Piggy Trader
An autonomous trading agent that learns from the full spectrum of market patterns and adapts its strategy to the market it's actually in — disciplined, compounded patiently, opened only when the maths agree.
01Overview
Piggy is an autonomous trading agent that learns from price, volume, volatility regimes, macro context and sentiment, and adapts its strategy to the market it's actually in. It's built like a piggy bank: disciplined deposits, compounded patiently, opened only when the maths agree.
visit piggytrader.cc ↗02The problem it solves
Most "AI trading" tools train on one signal — usually news sentiment — and hope it generalises. They also tend to be rule-based bots that commit to a single playbook and break the moment the market regime changes. Piggy is the opposite on both counts: it treats sentiment as one input among many, and it learns which kind of market favours which kind of strategy rather than betting everything on one.
Result strategies that have to earn their place out of sample, sized so a single bad trade can't sink the book.
03What we built
Every pattern that moves a market
Piggy reads a broad signal set rather than a single feed:
- Price & volume patterns — candlestick shapes, classical chart patterns, support & resistance, volume profile and breakouts, across timeframes from one minute to one week.
- Statistical structure — mean reversion, momentum factors, cointegration for pairs, volatility term structure.
- Market regimes — trend / chop, low-vol / high-vol, risk-on / risk-off, correlation and liquidity regimes.
- Cross-asset signals — what gold, oil, bonds, the dollar index and the VIX are doing, and how that historically rhymes with the instrument.
- Macro & positioning — economic and rate cycles, central-bank stance, commitments-of-traders positioning, large-fund flows, event calendars.
- Sentiment & flow — news tone, social mood and options flow, read as a tilt, never the whole thesis.
One agent, six strategy families
Piggy is a meta-controller: it learns which regime favours which approach and allocates capital between them in real time — trend following, mean reversion, breakout, momentum, volatility and event-driven. It identifies the live regime first (for example, "trending market, low volatility regime"), then weights the families accordingly.
Risk first — sizing built in, not bolted on
A strategy is only as good as the size it's traded at. Piggy sizes every position with Kelly-style scaling to expected edge, volatility scaling so a quiet bond pair and a wild crypto pair don't size the same, and correlation awareness — if two positions are quietly the same trade, the book trims one before drawdown does it.
04Validation discipline
Every strategy that ships through Piggy passes walk-forward validation against years of out-of-sample data before any capital — paper or real — goes near it. The internal acceptance bars are a Sharpe above 1.5, max drawdown under 20%, win rate above 55%, and all eight walk-forward folds positive. These are acceptance criteria, not promises — past simulated performance is not a guarantee of future results, and trading involves risk of loss.
05Tech
06Highlights
- A meta-controller over six strategy families, not a single-playbook bot.
- Reads 300+ signals across price, volume, volatility, regimes, intermarket and macro.
- Knows which regime it's in before it picks a strategy, then allocates capital live.
- Risk-first position sizing: Kelly-style, volatility-scaled and correlation-aware.
- Nothing ships without passing walk-forward, out-of-sample validation.