BSW: The Case for Targeted Recounts

2026-03-14 | Dashboard

Executive Summary

BSW received 4.981%, missing the 5% threshold by 9,529 votes (0.019pp). This report examines whether the margin is small enough to justify targeted recounts.

Key findings:

Official corrections recovered +4,277 BSW votes (44.9% of deficit)
784 low-tail precincts show 5,145 excess missing votes (p=0.005)
4 powered forensic tests have 0% power to detect 9,529 votes spread across precincts
All 3 publicly specified affidavit cases matched to anomaly registry (8 filed total)
Strict BSW model (no 2025 Erststimmen) confirms predictions (R²=0.64)

The Official Gap

BSW received 2,472,947 valid Zweitstimmen out of 49,649,512 total (4.981%). The 5.000% threshold requires 2,482,476 — a deficit of 9,529 (0.019pp).

Official Corrections

Between preliminary and final results (Arbeitstabelle 9), BSW gained +4,277 Zweitstimmen — 44.9% of the deficit. BD lost −2,640.

Power Analysis

4 forensic tests (skew, Benford, geographic, negative-residual fraction) have 0% power to detect 9,529 votes spread across precincts. Even concentrated 953x10 patterns are only 15% detectable (skew test alone).

BSW Claims Assessment

1: BSW↔BD Ballot Confusion

Residual r=+0.028. No swap detected. Would need ~12.5% of BD.

2: Zero-Vote Precincts

784 low-tail precincts (81 zeros + 703 BSW>0). Null-calibrated excess: 5,145 votes (p=0.005).

3: Recount Extrapolation

50 BSW-selected recounts. Selection bias limits national extrapolation.

4: Official Corrections

+4,277 BSW (44.9% of deficit) via normal verification process.

Model Performance

95,046 precincts, GroupKFold(10) by Wahlkreis. LR: BSW uses strict features (no 2025 Erststimmen); other parties use base features.

Party	LR R²	XGB R²	Δ
BSW	0.6444	0.8085	+0.1640
AfD	0.9572	0.9896	+0.0324
CDU	0.9638	0.9863	+0.0225
SPD	0.8540	0.9276	+0.0736
GRÜNE	0.9057	0.9619	+0.0563
Die Linke	0.8577	0.9451	+0.0874
FDP	0.5944	0.7014	+0.1071
CSU	0.9824	0.9964	+0.0140
FREIE WÄHLER	0.7726	0.8772	+0.1046

Residual Distributions

Unimodal, no hidden subpopulation.

Turnout Correlation

Benford's 2nd Digit

Party	χ²	p
BSW	26.9	0.0014
FDP	56.5	0.0000
Die Linke	13.6	0.1390

Distribution Shape

Party	Skew	Kurtosis
BSW	+0.620	5.84
AfD	+2.433	21.66
CDU	+0.060	6.59
SPD	-0.176	4.73
GRÜNE	+0.892	7.28
Die Linke	-0.596	14.11
FDP	+0.364	7.52
CSU	+2.283	25.71
FREIE WÄHLER	+1.092	17.27

BSW skew=+0.62 (positive right tail). Missing votes would produce negative skew. Kurtosis 5.8 (leptokurtic; normal=3.0).

Conclusion

The 9,529-vote deficit is small enough that targeted recounts are justified. Official corrections recovered 44.9%, 5,145 excess low-tail missing votes are statistically significant (p=0.005), and forensic tests lack power to detect diffuse errors. The question is not settled — it requires recounts.