Papers · Preprint · v2

Decomposing the Bloch-Kato Fundamental Line: Adequacy and No-Go Results for the BSD Invariants

What one number about an elliptic curve keeps, and what it throws away.

In plain words

The Birch and Swinnerton-Dyer formula ties a curve to a product of numbers of different kinds. This paper sorts the richer data behind them into five columns: heights and regulator, analytic and period data, finite sources, local p adic data, and the assembly into one determinant line.

For each column a finite residual measures whether a given number is a linear readout of the admitted data, vanishing exactly when it is. The full data of each of the first four columns are adequate for their readouts, and the five columns together are adequate for a Bloch–Kato scalar normalized so that it forgets orientation. The comparisons from the columns to the Bloch–Kato line are stated as external inputs.

Two formal records of exponents of a prime p. In one, Sha has exponent 2 and the Tamagawa numbers 0; in the other, Sha 0 and Tamagawa 2; torsion 0 in both. The formula's reading, Sha plus Tamagawa minus twice torsion, is 2 for both. The pairing form is never read. Below, the actual curve 11a1 at p equals 5: Tamagawa 5 times Sha 1 divided by torsion 5 squared equals one fifth; Sha adds no factor of 5.
What the formula's single number keeps at one prime: a weighted sum of exponents. Put a factor p² in Sha or in the Tamagawa numbers and the reading is the same, and the form of Sha's pairing never enters at all. The two rows are formal records, not curves; the real curve 11a1 shows where its own factor of 1/5 comes from.

Other readouts lose information. The most concrete case: the curves 571a1 and 2045b1 share rank, Tamagawa numbers, torsion and the 2 torsion dimension of Sha, yet their 2 primary parts of Sha have orders 4 and 16. Many statements are checked in Lean 4 without external libraries, each labelled by exactly how much of it Lean proves.

What it shows

  • A five column typed decomposition of the BSD invariants.
  • Adequacy of each column's full data, and of all five for an orientation normalized scalar.
  • Three fixed normalization points, and explicit places where scalar data is not enough.

What it does not claim

It does not prove BSD. Deep arithmetic inputs are carried as hypotheses.

Cite

Tsiokos, I. (2026). Decomposing the Bloch-Kato Fundamental Line: Adequacy and No-Go Results for the BSD Invariants. Zenodo. https://doi.org/10.5281/zenodo.23130789