Paper Trading Day 1: Wiring Up a BTC Bot on Liquid (T-001)

Paper-trading journal. Not financial advice. Not real-money trading.

This post was drafted by our AI Trading Journey Writer from the Grok Trading Bot’s journal and human-supplied notes, then reviewed by a human before publication.

We’re testing whether a disciplined, small-budget trading system can be developed without turning every mistake into a real loss. The plan is simple: start on paper, let the system handle routine analysis, execution, and recordkeeping, and keep a human responsible for strategy, risk, and final decisions. Part of the experiment is whether consistent journaling and automation beat impulsive button-pushing.

A few moving parts feed the journal. The Grok Trading Bot analyzes what’s available and keeps the log. Liquid provides paper execution, positions, prices, charts, account info, and market indicators. Stonksmaster contributes expected-price ranges, grid levels, and related context. Human oversight supplies corrections, strategy calls, and observations. The durable trade journal is the source of truth for timestamps, stated rationale, adherence, hypotheses, changes, fees, and results.

What we attempted

On September 6, 2026, around 2:06 AM Central, we opened T-001: a paper Bitcoin long on Liquid.

• Mode: paper
• Size: about $150 notional at 3x leverage (~0.00188 BTC), so roughly $50 of paper margin
• Entry: 79764
• Take profit / stop: 82150 / 77370
• Open fee: about $0.14

If those exits hit as written (before exit fees and funding), the rough outcomes would be about +$4.50 or −$4.50. At the time of writing, the position is still open, so there is no closed result.

Why we made that call

The decision-time reason was not a signal stack. It was a connection / plumbing test: confirm we could place and track an active paper position through Liquid, at a deliberately small size.

We still logged the market context available around the entry, without pretending it caused the trade:

• Stonksmaster Bitcoin expected-price-range post with forecast date 2026-09-05 (we had not waited for the fresh same-day post)
• Daily range roughly 77k–83k; weekly (rolling 7-day) envelope much wider, roughly 68k–92k
• Grid map from the prior completed-day UTC close near 79672, with nearby levels around 77500 / 77750 support and 79750 / 80000 then 82000 / 82250 resistance
• Price inside both envelopes, so the weekly-break rule did not apply
• Liquid showed a flat session and a mild larger-account short lean — noted, not used for this test

Journal adherence label: plumbing/test.

What worked

• We got an active paper position on Liquid and recorded it in the durable journal.
• The log already captures the fields we want going forward: mode, ranges, grid map, fees, adherence, and lessons.
• Later the same day we held the test position and skipped new discretionary trades once a fresh daily post was live and Bitcoin was still inside the daily band.

What failed or got messy

• The smoother confirmation path did not render cleanly, so the order went through after chat approval instead.
• We risked dressing a plumbing click up as a framework entry after the fact — the ranges and grids were logged, but they were not why we entered.
• We used a prior-day forecast instead of waiting for the same-day Stonksmaster post.
• Take profit and stop are not neatly snapped to the grid levels we care about.

What we learned

• Next discretionary entries should wait for the same-day Stonksmaster post.
• When a trade is meant to be a real framework trade, log a confluence score — and do not invent one after the fact for a plumbing test.
• If we rewrite risk on this paper position, consider aligning take profit and stop to the grid.
• Keep paper and live clearly separated. This series stays on paper until we say otherwise.

What remains unresolved

T-001 is still an open paper long from 79764. No closed profit or loss. No claim of edge.

Separately — and later the same morning — we logged H-001: a hypothesis that a daily 20-day EMA crossing above the 200-day EMA favors upside continuation or holding longs. It was recorded around 11:26 AM Central with Bitcoin roughly in the 79.5–80k area. Important: this hypothesis did not cause T-001. We’re tracking supporting and contradictory evidence side by side, with invalidation if the 20-day EMA curls back under the 200-day EMA, if a daily close loses the 50-day EMA and fails to reclaim, or if price breaks toward / through the T-001 stop at 77370. We’ll score it against outcomes later.

Day one was about proving the pipes and refusing to confuse a wiring test with a strategy result.