CryptoEdgeExplorerslab notes by Ξmars
Field notes from the trading internet

Most trading claims don't survive fees. This ledger shows which ones do.

Crypto, stocks and ETFs: signals, bots, copy trading, arbitrage, "income" funds, "trending" tokens. Each claim gets real data, real costs and rules written down before the results come in. Wins and losses are both published. No signals, no promises.

The ledger

20 claims tested · 16 failed
  1. 01Copy the top Binance futures leadstested 2026-10-22 · full write-up~1 in 5 of 6,753 leads had copiers in profit over 30 daysFAIL
  2. 02Buy tokens when a Telegram channel says "trending"tested 2026-10-21 · full write-up300 tokens: 1h hold averaged -23% after costs; 54% were down 90% within 24hFAIL
  3. 03Buy new tokens right at launchtested 2026-10-2192% of new BSC pairs and 84% of new Solana tokens never traded after the postFAIL
  4. 04Spot triangular arbitrage on Binancetested 2026-10-220 profitable cycles out of ~4,700 per snapshot with real feesFAIL
  5. 05Arbitrage between exchanges (9 venues incl. a DEX)tested 2026-10-22Price gaps 0.01% to 0.16%, smaller than two taker feesFAIL
  6. 06Funding-rate arbitrage like exchange "arbitrage bots"tested 2026-10-21Pre-registered: out-of-sample +0.37% a week, lower bound -0.10%FAIL
  7. 07Grid bots beat holding BTCtested 2026-10-213 years: $150 to $173 per $100 vs $285 for holdingFAIL
  8. 08Paid Telegram futures signal channelstested 2026-10-174 channels, 12 months of calls replayed: no edge after feesFAIL
  9. 09Intraday indicators and scalping on BTCtested 2026-10-20MACD, RSI, EMA, VWAP rules on 1m to 1h: all lose after feesFAIL
  10. 10Buy the favourite late in 5-minute BTC prediction marketstested 2026-10-21Pre-registered, 152 trades: 88% win rate, still -5.1% per USDTFAIL
  11. 11High-yield "income" ETFs pay you moretested 2026-10-21Every one with 2 years of data lagged its own index on total returnFAIL
  12. 12Tokenized stocks pop on listing daytested 2026-10-2195 listings: -1.5% after 24h net; they just track the stockFAIL
  13. 13Dated gold and "dollar reset" forecasts in promo newsletterstested 2026-10-25 · full write-up20 due: 9 came true vs about 8 by chance; 0 of 4 dated dollar resets happenedFAIL
  14. 14High-yield stocks that pay a dividend every monthtested 2026-10-26 · full write-up10 years: 8.1% a year vs 12.4% for a plain dividend ETF; 7 of 13 cut their dividendFAIL
  15. 15Buy a stock right after it doubles in a daytested 2026-10-26 · full write-up213 cases in 5 years: 2 in 3 lost; typical trade -13.6% after 5 daysFAIL
  16. 16An AI-built trading bot beats holding cryptotested 2026-10-26 · full write-upTuned years 58% a year; unseen 2023-2026 21% vs 54% holding BTCFAIL
  17. 17A daily 200-day trend filter on BTCtested 2026-10-20 · full write-upSimilar return to holding since 2020, worst drawdown 56% vs 77%HOLDS
  18. 18Copy only long-lived, low-drawdown leadstested 2026-10-21Paper test running, review on 2026-11-18PENDING
  19. 19Football favourites are underpriced on prediction marketstested 2026-10-22Pre-registered paper test, needs 150 matchesPENDING
  20. 20Solana finality is 150 ms after Alpenglowtested 2026-10-22 · full write-upInterim, 86 h: median 217 ms, 95% within 481 ms; verdict on 2026-10-29PENDING

Want a claim tested? Send it to @heyemars on X.