BSAHI — SCCR hardening (P2) Produced: 2026-09-17 · Instrument: tools/research/sccr_sensitivity.py
The current reading is SCCR 0.2883 (2026-09-17, 140 blocks, N=32,000, T=10, C=$925). Across the assumption grid the index spans 0.028 – 3.413 (median 0.288), and stays below 1.0 in 22 of 27 combinations (81.5%) — it crosses only at the corner most favourable to coverage (high N, long T, high C).
The node count N dominates the uncertainty (10.0× across the tested range), so the honest statement is "SCCR is below 1× across essentially all plausible assumptions, dominated by uncertainty in N".
The SCCR is analytic in its assumptions:
R_blocks = 365.25 × 24 × 6 (blocks/year)
L_net = N × T × C / R_blocks (USD/block)
SCCR = fee_USD_per_block / L_netSo SCCR is inverse-linear in N, T and C, and linear in the fee. The bands are exact for the stated ranges; no Monte-Carlo approximation is needed for the grid. (A bootstrap is used only for the distribution CI below, where it is the right tool.)
| N (nodes) | T (yr) | C (USD) | L_net (USD/block) | SCCR | vs 1× |
|---|---|---|---|---|---|
| 10,000 | 5 | $500 | 475 | 3.413 | above |
| 10,000 | 5 | $925 | 879 | 1.845 | above |
| 10,000 | 5 | $1500 | 1,426 | 1.138 | above |
| 10,000 | 10 | $500 | 951 | 1.707 | above |
| 10,000 | 10 | $925 | 1,759 | 0.922 | below |
| 10,000 | 10 | $1500 | 2,852 | 0.569 | below |
| 10,000 | 20 | $500 | 1,901 | 0.853 | below |
| 10,000 | 20 | $925 | 3,517 | 0.461 | below |
| 10,000 | 20 | $1500 | 5,704 | 0.284 | below |
| 32,000 | 5 | $500 | 1,521 | 1.067 | above |
| 32,000 | 5 | $925 | 2,814 | 0.577 | below |
| 32,000 | 5 | $1500 | 4,563 | 0.355 | below |
| 32,000 | 10 | $500 | 3,042 | 0.533 | below |
| 32,000 | 10 | $925 | 5,628 | 0.288 | below |
| 32,000 | 10 | $1500 | 9,126 | 0.178 | below |
| 32,000 | 20 | $500 | 6,084 | 0.267 | below |
| 32,000 | 20 | $925 | 11,256 | 0.144 | below |
| 32,000 | 20 | $1500 | 18,252 | 0.089 | below |
| 100,000 | 5 | $500 | 4,753 | 0.341 | below |
| 100,000 | 5 | $925 | 8,793 | 0.184 | below |
| 100,000 | 5 | $1500 | 14,260 | 0.114 | below |
| 100,000 | 10 | $500 | 9,506 | 0.171 | below |
| 100,000 | 10 | $925 | 17,587 | 0.092 | below |
| 100,000 | 10 | $1500 | 28,519 | 0.057 | below |
| 100,000 | 20 | $500 | 19,013 | 0.085 | below |
| 100,000 | 20 | $925 | 35,174 | 0.046 | below |
| 100,000 | 20 | $1500 | 57,039 | 0.028 | below |
From the frozen reproduction capture — 155 real blocks (not a model):
| statistic | value |
|---|---|
| mean | 0.2406 |
| median | 0.1933 |
| IQR | 0.1158 – 0.2954 |
| P5–P95 | 0.0687 – 0.5831 |
| min–max | 0.0490 – 1.0757 |
| bootstrap 95% CI for the mean | 0.2131 – 0.2707 |
legitimate CI for the mean of a real 155-block sample; NOT a claim about the population of all blocks. (bootstrap n=10000, seed 20260917).
| assumption | range tested | SCCR at low | SCCR at high | spread |
|---|---|---|---|---|
| N | 10,000–100,000 | 0.922 | 0.092 | 10.0× |
| T | 5–20 | 0.577 | 0.144 | 4.0× |
| C | 500–1,500 | 0.533 | 0.178 | 3.0× |
Verdict: SCCR is below 1.0 in 22/27 combinations; it crosses only at the most favourable-for-coverage corner (high N, long T, high C).
# the full sensitivity + CI + stress test, deterministic
python3 tools/research/sccr_sensitivity.py
# the per-block reproduction that feeds the distribution (JS = Python = C)
bash research/reproduce/cross_check.shThe script reads data/sccr.json (the dated reading) and the frozen kit output, so both the grid and the CI are reproducible from the repository alone.
data/sccr_sensitivity.json — deterministic, frozen-input instrument; re-runnable offline.