Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/theonetrade/backtest-ollama-crontab/llms.txt

Use this file to discover all available pages before exploring further.

This page documents the statistical research that motivated the LLM risk filter in backtest-ollama-crontab. A full export of a Telegram crypto trading-signals channel (April 2025 – April 2026) was parsed into machine-readable JSONL, and three separate datasets were analysed: signals.jsonl (raw signal posts), proof.jsonl (result posts), and entry.jsonl (matched pairs). The findings reveal a systematic, three-layer manipulation scheme that operates entirely within the laws of selective publishing and affiliate marketing — no market manipulation required.
This analysis is research and educational material. It is based on publicly observable signal posts from a specific channel during a specific time window. Nothing on this page constitutes financial advice. Past channel behavior does not guarantee future behavior. Do not use this analysis to make trading decisions.

​
Overview

416 total signals

Published across April 2025 – April 2026, covering LONG and SHORT directions on multiple crypto assets.

191 results published

Only 191 of 416 signals (46%) ever received a public outcome post — the rest disappeared silently.

65 SHORT signals parsed

All 65 have a risk/reward ratio below 1:1 to TP1. Not a single SHORT signal in the dataset has favorable R:R.

17 SHORT→LONG pairs

Documented cases where a SHORT is stopped out and a LONG immediately follows from the same price level.

​
Section 1 — The Mechanism: SHORT Signals With Guaranteed Liquidation

The channel publishes leveraged trading signals in a consistent format. Here is a verbatim example from January 2026: 
2026-01-06T10:16:16Z

SIGNAL #TRX/USDT

Open SHORT in the zone $0.2898 – $0.2930 with 25x leverage.

Take-profit targets:
  TP1: $0.2875
  TP2: $0.2864
  TP3: $0.2838
  TP4: $0.2809
  TP5: $0.2765

STOP-LOSS: $0.3027
The mathematics of this single signal are already damning:
ParameterValue
Average entry price$0.2914
Stop-loss$0.3027
Spot risk+3.88%
Target TP1$0.2875 (−1.34%)
Risk/Reward to TP10.345 : 1
Risk at 25x leverage96.9% of deposit
Price must fall 1.34% to reach TP1. If price rises 3.88% — the entire deposit is liquidated. The signal is placed at the bottom of a trading range where liquidity is thin and stop-hunt wicks are common. This is not risk management; it is a mathematically programmed wipeout.

​
It Is Not a One-Off — It Is a System

Across all 65 parsed SHORT signals in the full dataset:
MetricValue
Average spot risk4.24%
Average R:R to TP10.375 : 1
Average risk at 25x leverage106% of deposit
Minimum R:R observed0.333
Maximum R:R observed0.833
Not a single one of the 65 SHORT signals has R:R above 1:1. An average of 0.375 means the expected value is negative regardless of entry-timing accuracy.

​
Worst Signals by R:R

Date (UTC)TickerSpot riskR:RRisk at 25x
2026-01-28T14:18:01ZPOL3.88%0.33396.9%
2025-12-15T13:05:04ZSOL3.87%0.34096.8%
2026-03-04T08:40:41ZPOL3.84%0.34296.0%
2026-03-02T14:20:28ZPOL3.89%0.34297.1%
2026-01-09T14:10:57ZNEAR3.84%0.34496.0%
Every entry in this table wipes out more than 96% of a 25x-leveraged position on a stop-loss trigger.

​
Section 2 — Selective Result Reporting

After SHORT subscribers are stopped out, the channel only publishes outcome posts for the winning subset of trades. This creates survivorship bias in the public record.

​
Coverage by Direction

DirectionSignals sentResults publishedCoverage
LONG22611350.0%
SHORT1887037.2%
Only 3–4 out of every 10 SHORT signals ever receive a result post. For LONGs, every second one does. Within a strict 5-day post-signal window, the disparity is even clearer:
DirectionWith a published result (5-day window)
LONG124 / 226 (54.9%)
SHORT29 / 65 (44.6%)

​
Published SHORT Profit Is Anomalously High

If SHORT results are only published when they win, the average published profit for SHORTs should be inflated relative to LONGs, which are reported more consistently. This is exactly what the data shows:
MetricLONGSHORTDelta
Results published11370—
Mean profit51.85%55.09%+3.24%
Median profit47.92%53.51%+5.59%
Avg take-profit targets hit2.933.17+0.24
All targets hit20.4%35.7%+15.3%
The profit distribution makes the cherry-picking signature visible:
Profit rangeLONG resultsSHORT results
0–20%51
20–40%4324
40–60%3319
60–80%1822
80–100%61
100%+83
SHORT results are heavily concentrated in the 60–80% bucket — 22 cases vs. 18 for LONGs, despite LONGs having 1.6× more results published overall. Only the largest moves ever get a result post.
TickerLONG avg%SHORT avg%Delta
FARTCOIN14.3%116.6%+102.3%
ETH54.3%113.3%+59.0%
ADA45.1%57.3%+12.2%
SOL45.8%55.5%+9.7%
FARTCOIN has 4 published SHORT results with a mean profit of 116.6%, while its single LONG result clocks in at 14.3%. This does not mean shorts on FARTCOIN performed better — it means only the extreme outlier runs were ever published.Several tickers have LONG results published but zero published SHORT results: DOGE, IOTA, LTC, PUMP, RIVER, TAO, TRUMP, XAUT, ZEC.BTC has exactly 1 published SHORT result at 7.6% profit — the lowest of any ticker — against 9 LONG results averaging 43.4%.

​
Section 3 — The SHORT → LONG Pattern

Signal pair analysis identified a recurring structural pattern: a SHORT signal is published near range lows; after the stop-loss is triggered, a LONG signal on the same ticker appears with an entry zone placed close to the short’s stop-loss price. Seventeen such pairs were found and documented in proof.jsonl. Selection criteria: a LONG on the same ticker must appear within 7 days of the SHORT, and the midpoint of the long entry zone must be within 3% of the short’s stop-loss level.

​
All 17 Documented Pairs

SHORT date (UTC)TickerShort SLLong entrySL/entry gapHours laterRisk at 25x
2025-12-25T03:01ZNEAR$1.5601.572–1.572–1.5891.31%100.095.7%
2025-12-25T14:06ZSOL$127.00122.9–122.9–124.22.72%95.097.1%
2025-12-26T18:09ZTRX$0.28880.2811–0.2811–0.28422.13%62.996.2%
2025-12-29T06:51ZHYPE$26.3625.52–25.52–25.802.66%30.396.5%
2026-01-09T14:06ZTRX$0.30680.3027–0.3027–0.30610.78%142.196.0%
2026-01-12T13:08ZTRX$0.30990.3027–0.3027–0.30611.77%71.095.9%
2026-01-13T11:05ZTRX$0.31240.3027–0.3027–0.30612.56%49.196.8%
2026-02-19T22:59ZHYPE$30.3129.34–29.34–29.672.66%15.596.8%
2026-02-20T18:25ZPOL$0.11000.1107–0.1107–0.11191.18%96.795.6%
2026-02-24T20:01ZHYPE$28.3028.08–28.08–28.390.23%19.296.3%
2026-02-26T12:06ZPOL$0.11430.1148–0.1148–0.11611.01%22.895.4%
2026-02-27T12:08ZSOL$86.9488.15–88.15–89.131.96%75.995.8%
2026-03-19T16:05ZTRX$0.31200.3022–0.3022–0.30552.61%20.396.1%
2026-03-23T14:03ZHYPE$40.4839.11–39.11–39.542.85%28.596.5%
2026-03-24T12:34ZTRX$0.32290.3151–0.3151–0.31861.87%147.795.7%
2026-04-09T16:03ZPENGU$0.0068010.006859–0.006859–0.0069351.41%122.096.2%
2026-04-13T02:01ZHYPE$43.5643.00–43.00–44.600.55%34.095.9%

​
Three Most Telling Cases

1

Case 1 — HYPE, February 2026 (gap 0.23%, 19 hours)

2026-02-24T20:01:15Z — SHORT HYPE
SIGNAL #HYPE/USDT

Open SHORT in zone $27.1 – $27.4 with 25x leverage.

Targets:
  TP1: $26.88
  TP2: $26.77
  TP3: $26.54
  TP4: $26.26
  TP5: $25.85

STOP-LOSS: $28.30
R:R = 0.352 · Risk at 25x = 96.3% of deposit
2026-02-25T15:10:54Z — LONG HYPE (19 hours later)
SIGNAL #HYPE/USDT

Open LONG in zone $28.08 – $28.39 with 25x leverage.

Targets:
  TP1: $28.62
  TP2: $28.73
  TP3: $28.98
  TP4: $29.26
  TP5: $29.69

STOP-LOSS: $27.15
The long entry midpoint (28.24)is∗∗0.2328.24) is **0.23% away** from the short stop-loss (28.30). The long’s own stop-loss (27.15)issetjustbelowtheshort′soriginalentryzone(27.15) is set just below the short's original entry zone (27.1–$27.4). The channel is implicitly acknowledging that short subscribers are already out and is opening the opposite position from their liquidation level. The short result is never published. The long result is.
2

Case 2 — TRX, January 2026 (three consecutive shorts, one long)

Three SHORT signals on TRX are published between January 9–13, 2026, all targeting the same approximate stop level near $0.31. Then a single LONG:
Date (UTC)DirectionEntry zoneStop-lossResult published
2026-01-09T14:06ZSHORT0.2938–0.2938–0.2971$0.3068No
2026-01-12T13:08ZSHORT0.2968–0.2968–0.3001$0.3099No
2026-01-13T11:05ZSHORT0.2991–0.2991–0.3024$0.3124No
2026-01-15T12:10ZLONG0.3027–0.3027–0.3061$0.2927Yes ✅
Three sequential shorts are stopped out as price climbs through their stop levels. The long is opened exactly where price arrived after all three stop-losses were triggered. None of the short results are published. The long result is.
3

Case 3 — HYPE, April 2026 (gap 0.55%, 34 hours)

2026-04-13T02:01:31Z — SHORT HYPE
SIGNAL #HYPE/USDT

Open SHORT in zone $41.72 – $42.18 with 25x leverage.

Targets:
  TP1: $41.38
  TP2: $41.21
  TP3: $40.84
  TP4: $40.43
  TP5: $39.79

STOP-LOSS: $43.56
R:R = 0.354 · Risk at 25x = 95.9% of deposit
2026-04-14T12:02:04Z — LONG HYPE (34 hours later)
SIGNAL #HYPE/USDT

Signal risk: 6/10

Open LONG in zone $43.00 – $44.60 — 5% of deposit, 10x leverage.

Targets:
  TP1: $46
  TP2: $48
  TP3: $50
  TP4: $52
  TP5: $54

STOP-LOSS: $39
The long entry midpoint (43.80)is∗∗above∗∗theshortstop−loss(43.80) is **above** the short stop-loss (43.56) by 0.55% — the long is only published after short subscribers’ stops have certainly been hit. Note that by April 2026 the channel has started adding a Signal risk: N/10 label and reduced leverage to 10x for longs — the pattern is unchanged while the presentation has become more sophisticated.

​
Section 4 — Business Model

The scheme does not require the channel to move markets — its subscriber count is too small for that. It operates through three compounding mechanisms.

Survivorship Bias

Subscribers see only profitable trades in their feed. Losing shorts disappear without a result post. This creates the persistent illusion that the channel has a high hit rate, anchoring subscriber trust even as money is lost.

Exchange Referral Commission

Every subscriber liquidation generates a trading fee for the exchange. The channel operator receives a percentage of that fee as a referral partner. A referral link appears in every single signal post: Trade here | Join VIP channel.

VIP Funnel

A subscriber loses money on a short, sees that longs are working, concludes they “entered wrong,” and pays for VIP access to get “more precise” signals. The cycle repeats. Each churn event is a new conversion opportunity.

​
Summary of Findings

FindingData
Average R:R of SHORT signals0.375 — negative expected value at any win rate
Average risk per SHORT trade at 25x106% of deposit
SHORT signals without a result post44.6% vs 54.9% for LONGs (5-day window)
LONGs reported more consistently50.0% overall coverage vs 37.2% for SHORTs
Mean profit of published SHORTs55.09% — higher than LONGs (51.85%) despite worse underlying R:R
SHORTs in the 60–80% profit bucket22 cases vs 18 for LONGs in a sample half the size
SHORT → LONG pattern documented17 pairs, SL/entry gap ranging from 0.23% to 2.85%
This analysis directly motivated Rule 1 of the LLM risk filter (packages/core/src/logic/outline/risk.outline.ts): the sleeping-coin SHORT veto. When avgRangePct < 0.07%, a SHORT signal is automatically vetoed — thin candle range indicates an illiquid, low-volatility market where a SHORT placed at range lows is exactly the stop-hunt setup documented above. The January 2026 backtest confirmed that four of the six vetoed signals (TRX SHORT Jan 06, TRX SHORT Jan 12, and others) matched this profile precisely.

Build docs developers (and LLMs) love

Get started for freeTalk to us