch with versioned RTD builds by @igerber in #149

Full Changelog: v2.3.0...v2.3.1

v2.3.0

What's Changed

• Add Borusyak-Jaravel-Spiess (2024) Imputation DiD estimator by @igerber in #141
• Add MultiPeriodDiD vs R (fixest) benchmark by @igerber in #135
• Shorten test suite runtime with parallel execution and reduced iterations by @igerber in #140
• Fix TWFE within-transformation bug and add methodology review by @igerber in #139

Full Changelog: v2.2.1...v2.3.0

v2.2.1

What's Changed

• Fix NaN t-statistics across 7 locations for consistent undefined inference by @igerber in #118
• Align TROP lambda conventions with paper (Athey et al. 2025) by @igerber in #119
• Add git worktree management slash commands by @igerber in #120
• Add branch-cleanup slash command by @igerber in #122
• Fix worktree-new Rust backend rebuild: add maturin to dev deps and fix manifest path by @igerber in #121
• Fix worktree-rm: detect squash-merged branches via GitHub PR status by @igerber in #123
• Remove jackknife variance estimation from TROP by @igerber in #124
• Replace hardcoded GitHub URL in bump-version skill by @igerber in #127
• Remove LOOCV control observation subsampling from TROP by @igerber in #126
• Add /review-plan skill for staff-engineer plan review by @igerber in #128
• MultiPeriodDiD: full event-study specification with pre-period coefficients by @igerber in #125
• Update methodology review for MultiPeriodDiD by @igerber in #130
• Fix invalid baseRepository field in push-pr-update skill by @igerber in #131
• Add backend-aware test parameter scaling for pure Python CI by @igerber in #129
• Bump version to 2.2.1 by @igerber in #132

Full Changelog: v2.2.0...v2.2.1

v2.2.0

What's Changed

• Add Windows wheel builds using faer pure Rust linear algebra by @igerber in #115

Full Changelog: v2.1.9...v2.2.0

v2.1.9

What's Changed

• Add unified LOOCV for TROP joint method with Rust acceleration by @igerber in #113

Full Changelog: v2.1.8...v2.1.9

v2.1.8

What's Changed

• Align TROP with paper methodology (Equation 5, D matrix semantics) by @igerber in #110
• Add /push-pr-update skill for pushing PR revisions by @igerber in #111
• Release v2.1.8 by @igerber in #112

Full Changelog: v2.1.7...v2.1.8