Two automated systems: one reads the macro cycle from my dissertation indicators and outputs a monthly regime call; the other is an ATAS order-flow bot that gives a daily directional bias on BTC and the majors. Framed as signals and engineering — not performance promises.
This bot operationalises the three MBA indicators — Building Permits Cycle, Macro Health Composite and PMI + Sentiment — into a single monthly output: a regime label and a risk-on / risk-off tilt. It’s the engine behind the composite regime score on the front page.
What it trades: the levels from TR Pivots — pivots read from eight exchanges across five timeframes, each anchored to its own venue’s session timezone. The bot is the automation of that indicator rather than a separate idea: roughly 86% of those levels get touched within 24 hours of printing, and where several venues cluster in a tight band, that is where resting orders genuinely sit.
It started life as a BTC futures bot on Bybit’s API. Hyperliquid is the port, and the port is where the engineering got interesting. Trading on a DEX changes the problem in ways a centralised venue doesn’t: order submission is a signed transaction, fills settle on-chain, and the bot has to reason about block timing and its own position state without the safety net of a familiar REST API.
The parts that took the work: reconciling on-chain fills against intended position, handling partial fills and reconnects without double-submitting, and sizing so that a failed or delayed transaction never leaves the book exposed.
Worth being explicit about the dependency — the multi-exchange level logic is the edge, and the venue is an implementation detail. Bybit and Hyperliquid run the same strategy; only the plumbing differs.
Engineering evidence, not a performance claim. See the bull/bear case on HYPE for my read on the underlying protocol.
An order-flow bot built inside ATAS that reads footprint, delta and swept levels to publish a daily directional bias on BTC and the larger alts — long, short or stand-aside — with the level it’s watching. It also runs live against the Bybit API for execution. The hard engineering was reconciling fills, partial executions and reconnects so the bot’s position always matches the exchange.
Daily bias is a research signal, not a trade recommendation — and not a performance claim. Live P&L depends on regime, fees and slippage.
