ALERTAUDIT · CASE FILE REPORT AA-2026-001 (PILOT SAMPLE) WINDOW 2026-04-20 → 2026-07-20 · 90 DAYS ISSUED 2026-07-21

Audit of “Binance Killers”

The public feed of Telegram channel Binance Killers (~319,000 subscribers) — every post in the window collected, classified, and scored against real exchange data. Every quote and signal below links to its original public post, so any line can be verified in one click.

Its two real public calls both won — and every advertised profit figure, from “+107%” to “almost 1800% a month”, cannot be checked against anything the channel has publicly posted.

Public signals
Pass

2 of 2 advance calls hit Target 1 and ran to Target 5 — while BTC fell 11.8%. Sample of two: 95% CI 34–100%.

Marketing claims
Unverifiable

36 profit/accuracy posts. 0 of 9 coin-specific claims preceded by any public advance call. “Zero losses” — untestable.

Feed quality
0.5% signal

2 actionable posts in 859. The rest: 355 commentaries, 36 profit claims, 466 promos & misc.

859
public posts in 90 days
2
unique advance signals
36
posts claiming profits or accuracy
0/9
claims checkable from the public feed
Exhibit A · The feed

859 posts. Two of them were signals.

859 public posts in 90 days — classified every post in the window, one bar per category
Advance signals — 2 unique (+2 reposts)4Profit claims & accuracy ads36Market commentary (untradeable)355Promos & misc464

Every post in the window, classified. An advance signal = instrument + direction + entry + posted before the move. Everything else cannot be traded and cannot be scored.

Exhibit B · The two real signals

Credit where due: both public calls worked.

Signal 1 — ZIG longposted 2026-05-04 13:12 UTC · Bybit spot, 1h closes
0.031810.040340.048860.05739May 04May 30Jun 25Jul 20 TP1 ✓TP3 ✓TP4 ✓TP5 ✓TP6SL — never touched ENTRY 0.02812
Posted2026-05-04 13:12 UTC
Entry (“below 0.03”)filled 0.02812
Target 1 (+13.8%)hit 05-10, stop untouched
Best target5 of 9 · mark +50.4%
Signal 2 — BULLA longposted 2026-07-17 12:25 UTC · Binance futures, 1h closes
0.0073090.00870.010090.01148Jul 17Jul 18Jul 19Jul 20 TP1 ✓TP3 ✓TP4 ✓TP5 ✓TP6SL — never touched ENTRY 0.008605
Posted2026-07-17 12:25 UTC
Entry (“below 0.0089”)filled 0.008605
Target 1 (+5.75%)hit in 5 hours
Best target5 of 8 · mark +34.0% (3 days old — not final)

The $10K test: $10,978 exiting each at Target 1, or $14,220 riding the ladder to window end — against $8,820 just holding BTC. Both readings beat the market. Both rest on exactly two trades — a 2/2 hit rate carries a 95% confidence interval of 34% to 100%. Two wins is consistent with skill and with luck.

Exhibit C · Claims vs. public record

36 profit claims. We looked for the trades.

For every coin-specific profit claim in the window, we searched the channel's own public history for the advance call it refers to:

DateCoinClaim (verbatim)Public advance call
2026-04-24$CHZ“+107% profit”NONE FOUND
2026-04-28$XTZ“88% profit”NONE FOUND
2026-05-09$ICP“1225% + total profit for May”NONE FOUND
2026-05-29$WLD“🔥 44.2% Profit (5x) 🔥”NONE FOUND
2026-07-06$SKY“We called this bounce early and we're currently sitting on a 77% profit (5x) fro”NONE FOUND
2026-07-07$SOL“We're currently sitting on 108% profit (5x) from this call, and price has alread”NONE FOUND
2026-07-08$UNI“We're sitting on 227% profit (5x) from this call, with just the final target lef”NONE FOUND
2026-07-15$LINK“We're sitting on 24% profit with the first target already touched, and price jus”NONE FOUND
2026-07-19$LINK“✔ $LINK up 50% from our entry ✔”NONE FOUND
Finding: every claim refers to trades whose entry calls never appeared in the public feed — attributed to a private VIP service. From the public record, none of the 9 coin-specific claims, nor the month totals of “770%”, “1035%”, “1225%” and “almost 1800%”, can be substantiated or refuted. A track record that cannot be checked is not a track record.
Exhibit D · The language

Claims built to be unfalsifiable.

“Zero losses. Not a single one. Every signal. Every setup. Still green.” Post #19744 · 2026-04-21
“May signals alone. Almost 1800% profit. Not counting March. Not counting April.” Post #20124 · 2026-05-29
“✅ 92%+ Historical Accuracy” Recurring advertisement · 5 appearances in window · e.g. post #19846

Structural flags on the signals themselves: entry zones with no lower bound (“below 0.03”) and target ladders ending in “OPEN” — which makes any per-signal “total profit” figure non-computable even in principle.

Right of reply: Binance Killers — or anyone — may dispute any line of this report at audit@alertaudit.io. Every factual claim above links to its source; demonstrated factual errors are corrected within 48 hours with a visible changelog. Nothing here alleges wrongdoing — it documents what the public record does and does not support.
Reading this report

What 90 days can and cannot prove.

The data supports

  • The public feed is 0.5% actionable signal by volume.
  • Both real public calls won, in a falling market.
  • Advertised accuracy figures have no public evidence behind them.
  • 4% of all posts are performance marketing.

The data cannot tell you

  • Whether the private VIP record is real — nobody outside can check it.
  • Whether 2/2 is skill or luck (CI: 34–100%).
  • How the channel performs in other market regimes.

If you are paying for the VIP tier on the strength of the advertised numbers, you are paying for figures the public record cannot confirm. That is this audit's verdict.

Methodology — identical in every report
  1. Source material. Every public post in the window collected and counted. Nothing excluded.
  2. Scoring rule. A post is a scoreable advance signal if it names an instrument, a direction, an entry level, and was posted before the move it describes. Exits follow the signal's own targets/stop.
  3. Price data. 1-hour exchange OHLC from the post's timestamp — no look-ahead. Entry rule: first 1h open after the post if inside the stated zone, else first touch of the zone edge.
  4. Conservative tie-break. A candle touching both target and stop counts the stop first.
  5. The $10K test. Equal split across all scoreable signals, no leverage. Benchmark: BTC buy-and-hold, same window.
  6. Uncertainty. Hit rates carry Wilson 95% confidence intervals. Small samples are stated, not hidden.
  7. Independence. No relationship with any signal provider. Educational statistical analysis, not financial advice.
Appendix A — the scoreable signals, verbatim

Post #19864 · 2026-05-04 13:12 UTC — “GEM SIGNAL: $ZIG — Exchanges: ByBit, Bitget, MEXC, Gate … ENTRY: Below 0.03 · TARGETS: 0.032 – 0.034 – 0.037 – 0.041 – 0.048 – 0.064 – 0.088 – 0.104 – 0.128 – OPEN · SL: Below 0.024”

Post #20580 · 2026-07-17 12:25 UTC (reposted #20597, #20600) — “SIGNAL ID #2183 · COIN: $BULLA/USDT (2-5x) · Direction: LONG (Binance Futures) · ENTRY: Below 0.0089 · TARGETS: 0.0091 – 0.0093 – 0.0096 – 0.01 – 0.011 – 0.0125 – 0.0145 – 0.017 – OPEN · STOP LOSS: Below 0.0061 · ‘Same pattern as other coins that pumped 10-20x’”

A customer report includes the full 859-post census with per-post classification.

Appendix B — data sources & parameters
  • Bybit spot ZIGUSDT 1h OHLC (1,858 candles); Binance futures BULLAUSDT 1h OHLC (83 candles).
  • Cross-check: Gate.io ZIG_USDT daily close vs Bybit, 2026-05-10 — deviation 0.27%, within tolerance.
  • Benchmark: Bybit BTCUSDT daily, 2026-04-20 open 73,803 → 2026-07-20 close 65,125 (−11.8%).
  • Confidence intervals: Wilson score, 95%.