The Danish data shows what should never have happened: Identical product, completely different outcomes. Some batches cause illness. Some do — nothing at all. #
In December 2025, Hamburg mathematics professor Hans-Juergen Bandelt published the first part of a three-part analysis on tkp.at. His starting point: the Danish study by Schmeling and Manniche, peer-reviewed and published in the European Journal of Clinical Investigation in 2023 [1]. His finding: More than 30% of BioNTech batches were practically ineffective — and the PEI knew it [2][3][4].
The Danish Study: Three Colors of Death #
The study by Max Schmeling and Vibeke Manniche examined all BioNTech doses administered in Denmark and classified the batches by their adverse event rate into three categories [1]:
Blue — Early batches with extremely high adverse event rates. In Germany, given as priority in nursing homes to the very elderly.
Green — Batches with a “normal” adverse event rate. The standard product, as expected.
Yellow — Batches with virtually zero adverse events. No reaction. No effect. Nothing.
The problem: With an identical pharmaceutical product, this variation should not exist. Every batch should be identically composed. That is the fundamental prerequisite of drug approval.
The Mathematics Behind It #
Prof. Bandelt does the math [2][3][4]:
- In the later phase of the campaign, batches were in circulation at a ratio of 2:1 (Green:Yellow)
- This yields an efficacy rate q = 0.6 — only 60% of doses were actually effective
- A triple-boosted individual had a 3.7% probability of having received no effective dose at all
- The probability of actually being effectively boosted was only 29.6%
The formula is simple: With q = 0.6 and three injections, the probability that all three were effective is 0.6 cubed = 21.6%. The probability that none were effective: 0.4 cubed = 6.4%.
This explains why so many boosted individuals say: “The vaccine didn’t do anything to me.” Correct — because they may have received a placebo.
The PEI Knew — and Did Nothing #
The Paul-Ehrlich-Institut (PEI) — Germany’s federal authority for vaccine safety — responded to the Danish study with an official statement: “No batch-specific clusters of suspected adverse event reports” [5].
That sounds reassuring — until you learn how the PEI received its batch samples: directly from the manufacturer [2]. No independent sampling. BioNTech delivered the samples to the PEI that the PEI was supposed to test. You let the defendant select the evidence.
Then, on November 28, 2024, the PEI quietly uploaded adverse event data broken down by batch. This data confirmed exactly what the Danes had shown in 2023: massive batch-dependent differences [2].
The most frequently reported Moderna batch: 3001651 — with 4,710 suspected adverse event reports. At an estimated underreporting factor of 10, that would be over 47,000 actual cases from a single batch [2].
Attorney Tobias Ulbrich (Rogert & Ulbrich) analyzed the monthly batch distribution and stated: “Anyone who sees the numbers in these tables must feel sick, because they suggest the systematic planning of harm” [2][6].
IgG4: The Biochemical Proof #
In Parts II and III, Bandelt connects the batch problem with the IgG4 class switch — an immunological phenomenon after repeated mRNA injections [3][4]:
- At q = 0.6, Bandelt calculates that 58% of effectively vaccinated individuals exhibit an IgG4 class switch
- IgG4 is a “tolerogenic” immunoglobulin — it signals the immune system: Ignore this antigen
- The immune system learns to tolerate rather than fight spike proteins
- “Hybrid immunity” (infection + vaccination) is promoted by STIKO as an advantage — Bandelt shows it is a hybrid injury [4]
A case from the 3sat NANO documentary (October 2025): A nine-year-old girl overcame COVID twice. 14 days after recovery, she received a BioNTech injection. Afterward: bedridden for weeks, permanently ill. A Frankfurt hospital diagnosed: “Must be psychologically caused” [4].
What This Means for “95% Efficacy” #
The famous Pfizer number — 95% efficacy — comes from the approval study. But if over 30% of batches were placebos:
- The control group unknowingly contained “vaccinated” individuals who received placebos
- The vaccine group contained people who received placebos and were counted as “vaccinated”
- The actual efficacy is mathematically not derivable from the published study data
The 95% assumes every dose was identical. This assumption is demonstrably false.
The Craig Paardekooper Database #
Since 2022, British researcher Craig Paardekooper has documented batch-specific adverse event rates using the US VAERS database [7]:
- Pfizer batch FC3661: 51 reported deaths
- Janssen batch 212C21A: 60 reported deaths
- Moderna batch 032H20A: 39 reported deaths
Data publicly accessible at HowBadIsMyBatch.com.
Connection to the Enquete Commission #
On March 19, 2026, former Pfizer toxicologist Helmut Sterz testified before the Bundestag’s Corona Enquete Commission:
“According to RKI protocols, approval was fast-tracked. I know of no other case in which studies were omitted in this manner.” [8]
He spoke of 60,000 deaths from vaccinations in Germany and demanded the immediate ban of mRNA technology [8].
The batch data provides the statistical foundation for Sterz’s statements. If batches were differently composed, the entire approval study was worthless — it presupposed product homogeneity.
The Questions the PEI Must Answer #
- Why did over 30% of BioNTech batches show no measurable effect?
- Why did the PEI receive test samples directly from the manufacturer?
- Why were batch-specific reports not published until November 28, 2024 — two years after the Danish findings?
- Why did the PEI claim “no batch-specific clusters” when its own data shows the opposite?
- Who decided which batches were delivered to which countries?
Bandelt suspects Asian countries did not receive placebo batches — the IgG4 class switch occurred faster there [4]. If true, the distribution was not a manufacturing defect, but a distribution plan.
Sources #
[1] Schmeling, M. & Manniche, V. (2023): Batch-dependent safety of the BNT162b2 mRNA COVID-19 vaccine. European Journal of Clinical Investigation, 53(8). doi.org/10.1111/eci.13998
[2] Bandelt, H.-J. (2025): Teils wirkungslose Dosen. Teil I. tkp.at. tkp.at
[3] Bandelt, H.-J. (2026): Teil II. tkp.at. tkp.at
[4] Bandelt, H.-J. (2026): Teil III. tkp.at. tkp.at
[5] Paul-Ehrlich-Institut (2023): Statement. pei.de
[6] Ulbrich, T. (2025): X/Twitter. @AnwaltUlbrich
[7] Paardekooper, C. (2022): HowBadIsMyBatch
[8] Welt.de (20.03.2026): Corona-Enquete-Kommission. welt.de
[9] 3sat NANO (2025): YouTube
[10] Sano et al. (2026): PMID 41759245
If a product fails to work in over 30% of its batches, you normally call that a manufacturing scandal. If it is a vaccine that 52 million Germans received as a booster, you call it — Pfizer Roulette. The bullet didn’t hit everyone. But the cylinder was deliberately spun.
