SentinelAudit a channel

The Dead Edges Ledger

In November 2026 we tested 101 trading rules across 40 ideas that traders are told work: whale alerts, ETF flows, funding arbitrage, Telegram calls, on-chain indicators and more. 91 failed. Every one is listed here, including the ones that hurt.

Each rule was written into a lock file before any result existed, and the first 16 characters of that file's sha256 fingerprint are shown beside it. The score is the t-statistic on data the rule had never seen, after exchange fees and funding. A rule needed 3.0 or more to pass, plus a set of checks: still positive with higher costs, still positive without its five best weeks, and better than random entries at the same moments.

The two rules marked passed are the same trade measured two ways. We then tested that trade again on other exchanges, and it failed there. Read what happened. So of 101 rules, none is a proven edge today.

91
failed
7
too little data to call
3
passed or came close, all one idea, and it failed its second test
IdeaRule testedtVerdictLocked
Academic trend factorH1 CTREND-style trend factor (JFQA 2025): weekly long top/short bottom quintile of forecast from 7 price-MA + 6 volume-MA rank signals, trailing-52wk CS-regression weights, hold 7d, week-clustered-0.04failed2026-11-11
f9482729d1b2fca4
Binance monitoring tagH1 short Binance Monitoring-Tag additions vs top-50 alt basket, 7d, entry 1-5h after announcement (12 TEST week-clusters)-1.20failed2026-11-11
47808e93de5c0fad
Binance monitoring tagH2 short Binance Monitoring-Tag additions vs top-50 alt basket, 30d (11 TEST week-clusters)-1.06failed2026-11-11
98623debcec34ce8
Binance monitoring tagH3 long Monitoring/Seed tag removals vs top-50 alt basket, 7d (4 TEST week-clusters)-0.14failed2026-11-11
3ee0a9aa72dd5e01
Bitcoin ETF flowsH1 follow US BTC ETF net-flow surprise |z20|>=1, BTCUSDT 24h from 12:00 UTC next day-0.26failed2026-11-11
6925b2ecb9e91754
Bitcoin ETF flowsH2 same signal, BTCUSDT 72h, week-clustered+0.16failed2026-11-11
7fafc53893b11a04
Bitcoin ETF flowsH3 same signal, top-20 turnover alts 72h equal-weight, week-clustered+1.26failed2026-11-11
33d98b9e29287cc7
Bitcoin on-chain indicatorsH1 hash ribbon weekly BTC timing: long if SMA30>=SMA60 hash rate else short, 7d, Monday entry, 2d pub lag+0.97failed2026-11-11
bee2cd01f9d4c53b
Bitcoin on-chain indicatorsH2 Puell (7d/365d miner rev USD) weekly BTC timing: long if <1 else short, 7d-1.14failed2026-11-11
0db943f30ad14364
Bitcoin on-chain indicatorsH3 difficulty-adjustment sign: long BTC after negative adj, short after positive, hold to next adj (<=14d); TRAIN t+0.85 opposite sign-2.23failed2026-11-11
f512c092e4df3a13
CME gaps and options expiryCME gap fade: |gap|>=1% at Sun reopen, TP at Fri CME close, time exit next Fri close (BTCUSDT 1h)+0.18failed2026-11-10
03ee8f4339dbfff2
CME gaps and options expiryDeribit monthly expiry: fade 7d pre-expiry BTC move, Fri 08:00 -> Mon 08:00 UTC-0.10failed2026-11-10
9ef8e03d2827a59d
CME gaps and options expiryWeekend long BTC Sat 00:00 -> Mon 00:00 UTC+0.17failed2026-11-10
3b90bed77aa88b5a
Coinbase premiumH1 Coinbase BTC/ETH 24h premium z(30d) >=+1.5 long / <=-1.5 short perp, hold 24h, week-clustered+0.24failed2026-11-11
5190a887a3050f18
Coinbase premiumH2 daily XS: long 5 highest / short 5 lowest 3d Coinbase-vs-Binance spot premium alts, hold 24h, day-clustered+0.44failed2026-11-11
b35e7444529f6f49
Copy-trading leaderboardsCEX leaderboards: only OKX has an official public API, giving current leaders only (survivor-biased) with 12w weekly PnL / 3m positions, so <30 clean clusters; no prereg—not testable2026-11-11
none
Corporate token treasuriesH1 short token perp/long BTC 14d from 00:00 UTC after first SEC 8-K announcing a <token> treasury (DAT), week-clustered (21 TEST clusters; TRAIN t-1.87 opposite)-0.31failed2026-11-11
fa8714579a1a3a36
Developer activityH1 GitHub main-repo commit growth (91d vs prior 91d) top-vs-bottom tercile, monthly market-neutral, ~61 coins+0.16failed2026-11-11
ea7e3095cfa81eda
Developer activityH2 GitHub main-repo distinct-author growth (91d vs prior 91d) top-vs-bottom tercile, monthly market-neutral+0.50failed2026-11-11
1fe357b1ea058631
Do good Telegram channels stay goodH1 follow top-3 TG providers by trailing-30d basket-hedged mean, next week, 24h hold (11 TEST weeks; fails drop-best-5 and 1-bar-late)+1.41inconclusive2026-11-11
9e3629db654f631a
Do good Telegram channels stay goodH3 weekly Spearman IC trailing-30d provider score vs next-week adj return (mean IC +0.11 TEST, +0.03 all 24 wks; inside null band)+1.26inconclusive2026-11-11
a0e0fd49cde4e656
Do good Telegram channels stay goodH2 fade bottom-3 TG providers by trailing-30d basket-hedged mean, next week, 24h hold-0.04failed2026-11-11
748d4aaa7e471c50
Fear and Greed IndexH1 FNG<=25 -> long BTCUSDT 14d, entry D 04:00Z, week-clustered-0.16failed2026-11-11
84030a7564ee5328
Fear and Greed IndexH2 FNG>=75 -> short BTCUSDT 14d, entry D 04:00Z, week-clustered-0.90failed2026-11-11
f48d0a048457fdb0
Fear and Greed IndexH3 FNG(D)-FNG(D-7)<=-15 -> long top-20 liquid alts 7d, week-clustered+0.88failed2026-11-11
55c3398b27e94dcf
Fed, CPI and jobs daysH3 pre-FOMC drift: long BTC+ETH T-24h -> T, FOMC only; 25 TEST clusters <30 (point estimate zero)-0.04inconclusive2026-11-11
becbc3360d3fd9b3
Fed, CPI and jobs daysH1 long BTC+ETH perps T-24h -> T before every FOMC/CPI/NFP release (30m bars, event-clustered)-0.14failed2026-11-11
50354896b33d3cef
Fed, CPI and jobs daysH2 trade sign of BTC first-1h move after FOMC/CPI/NFP, BTC+ETH T+1h -> T+24h+1.43failed2026-11-11
dcdea55ab879f79a
Free-token listings, other exchangesX1 replication on Bybit Launchpool/OKX Jumpstart: short farmed token vs BTC 14d, n23 G20 mean +8.97% median +31.3%; pre-2025 t-0.03, 2025+ t+5.77 (n8); null 95th +2.31 -> 2025 regime effect+0.62failed2026-11-12
7846a144dedd1eaa
Free-token program listingsH1 short Binance Launchpool/HODLer/Megadrop token perp vs long BTC equal notional at listing-bar close, hold 14d, 0.214% RT + funding, week-clustered+4.15passed2026-11-11
6243901b1b367573
Free-token program listingsH2 naked short farmed-token perp at listing-bar close, hold 30d, 0.107% RT + funding (same 59 TEST events as H1)+5.04passed2026-11-11
b1ca98ce3087e804
Free-token program listingsH3 pre-market perp (>=24h before spot) short vs BTC from spot-listing bar, hold 7d+0.24failed2026-11-11
8aebad71e051f252
Funding rate arbitrageH1 Binance-Bybit funding arb: |72h funding diff|>=0.30% -> short high venue/long low, 7d, 4x0.055% (no TRAIN: Bybit daily klines only from 2024-02-10)+0.73failed2026-11-11
5733ef70625b1a3d
Funding rate arbitrageH2 Binance-Hyperliquid funding arb, same rule (TRAIN t+11.59 2023-24, decayed to ~0 in 2025-26)+0.92failed2026-11-11
435e20c10e09a4de
Funding rate arbitrageH3 long top-3/short bottom-3 by HL-minus-Binance 24h funding, Binance perps 24h hold+0.34failed2026-11-11
57822c851ca5d85b
Futures basisH1 BTC/ETH perp contrarian on COIN-M quarterly ann. basis vs trailing-365d p10/p90, hold 7d-1.57failed2026-11-11
03dcc219c90ccce5
Futures basisH2 BTC/ETH perp on (quarterly ann. basis - 7d perp funding) p10 short / p90 long, hold 7d+0.22failed2026-11-11
deefd2780f5bf1f9
Futures basisH3 short BTC/ETH perp after 7d basis change < trailing p10 and <0, hold 7d-0.03failed2026-11-11
c4f4adf85aeadd76
Hyperliquid vaultsH1 HL vault persistence: every 14d deposit top-5 vaults by trailing 30d return (AV>=$50k), hold 14d, net of leader commission-2.74failed2026-11-11
caee3edc16a21763
Ideas from forumsH1 long Binance perp 72h after funding interval shortened on negative cap (prev rate<=-0.25%), naked (median -4.3%, drop-best5 t-1.53; lottery tail)+1.33failed2026-11-11
2d6a9bca8d398690
Ideas from forumsH2 same events, long coin/short BTC 72h (TRAIN t+2.19; drop-best5 t-1.32; median -4.1%)+1.46failed2026-11-11
235ff056f3412df2
Ideas from XH1 long Binance perp at first hourly-funding switch w/ negative funding (floor hit), hold 72h, week-clustered (median -4.4%, drop-best5 t-1.59)+1.14failed2026-11-11
a8f7296243491bd6
Ideas from XH2 same events, long perp 24h, day-clustered+0.37failed2026-11-11
fb5e0995c3d89f6e
Ideas from XH3 same events, long perp / short BTC 72h, week-clustered (inside null 5-95 band -0.45..2.06)+1.27failed2026-11-11
fdb9bf17861b2cc9
Kimchi premiumH1 SHORT coin when Upbit rel premium (vs BTC premium) >= 30d median +5pp, hold 72h+0.00failed2026-11-11
3653c9df120f24bd
Kimchi premiumH2 BTC kimchi premium dev vs 30d median >=+2pp SHORT / <=-2pp LONG BTCUSDT 72h-0.54failed2026-11-11
f386a955cf8e0f63
Kimchi premiumH3 SHORT coin when Upbit/Binance volume share >= 5x 30d median, hold 72h-0.38failed2026-11-11
9eab713a64faf737
Korean exchange listingsH1 long Binance perp ~1-2min after Upbit KRW listing notice (close bar 1), exit +60min-0.06failed2026-11-11
c29950780ee19f06
Korean exchange listingsH2 long Binance perp ~5min after Upbit KRW listing notice (close bar 4), exit +4h+0.01failed2026-11-11
3e226cf24e00d108
Korean exchange listingsH3 short Binance perp 1h after Upbit KRW listing notice, exit +24h (median +5.7% but left tail to -147%)+0.03failed2026-11-11
6343631bd3bc2617
Nasdaq, VIX and bond yieldsH1 Nasdaq-100 vs BTC same-window beta-gap |z|>1 -> trade BTC toward catch-up, enter 00:00 UTC, hold 24h-0.58failed2026-11-11
77057e66f9a4772c
Nasdaq, VIX and bond yieldsH2 VIX close +10% d/d -> long BTC 00:00 UTC, hold 72h-0.11failed2026-11-11
ab225dfea717fff6
Nasdaq, VIX and bond yieldsH3 US10y |dy|>=7bp -> BTC opposite to yields, enter 00:00 UTC, hold 24h-0.20failed2026-11-11
b3d6bea710f5560c
New spot listingsH2 same events, short perp hold 7d (TRAIN mean -1.55% opposite sign -> not CONFIRMED; +0.2% cost t+2.94; drop-best5 t+1.85)+3.02candidate2026-11-11
1ea4853568c9ca1e
New spot listingsH1 short Binance perp 1-5h after Binance/Coinbase spot listing start of an existing perp, hold 24h+0.52failed2026-11-11
6f83b63f1ad998b2
News tone and volumeH1 GDELT bitcoin tone z<=-1.5 & volume-share z>=+0.5 (60d) -> long BTC 04:00 UTC next day, hold 72h-0.44failed2026-11-11
1fc7035dee5d8513
News tone and volumeH3 GDELT bitcoin tone 7d mean vs prior 28d: above -> long BTC 7d, below -> short, weekly Monday 04:00 UTC+0.91failed2026-11-11
6bcaa9dee818774d
News tone and volumeH2 GDELT per-coin news-volume-share spike (z>=3 vs 30d, count>=20) -> short coin / long BTC 04:00 UTC next day, hold 72h, 47 alts-0.50failed2026-11-11
cb339b023cd218cb
Options volatilityH1 VRP pct>=0.8 long / <=0.2 short BTC,ETH perps 7d weekly+1.08failed2026-11-10
d9c599bf6a22108d
Options volatilityH2 DVOL 1d jump >=p95 -> long same perp 72h+0.92failed2026-11-10
ee95d794048a3286
Options volatilityH3 BTC VRP regime -> top20 alt basket long/short 7d-0.39failed2026-11-10
6cf04ce316596ecc
Pairs tradingH1 Engle-Granger pairs (GitHub stat-arb repos): top-30 liquid perps, 90d formation, 5 most-cointegrated pairs, trade |z|>=2 -> 0 (stop 4) over next 30d, daily closes, week-clustered-0.12failed2026-11-11
98b1359780f31d01
Pairs tradingH2 distance-method pairs (Gatev): 5 min-SSD pairs of top-30 perps, 90d formation, |z|>=2 -> 0 over next 30d (TRAIN t+1.61 decayed)-0.81failed2026-11-11
a9c94812beb6bb80
Prediction marketsH1 Kalshi daily BTC ladder longshot: 6h pre-settle take the expensive side of strikes priced 0.02-0.15, taker fee, week-clustered (TRAIN t-0.48 opposite; +$0.01/contract stress t+0.46; not tradable from AU)+3.33failed2026-11-12
faa0925b9d322123
Prediction marketsH2 Kalshi vs DVOL-lognormal: 6h pre-settle take model side when |P-mid|>0.10, taker fee, week-clustered (TRAIN t-3.00; model loses to Kalshi)-1.93failed2026-11-12
0b2642bfadc5158c
Prediction marketsH3 Kalshi-implied median vs Binance BTC perp: z>TRAIN p90 long / z<p10 short, 4h hold, 0.107% RT + funding, day-clustered (TRAIN t-1.42)-1.80failed2026-11-12
af1d6cf03f215a5d
Pump groupsH2 short low-turnover perp up >=15%/24h at TG LONG call, hold 72h (only 5 events in 7 months; no testable sample)—inconclusive2026-11-11
b8c8e8089e8385da
Pump groupsH1 short low-turnover perp after 4h bar >=+20% on >=10x avg turnover, hold 72h (median +9.7% but left tail to -178%; drop-best-5 t=0.25)+1.37failed2026-11-11
6484fdacfee49020
Sector rotationH1 sector-neutral reversal: long bottom/short top 20% of 3d return minus leave-one-out sector mean, hold 3d, 8 hand-made sectors-0.81failed2026-11-11
1acd08f75fdfbf87
Sector rotationH2 sector momentum: long top-2/short bottom-2 sector EW baskets by 14d return, hold 7d weekly+0.28failed2026-11-11
306e647711521601
Sector rotationH3 sector-neutral reversal 7d residual, hold 7d weekly-0.38failed2026-11-11
442efa94bd0106ea
Speed-following Telegram callsH1 follow TG call: enter call direction at first 1m open >= capture+60s, exit +15m, 0.157% RT (Binance UM 1m)-7.74failed2026-11-11
2514746669a94b2d
Speed-following Telegram callsH2 fade TG call: enter opposite at message minute +30m, exit +240m, 0.157% RT + funding-5.27failed2026-11-11
8ed487f679c93bf4
Spot versus futures volumeH1 within top-third 3d rallies, long top / short bottom quintile of 3d-vs-30d spot-share change, hold 3d, every 3rd day-1.64failed2026-11-11
ee9ec0dad74086e9
Spot versus futures volumeH2 long top / short bottom decile of spot taker-buy share 7d minus 60d, hold 7d, weekly+0.92failed2026-11-11
510b08b4bec7d9eb
Spot versus futures volumeH3 within rallies, long spot-led (dS>0, Bybit dOI<=median) / short leverage-led (dS<0, dOI>median), hold 3d-1.86failed2026-11-11
6f6fa4c6bde31ca2
Telegram calls after a run-upH1 fade TG LONG calls w/ excess pre72 >= TRAIN P80: short coin/long BTC 72h+0.55failed2026-11-10
df9cde882a37a52b
Telegram calls after a run-upH2 fade TG SHORT calls w/ excess pre72 <= TRAIN P20: long coin/short BTC 72h-0.69failed2026-11-10
80a7c61d3203cfb7
Telegram calls after a run-upH3 fade TG LONG calls w/ excess pre24 >= TRAIN P80: short coin/long BTC 72h+0.30failed2026-11-10
85b813bb6f23467c
Telegram copy networksH1 ride TG network leader call (+60s) until first linked-follower re-post +30m (else 6h), Binance 1m, 0.107% RT + funding, day-clustered+0.27failed2026-11-11
ae76cfb590077cca
Telegram copy networksH2 fade first linked-follower re-post (+60s) for 4h, Binance 1m, 0.107% RT + funding, day-clustered-1.50failed2026-11-11
d31c77a6bbcb179a
Telegram mention countsDaily at F: SHORT A_72h>=2, LONG A_72h==0, hold 72h. Cluster by WEEK (holds overlap). Stated upfront: ~11 TEST+1.23inconclusive2026-11-10
e97fb1f606fe1145
Telegram mention countsDaily at F: SHORT equal-weight all universe coins with A_72h>=2; LONG equal-weight all universe coins with A_7+1.29failed2026-11-10
eba2b2d3d9447d41
Telegram mention countsDaily at F: LONG equal-weight universe coins with A_24h>=1 and zero mentions by fixed-set channels in [F-8d, F-2.18failed2026-11-10
e2fadb6002348ebf
Tether mintsH1 long BTC perp 72h after USDT Issue (ETH+Tron) >=$1B in trailing 24h, non-overlapping, week-clustered (TRAIN t-1.04 opposite; +0.2% cost t1.73; drop-5 t1.11; null 95th 1.69)+2.27failed2026-11-11
e9aa93cf008ebd27
Tether mintsH2 weekly USDT net issuance vs trailing 52w median -> long/short BTC 7d from Mon 04:00 UTC (TRAIN t-0.85 opposite)+1.37failed2026-11-11
ff9d675e8170440e
Tether mintsH3 long top-20 turnover alt basket 72h after USDT Issue >=$1B/24h (TRAIN t-0.29 opposite; drop-5 t0.16)+1.32failed2026-11-11
1e292fb47b0da538
Time of day and monthH1 long BTCUSDT perp 21:00->23:00 UTC daily (Quantpedia hours), day-clustered-1.79failed2026-11-11
d797c9cdd2f0cfbf
Time of day and monthH2 short top-30 alts in the 1h before funding settlement when prior rate >= 0.03% (gross t +1.23, fees kill it)-1.18failed2026-11-11
550d4d8664c3d529
Time of day and monthH3 turn-of-month: long BTC+ETH last day 00:00 -> day 4 00:00 UTC, 37 TEST months-1.08failed2026-11-11
c17ec05d5761be22
Token unlocksH1 short token/long alt basket from TGE(first-kline)+365d -7d to +7d (anniversary proxy for 12m cliff)+1.44failed2026-11-11
39551d4cef008153
Token unlocksH2 short token/long alt basket from TGE(first-kline)+365d -30d to +7d-0.25failed2026-11-11
143624e2cc561833
Token unlocksH3 short token/long alt basket from TGE(first-kline)+182d -7d to +7d-0.50failed2026-11-11
424f0132b0cec093
Top-trader positioningH1 L/S 10/10 by day-mean ln(top-trader position LSR / all-account LSR), hold 72h, 3-day rebal, top-100 Binance perps, split 2024-03-01+1.44failed2026-11-11
4f17eea1b026c92c
Top-trader positioningH2 L/S 10/10 by day-mean ln(taker buy/sell vol ratio) continuation, hold 72h (TRAIN t+3.39 did not hold)-0.01failed2026-11-11
5a9954d2ad4c8b4d
Top-trader positioningH3 long coins with top-trader LSR>1 and all-account LSR<1, minus EW universe, hold 72h+1.02failed2026-11-11
abe709f94e9017d4
Whale tradesH3 whale_hunter feed alerts follow, coin minus BTC 24h (8 day clusters <30)-2.65inconclusive2026-11-11
38669c0b1a9707db
Whale tradesH1 follow net whale taker flow (aggTrade>=$250k, |NET|>=$1M & >=3% of hour vol) 4h outright, 6 alts-2.67failed2026-11-11
d89990f5b99a5717
Whale tradesH2 same events, coin minus BTC 24h pair, week-clustered-0.28failed2026-11-11
ca79a23ed39d357d
Wikipedia page viewsH1 weekly XSEC: short top-quintile Wikipedia attention shock (7d vs prior 30d median), long bottom, 7d hold, 61 mapped coins-1.17failed2026-11-11
414b2e456f2ea515
Wikipedia page viewsH2 short coin/long BTC 7d after 1-day Wikipedia pageview spike >=5x 30d median (33 TEST week-clusters)-0.17failed2026-11-11
390916830852eed5

Generated 2026-11-12. Rules are shown as written in the lock files, in research shorthand: “perp” is a perpetual future, “RT” a round trip, “TRAIN” and “TEST” the seen and unseen halves of the data. Research on past data, not financial advice.

The full report

The whole ledger written up as one document, about 8,000 words: for each of the 40 topics, why the idea should work, where the data came from, the exact rules, the result and what killed it. Launch price US$3. You get the report on the next page straight after paying; it reads on a phone and prints to PDF.

Buy the report, US$3

A description of how rules behaved on past data. General information, not financial advice.

Have a rule of your own?

We will test it the same way: locked first, fees charged, compared against random entries, and reported whichever way it comes out.

Get a rule testedRead the findingsAudit a signal channel