AI Assistant

AI Assistant monthly check September 2026

2026-09-01 to 2026-09-21

decided signals counted once per event. Take-profit vs stop-loss on decided signals. Open/NR is not a win. Next-bar direction is not this metric.

Signal outcomes

Decided

2623

TP vs SL

21.81%

95% interval

20.27–23.43

Open/NR

527

Take-profit vs stop-loss on decided assistant suggestions. Next-bar direction is not this metric.

TP vs SL by mode
ModeTP vs SL95% intervalDecided
default20.2315.78–25.57257
only_ai24.7917.85–33.33117
only_social22.1920.45–24.032064
only_ta17.8412.99–23.99185
CoinTP vs SL95% intervalDecided
ADA25.010.18–49.516
BNB25.2221.45–29.4456
BTC19.0215.24–23.48347
DOGE43.7523.1–66.8216
ETH16.5213.37–20.24448
LTC25.010.18–49.516
PAXG0.00–27.7510
RENDER6.451.79–20.7231
SOL31.427.43–35.67484
TRX0.00–1.54245
XAUT11.111.99–43.59
XRP28.3324.49–32.52480
ZEC16.929.72–27.8265

Market terms used here

CryptoCrispy reports score cryptocurrency trading signals on USDT spot pairs — Bitcoin (BTC), Ethereum (ETH), and altcoins such as SOL, XRP, BNB, and RENDER. Published Telegram buy/sell alerts are counted separately from in-app AI Assistant suggestions.

  • Take-profit (TP1–TP5) — staged profit targets on a long or short. A target hit counts only after the signal is decided.
  • Stop-loss (SL) — risk exit. A stop is a loss, not a timeout.
  • Hit rate / TP vs SL — wins among decided signals. Open or no-result (NR) rows are not wins.
  • AI Assistant — in-app long/short suggestions. Modes: default mix, technical analysis, social sentiment, and AI/on-chain data.
  • Telegram Official Signals — published channel alerts, often a rolling last-40 window.
  • Core vs tactical — investor-book split: persistent hold versus a smaller rotating sleeve.
  • ROI — return after pairing BUY then SELL on the same symbol, fees subtracted, on trading pages.
  • Risk appetite — separate risk modes; the headline is sample-weighted, not the best bucket.

See how these numbers are counted and the crypto trading reports library.

Methodology

Metrics follow methodology 2026-09-21. Each prediction is counted once at the stated horizon. Training accuracy is shown only when a labelled holdout set exists. See the methodology page.