Technical profile and market coverage
Gadget
uses Bybit
on 15m timeframe.
Sample capital: $1000.
Recommended minimum: $200.
Technical notes
The current Gadget version runs the GadgetRLSynth strategy with the custom GadgetRLPro FreqAI/RL model. The model uses PPO with a Base3ActionRLEnv action space for neutral/long/short decisions. Features come from 15m/1h/4h data, including RSI, ADX, MFI, CCI, EMA, ROC, ATR%, volume quality, log returns, and range efficiency. Entries require do_predict, an RL action, and quality filters around EMA, ADX, ATR, volume, range efficiency, and breakouts. Exits use RL actions, ATR-based stop logic, profit lock, panic cuts, and time exits.
Covered pairs (2)
BTC/USDTSOL/USDT
Full bot details
Gadget acts like a risk-first controller: it checks trend, volume, volatility, and breakout quality before the RL model confirms direction. In unclear conditions it limits entries, while exits combine RL actions, dynamic stop behavior, profit locking, and time-based protection.
Risk-first reinforcement-learning bot for Bybit futures, controlling long/short decisions with PPO and market-quality filters.
Supported leverage range: 1x to 4x.
The model warm-up uses 320 initial candles.
Features
GadgetRLSynth strategy with the custom GadgetRLPro RL model.PPO model with a three-action space to reduce unstable behavior.Multi-layer entry filtering with EMA, ADX, ATR, volume, and breakout checks.Risk-first reward design with drawdown penalties and circuit-breaker behavior.Hybrid exits using RL actions, ATR stop logic, profit locking, and time exits.