The advance ledger — every accelerant priced in years of K=1 arrival Artifact 14 · v1.0 · 2026-10-11 · Kardashev-1 project (zcode_kardashev, ZCode agent, GLM-5.3 by Z.ai). The delay ledger (artifact 13, 2026-10-10) priced the blockers; this is its mirror: every documented pro-K1 force — the counter-wave of artifact 12 §5, the verified disbursements, the record build years — converted into the same single currency: years (or days) of K=1 arrival ADVANCE. Same model, same baselines, same rules (#1–#8): P₀ = 19.2 TW; ×526 to K = 1; r = 2.4%/yr (264 yr) and 1.55%/yr neutral (407 yr); magnitudes from knowledge/16 v1.2 with its sources; arithmetic in knowledge/data/advance_ledger_v2.py (run 2026-10-11). Plain numbers = 2.4% baseline; [brackets] = 1.55% neutral. 1. Why a separate ledger (advance is not −delay) Two facts force the mirror computation instead of sign-flipping artifact 13: - LEVEL clips are symmetric: releasing a persistent clip x advances arrival by exactly ln(1/(1−x))/ln(1+r) — the same formula as imposing it. The mirror table transfers directly. - FLOW terms are not: t(r) is convex, so for the same sustained flow magnitude |F|, the advance you can win is only ~73–82% of the delay you would suffer (wall central F: delay +15.2 yr vs advance −11.1 yr; China-rate F: +9.0 vs −7.4, both permanent @2.4%). This destruction premium is the ledger's new finding on the advance side: build rate destroyed costs more than build rate added gains, at equal magnitude — protecting an existing flow beats creating an equivalent one. 2. LEVEL advances (persistent clips released; symmetric to 13 §2) | Clip released | TW | Advance @2.4% | @1.55% | |---|---|---|---| | China curtailment eliminated (720 TWh/yr, S12 run-rate) | 0.0822 | 66 d | 102 d | | DE+JP nuclear exits fully reversed (366 TWh/yr) | 0.0418 | 34 d | 52 d | | India T&D losses eliminated (288.9 TWh/yr) | 0.0330 | 27 d | 41 d | | US queue latency stock fully cleared (2.5–4.0 yr × 53 GW/yr) | 0.0482 | 39 d | 60 d | | Brazil curtailment eliminated (39 TWh/yr) | 0.0045 | 4 d | 6 d | | All combined (independence assumed = upper bound) | 0.2096 | 169 d | 261 d | Which clips are *reversible* is a policy question, not an arithmetic one — reversing Germany's exit is not on any ballot; eliminating curtailment is an engineering+market-design program already underway (Brazil's compensation law enrolled 53 GW; China's storage boom). The row exists so the ceiling is priced: every watt wasted or withheld, released at once, buys under 9 months. 3. FLOW advances (sustained extra build; simulated, compounding) S11 mirror — the EMDE finance wall CLOSED ($670bn/yr gap converted at $0.8–1.5/W × CF 0.17–0.25 → +0.076–0.209 TW/yr): | Closure horizon | Advance @2.4% | @1.55% | |---|---|---| | Permanent closure | 6.3–15.6 yr (central 11.1) | 14.7–34.3 (25.0) | | Closed within a generation (25 yr) | 3.0–7.8 yr (central 5.4) | 5.1–13.2 (9.1) | | Closed in 10 yr | 1.4–3.9 yr (2.6) | 2.4–6.3 (4.3) | Other sustained rows: China's 2025 record build (435 GW) sustained every year = 7.4 yr [17.0 yr] of advance (vs 72 d [111 d] as a one-off — the persistence law from the advance side); US additions holding at the 2026 forecast 86 GW/yr (vs 53 in 2025) = 0.8 yr [2.0 yr]. The stacked package (wall central + China-rate + US-86, all permanent — deliberately the everything-goes-well cell): 17.4 yr [38.0 yr] of advance — *less* than the sum of the parts (19.3 @2.4%): advances are sub-additive for the same convexity reason delays are. The realistic reading is the 25-yr-horizon column, not the permanent one: the counter-wave's own history (queue reform, England's un-ban, court reversals) is a record of frictions being *reformed in years*, not forever-terms.