Brief 02 · The case against
Is it a few apps, a relayer, or luck?
Every control here was written into the registered method before the out-of-sample data was read. Each reruns the same comparison with one assumption changed. If a control breaks the result, it says so in red.
Does the result survive every control?
The headline failed, and 1 of 8 reruns would pass the registered rule. A control that passes after the fact does not change the verdict: it was never the claim.
| Rerun | FC gain | DAA gain | Price gain | Weeks | Pass rule |
|---|---|---|---|---|---|
| Without the 10 most common function-call scopes | -13.1% | -3.9% | -8.6% | 46 | fails too |
| Without the 10 busiest function-call signers | -1.5% | -4.0% | -6.7% | 86 | fails too |
| FUNCTION_CALL only (meta transactions dropped from FC) | -4.7% | -4.1% | -7.0% | 84 | fails too |
| Every meta transaction counted as FC | 0.1% | -3.5% | -7.2% | 97 | would pass |
| Meta transactions removed from both sides | -2.4% | -4.1% | -7.0% | 84 | fails too |
| Every unknown key counted as FC | -5.3% | -3.6% | -5.5% | 94 | fails too |
| Every unknown key counted as full access | -5.1% | -3.6% | -5.5% | 94 | fails too |
| Flagged episode weeks left out of the scoring | -2.8% | -1.9% | -8.5% | 43 | fails too |
Source: data/out/controls.json; each rerun keeps the headline's in-sample lag and the baselines' lags.
Controls: analysis/controls.pyCould the timing be luck?
The real out-of-sample gain (-5.1 percent) beats 9 percent of 1000 timing placebos; the bar is 95 percent.
Source: Placebo: the FC log-change series circularly shifted by 1000 random offsets of at least 8 weeks (seed 20261114), out-of-sample gain recomputed each time. data/out/placebo.csv
Placebo: data/out/placebo.csvIs function-call activity just a few apps?
The ten contracts that function-call keys were most often scoped to carry 91.4 percent of all sampled function-call transactions.
| # | Key scoped to | Sampled FC tx | Share of FC | Weeks seen |
|---|---|---|---|---|
| 1 | app.nearcrowd.near | 2,495 | 40.8% | 127 |
| 2 | token.sweat | 1,419 | 23.2% | 170 |
| 3 | aa-harvest-moon.near | 486 | 7.9% | 119 |
| 4 | intents.near | 437 | 7.1% | 63 |
| 5 | v2_0_2.perp.spin-fi.near | 215 | 3.5% | 58 |
| 6 | asset-manager.orderly-network.near | 206 | 3.4% | 60 |
| 7 | inscription.near | 169 | 2.8% | 10 |
| 8 | v1.signer | 64 | 1.1% | 19 |
| 9 | aurora | 62 | 1.0% | 15 |
| 10 | tg | 44 | 0.7% | 30 |
| 11 | priceoracle.near | 42 | 0.7% | 33 |
| 12 | v2.ref-finance.near | 40 | 0.7% | 8 |
| 13 | login.learnclub.near | 31 | 0.5% | 29 |
| 14 | coin.abound.near | 29 | 0.5% | 16 |
| 15 | bridge-refuel.hot.tg | 28 | 0.5% | 20 |
Source: data/out/fc_receivers.csv: the receiver contract each function-call key was scoped to, across every sampled function-call transaction.
Scopes: data/out/fc_receivers.csvWhen was the signal wrong?
Function-call keys flagged 28 weeks: 13 hits and 15 false alarms, with 24 misses. DAA: 2 hits, 7 false alarms, 35 misses.
Function-call keys
Hits (13): 2025-01-06, 2025-02-03, 2025-04-28, 2025-06-30, 2025-07-28, 2025-08-04, 2025-09-15, 2025-09-29, 2025-11-03, 2026-04-13, 2026-05-04, 2026-07-13, 2026-07-27
False alarms (15): 2025-01-27, 2025-03-24, 2025-04-07, 2025-06-16, 2025-06-23, 2025-07-14, 2025-08-25, 2025-10-06, 2025-12-08, 2025-12-29, 2026-02-09, 2026-03-09, 2026-06-01, 2026-09-21, 2026-10-26
Misses (24): 2025-01-13, 2025-03-31, 2025-04-14, 2025-05-12, 2025-05-19, 2025-06-09, 2025-07-07, 2025-08-18, 2025-09-08, 2025-10-27, 2025-12-01, 2026-01-26, 2026-02-02, 2026-02-23, 2026-03-30, 2026-04-06, 2026-04-27, 2026-05-11, 2026-05-25, 2026-08-10, 2026-08-17, 2026-08-31, 2026-09-14, 2026-10-05
DAA only
Hits (2): 2025-01-06, 2025-05-12
False alarms (7): 2025-04-21, 2025-08-11, 2026-01-12, 2026-03-02, 2026-03-09, 2026-10-19, 2026-10-26
Misses (35): 2025-01-13, 2025-02-03, 2025-03-31, 2025-04-14, 2025-04-28, 2025-05-19, 2025-06-09, 2025-06-30, 2025-07-07, 2025-07-28, 2025-08-04, 2025-08-18, 2025-09-08, 2025-09-15, 2025-09-29, 2025-10-27, 2025-11-03, 2025-12-01, 2026-01-26, 2026-02-02, 2026-02-23, 2026-03-30, 2026-04-06, 2026-04-13, 2026-04-27, 2026-05-04, 2026-05-11, 2026-05-25, 2026-07-13, 2026-07-27, 2026-08-10, 2026-08-17, 2026-08-31, 2026-09-14, 2026-10-05
Price only
Hits (7): 2025-05-12, 2025-07-28, 2025-08-04, 2025-09-29, 2026-04-27, 2026-05-25, 2026-10-05
False alarms (10): 2025-05-26, 2025-06-02, 2025-10-06, 2026-03-23, 2026-06-08, 2026-06-15, 2026-09-07, 2026-09-21, 2026-09-28, 2026-10-12
Misses (30): 2025-01-06, 2025-01-13, 2025-02-03, 2025-03-31, 2025-04-14, 2025-04-28, 2025-05-19, 2025-06-09, 2025-06-30, 2025-07-07, 2025-08-18, 2025-09-08, 2025-09-15, 2025-10-27, 2025-11-03, 2025-12-01, 2026-01-26, 2026-02-02, 2026-02-23, 2026-03-30, 2026-04-06, 2026-04-13, 2026-05-04, 2026-05-11, 2026-07-13, 2026-07-27, 2026-08-10, 2026-08-17, 2026-08-31, 2026-09-14
Source: data/out/directions.csv: out-of-sample weeks; a signal or the outcome 'rose' when its log change beat its in-sample 67th percentile. The DAA rows are the committee that only had a DAA chart.
Directions: data/out/directions.csvWhich weeks look like bots or farming?
167 weeks are flagged: one function-call scope carried at least half the week's sampled FC transactions, or DAA reached three times its trailing 8-week median. No curated airdrop list is used, because none could be sourced for every week.
| Week | Top scope share of FC | DAA | Flag |
|---|---|---|---|
| 2022-W01 | 92.9% | 13,002 | one scope |
| 2022-W02 | 96.7% | 20,332 | one scope |
| 2022-W03 | 81.3% | 16,368 | one scope |
| 2022-W04 | 86.6% | 14,602 | one scope |
| 2022-W05 | 90.2% | 16,563 | one scope |
| 2022-W06 | 92.5% | 16,618 | one scope |
| 2022-W07 | 78.7% | 17,753 | one scope |
| 2022-W08 | 86.3% | 14,970 | one scope |
| 2022-W09 | 85.3% | 14,680 | one scope |
| 2022-W10 | 87.9% | 15,994 | one scope |
| 2022-W11 | 100.0% | 115,306 | one scope, DAA spike |
| 2022-W12 | 88.0% | 31,381 | one scope |
| 2022-W13 | 92.7% | 13,728 | one scope |
| 2022-W14 | 95.8% | 13,791 | one scope |
| 2022-W15 | 85.0% | 17,590 | one scope |
| 2022-W16 | 84.8% | 18,581 | one scope |
| 2022-W17 | 78.6% | 16,131 | one scope |
| 2022-W18 | 85.2% | 16,190 | one scope |
| 2022-W19 | 84.2% | 13,656 | one scope |
| 2022-W20 | 88.7% | 11,865 | one scope |
| 2022-W21 | 91.7% | 11,483 | one scope |
| 2022-W22 | 93.8% | 12,691 | one scope |
| 2022-W23 | 86.5% | 13,473 | one scope |
| 2022-W24 | 87.1% | 11,217 | one scope |
| 2022-W25 | 94.4% | 12,473 | one scope |
| 2022-W26 | 92.3% | 9,799 | one scope |
| 2022-W27 | 91.5% | 11,898 | one scope |
| 2022-W28 | 94.1% | 12,365 | one scope |
| 2022-W29 | 90.3% | 9,032 | one scope |
| 2022-W30 | 86.1% | 8,709 | one scope |
| 2022-W31 | 65.8% | 10,740 | one scope |
| 2022-W32 | 76.2% | 11,746 | one scope |
| 2022-W33 | 94.4% | 10,067 | one scope |
| 2022-W34 | 57.9% | 10,274 | one scope |
| 2022-W35 | 83.3% | 11,650 | one scope |
| 2022-W36 | 81.5% | 10,973 | one scope |
| 2022-W37 | 55.0% | 122,867 | one scope, DAA spike |
| 2022-W38 | 62.5% | 106,051 | one scope, DAA spike |
| 2022-W39 | 43.5% | 100,497 | DAA spike |
| 2022-W40 | 52.3% | 89,800 | one scope, DAA spike |
| 2022-W41 | 53.8% | 81,359 | one scope |
| 2022-W45 | 54.4% | 64,759 | one scope |
| 2022-W47 | 57.4% | 53,520 | one scope |
| 2022-W48 | 55.3% | 50,469 | one scope |
| 2023-W01 | 53.6% | 42,078 | one scope |
| 2023-W03 | 53.6% | 69,082 | one scope |
| 2023-W05 | 54.2% | 70,340 | one scope |
| 2023-W06 | 61.5% | 65,801 | one scope |
| 2023-W07 | 57.1% | 69,510 | one scope |
| 2023-W08 | 54.3% | 62,347 | one scope |
| 2023-W13 | 52.0% | 48,586 | one scope |
| 2023-W15 | 50.0% | 53,396 | one scope |
| 2023-W16 | 54.7% | 67,781 | one scope |
| 2023-W19 | 50.0% | 54,581 | one scope |
| 2023-W20 | 56.5% | 40,248 | one scope |
| 2023-W21 | 62.9% | 64,951 | one scope |
| 2023-W23 | 57.6% | 77,387 | one scope |
| 2023-W25 | 64.5% | 55,376 | one scope |
| 2023-W33 | 50.0% | 54,991 | one scope |
| 2023-W34 | 52.0% | 57,987 | one scope |
| 2023-W35 | 51.6% | 234,063 | one scope, DAA spike |
| 2023-W36 | 41.2% | 729,098 | DAA spike |
| 2023-W37 | 60.0% | 655,002 | one scope, DAA spike |
| 2023-W38 | 25.0% | 670,797 | DAA spike |
| 2023-W39 | 45.5% | 681,684 | DAA spike |
| 2023-W40 | 62.5% | 689,900 | one scope |
| 2023-W42 | 50.0% | 702,797 | one scope |
| 2023-W44 | 85.7% | 818,938 | one scope |
| 2023-W45 | 66.7% | 781,274 | one scope |
| 2023-W47 | 66.7% | 885,833 | one scope |
| 2023-W48 | 97.5% | 489,950 | one scope |
| 2023-W50 | 50.0% | 1,088,500 | one scope |
| 2023-W52 | 94.1% | 1,192,110 | one scope |
| 2024-W01 | 90.2% | 1,093,330 | one scope |
| 2024-W02 | 83.3% | 1,098,120 | one scope |
| 2024-W07 | 66.7% | 1,140,900 | one scope |
| 2024-W08 | 50.0% | 1,259,170 | one scope |
| 2024-W10 | 50.0% | 1,264,190 | one scope |
| 2024-W11 | 50.0% | 1,214,190 | one scope |
| 2024-W12 | 57.1% | 1,313,920 | one scope |
| 2024-W13 | 62.5% | 1,520,320 | one scope |
| 2024-W14 | 50.0% | 1,464,190 | one scope |
| 2024-W16 | 71.4% | 1,787,500 | one scope |
| 2024-W17 | 90.0% | 2,021,780 | one scope |
| 2024-W18 | 88.9% | 1,838,410 | one scope |
| 2024-W19 | 68.2% | 1,734,850 | one scope |
| 2024-W20 | 100.0% | 1,765,710 | one scope |
| 2024-W21 | 54.5% | 1,871,190 | one scope |
| 2024-W22 | 75.0% | 1,921,370 | one scope |
| 2024-W23 | 63.6% | 1,914,880 | one scope |
| 2024-W24 | 50.0% | 2,087,210 | one scope |
| 2024-W25 | 75.0% | 2,010,100 | one scope |
| 2024-W27 | 77.8% | 1,989,000 | one scope |
| 2024-W28 | 50.0% | 1,932,730 | one scope |
| 2024-W29 | 63.6% | 1,958,240 | one scope |
| 2024-W30 | 60.0% | 1,949,460 | one scope |
| 2024-W31 | 55.6% | 1,837,160 | one scope |
| 2024-W34 | 85.7% | 722,260 | one scope |
| 2024-W35 | 50.0% | 711,312 | one scope |
| 2024-W36 | 60.0% | 884,913 | one scope |
| 2024-W37 | 70.0% | 1,001,550 | one scope |
| 2024-W38 | 50.0% | 1,000,970 | one scope |
| 2024-W41 | 100.0% | 1,011,590 | one scope |
| 2024-W42 | 75.0% | 1,029,110 | one scope |
| 2024-W43 | 50.0% | 972,618 | one scope |
| 2024-W44 | 100.0% | 933,387 | one scope |
| 2024-W45 | 75.0% | 955,492 | one scope |
| 2024-W46 | 80.0% | 963,416 | one scope |
| 2024-W47 | 62.5% | 822,127 | one scope |
| 2024-W48 | 83.3% | 890,381 | one scope |
| 2024-W50 | 100.0% | 953,221 | one scope |
| 2024-W51 | 83.3% | 913,797 | one scope |
| 2024-W52 | 100.0% | 935,045 | one scope |
| 2025-W01 | 75.0% | 914,572 | one scope |
| 2025-W02 | 62.5% | 897,878 | one scope |
| 2025-W08 | 50.0% | 623,417 | one scope |
| 2025-W09 | 50.0% | 520,982 | one scope |
| 2025-W10 | 75.0% | 519,827 | one scope |
| 2025-W12 | 75.0% | 496,360 | one scope |
| 2025-W13 | 100.0% | 526,385 | one scope |
| 2025-W15 | 66.7% | 454,943 | one scope |
| 2025-W16 | 50.0% | 473,012 | one scope |
| 2025-W18 | 83.3% | 437,588 | one scope |
| 2025-W19 | 50.0% | 420,349 | one scope |
| 2025-W20 | 100.0% | 427,617 | one scope |
| 2025-W21 | 50.0% | 439,285 | one scope |
| 2025-W22 | 50.0% | 377,588 | one scope |
| 2025-W23 | 66.7% | 373,789 | one scope |
| 2025-W24 | 66.7% | 362,286 | one scope |
| 2025-W25 | 50.0% | 347,803 | one scope |
| 2025-W26 | 50.0% | 334,665 | one scope |
| 2025-W27 | 66.7% | 328,875 | one scope |
| 2025-W28 | 55.6% | 329,361 | one scope |
| 2025-W29 | 50.0% | 343,457 | one scope |
| 2025-W30 | 100.0% | 341,131 | one scope |
| 2025-W31 | 89.5% | 341,886 | one scope |
| 2025-W32 | 100.0% | 335,151 | one scope |
| 2025-W33 | 100.0% | 338,416 | one scope |
| 2025-W34 | 50.0% | 330,601 | one scope |
| 2025-W35 | 100.0% | 322,141 | one scope |
| 2025-W36 | 75.0% | 308,281 | one scope |
| 2025-W37 | 57.1% | 303,712 | one scope |
| 2025-W39 | 50.0% | 291,667 | one scope |
| 2025-W40 | 50.0% | 286,244 | one scope |
| 2025-W41 | 71.4% | 278,713 | one scope |
| 2025-W43 | 57.1% | 266,145 | one scope |
| 2025-W45 | 66.7% | 247,027 | one scope |
| 2025-W47 | 71.4% | 239,620 | one scope |
| 2025-W48 | 80.0% | 228,583 | one scope |
| 2025-W49 | 50.0% | 228,440 | one scope |
| 2025-W51 | 50.0% | 260,001 | one scope |
| 2025-W52 | 50.0% | 211,886 | one scope |
| 2026-W01 | 50.0% | 193,717 | one scope |
| 2026-W02 | 50.0% | 187,987 | one scope |
| 2026-W04 | 66.7% | 171,026 | one scope |
| 2026-W05 | 75.0% | 171,098 | one scope |
| 2026-W06 | 100.0% | 266,534 | one scope |
| 2026-W07 | 50.0% | 386,015 | one scope |
| 2026-W11 | 50.0% | 148,074 | one scope |
| 2026-W13 | 75.0% | 137,632 | one scope |
| 2026-W14 | 71.4% | 135,110 | one scope |
| 2026-W16 | 64.0% | 131,582 | one scope |
| 2026-W24 | 56.3% | 113,574 | one scope |
| 2026-W27 | 61.5% | 105,851 | one scope |
| 2026-W30 | 68.0% | 87,743 | one scope |
| 2026-W31 | 51.9% | 84,481 | one scope |
| 2026-W45 | 64.4% | 73,280 | one scope |
Source: data/out/episodes.csv
Episodes: data/out/episodes.csvWhat can this data not tell you?
- The key mix is a sample of 50,600 transactions, not a census. Weekly shares carry intervals; the full ADD_KEY ledger would cost roughly 20 TB to scan.
- USDC volume before 2023-10-16 is not in the census (bridged USDC.e logged no NEP-141 events until 2023-10-11), so those weeks are shown but not scored.
- Custodial apps sign with function-call keys for their users; function-call activity can measure a few large apps. Brief 2.3 shows how concentrated it is.
- The forward claim resolves on numbers the crank posts. Trust in them comes from make verify, which recomputes each from public data and compares hashes, not from the contract.
- Retention (signers who return within 4 to 8 weeks) and NEAR Intents volume are not computable within this build: the first is above the scan budget, the second needs a partner key the build did not have.
- Largest weekly share of keys nobody could read: 1.5%; never above 5 percent in a scored week.
Source: docs/METHOD.md sections 3 to 6; docs/RISK.md.