Run 4, Day Two: Seven Exits, One Winner, −$54.84
Day two of Run 4, from midnight to midnight CDT. Seven positions closed, two wins, five losses, and the bot goes to sleep holding nothing. All numbers straight from trades.log, nothing smoothed. Paper trading only. Simulated, hypothetical results. Not financial advice. Paper performance proves nothing about live trading.
Equity at open: $967.29 → Equity at close: $912.45
Day P&L: −$54.84 (−5.7%)
Run cumulative: −$87.55 (−8.8%) from the $1,000 start
Executions: 11 (4 buys, 7 sells) · Closed trades: 7 · Wins: 2 · Losses: 5 · Fees: $2.00
Open positions at close: 0
The trades, in order (times CDT)
12:00 AM — Overnight positions exit first. Pepper hits its one-hour mark: sold @ 0.000510, +1.1% (+$0.02) — entered 11:00 PM last night @ 0.0005044 (RawScan feed, liquidity $79,563). NMD exits its hour @ 0.001592, −0.6% (−$0.39) — entered @ 0.001601 (RawScan, liq $192,602). Then the ugly one: DARK hits the −50% stop and is out @ 0.000006, −91.0% (−$22.77) — entered @ 0.00006910 (RawScan, liq $34,954, tagged FAST-EXIT). The stop fires at −50% between scans (the bot checks every 10 minutes), so on a fast dump the actual exit can be far below the trigger. DARK went essentially to zero inside its hour.
All day — the daily-loss guard trips: "Daily −3.2%: no new entries for 24h." The rule blocks entries whenever the day's drawdown passes −3%. The scanner kept working all day (BESTBUYS lines every 10 minutes: DOT and ATOM leading the MEDIUM board, OP pinned at 50.0), but no signal was allowed in. SAFE and MEDIUM took zero trades — their Donchian breakout entries never triggered.
7:00 PM — the guard clears (the daily anchor resets at midnight UTC) and the bot buys three HIGH-tier positions: KURIBO @ 0.00016240 (RawScan feed, liq $33,748, tagged CONFLUENCE — flagged by both feeds, so the v4 conviction rule sizes it up to $40 instead of $25; FAST-EXIT band liquidity) · ODIE @ 0.00132800 ($25, RawScan, liq $110,276) · BITCHES @ 0.00361500 ($25, RawScan, liq $171,516). Entry rule: AlphaPulse ignition or RawScan composite score ≥ 70, v4 dead-zones skipped (age 3–12h, score 94+ unless under 3h old, velocity 30+).
8:00 PM — KURIBO blows through the stop: sold @ 0.000059, −63.8% (−$25.66). The sized-up confluence pick — the bot's "best" entry of the day — became its worst exit. Then the first true ignition entry of the evening: IRYS @ 0.08078278 ($25, AlphaPulse scanner ignition signal, liq $895,722).
8:10 PM — the hour expires on the two standard entries. ODIE out @ 0.001141, −14.1% (−$3.74). BITCHES out @ 0.002555, −29.3% (−$7.51). The universal 1-hour exit cutting losses before they can become DARK-style zeros.
9:00 PM — the day's one real winner. IRYS exits its hour @ 0.102300, +26.6% (+$6.34). The bot ends the day flat, holding nothing, at $912.45.
The ledger
| Trade | Tier | Entry → Exit | Paper P&L | Rule that fired |
|---|---|---|---|---|
| Pepper | HIGH | 0.0005044 → 0.000510 | +$0.02 (+1.1%) | 1-hour exit |
| NMD | HIGH | 0.001601 → 0.001592 | −$0.39 (−0.6%) | 1-hour exit |
| DARK | HIGH | 0.00006910 → 0.000006 | −$22.77 (−91.0%) | −50% stop (slipped) |
| KURIBO | HIGH | 0.00016240 → 0.000059 | −$25.66 (−63.8%) | −50% stop (slipped) |
| ODIE | HIGH | 0.00132800 → 0.001141 | −$3.74 (−14.1%) | 1-hour exit |
| BITCHES | HIGH | 0.00361500 → 0.002555 | −$7.51 (−29.3%) | 1-hour exit |
| IRYS | HIGH | 0.08078278 → 0.102300 | +$6.34 (+26.6%) | 1-hour exit |
What the day actually taught us
Five of the seven exits closed on the clock. ODIE −14.1% and BITCHES −29.3% would both have been allowed to keep falling under a longer hold; the hour mark turned them into measured losses instead of potential DARKs. Two days running, the exit rule is the only part of the strategy with a consistent story.
The v4 package sizes confluence picks at $40 — the idea being that dual-feed agreement deserves more capital. KURIBO was exactly that pick, and it stopped out −63.8%, the second-worst loss of the run so far. More capital on the "best" signal bought more loss. That's a real data point for the v5 point system currently being built: a 100 has to be genuinely hard to reach, because "high confidence" has cost us $25.66 today.
Two wins (+$6.36) against five losses (−$60.07). The stops slipped: DARK printed −91.0% and KURIBO −63.8% against a −50% trigger, because the bot scans every 10 minutes and fast dumps don't wait. A tighter scan cadence — the 22-minute scanner cycle just approved this week only makes it worse, not better — or an intra-scan price watch would be needed to make the stop a real stop instead of a suggestion.
Also worth logging honestly: the bot's plumbing showed strain today. Several Kraken data pulls failed with internal errors and remote-end disconnects during the daytime scans (DATA-SKIPs on SOL, DOGE, SUI, POL and others), and the STATUS log tag flipped from v3 to v5 mid-evening as the in-progress v5 candidate pipeline went live. The trades above still settled from the log as written.
The dashboard holds all four runs — wins, losses, fees, halts — for anyone to audit. Run 4 continues tonight holding nothing, down −8.8% for the run. No real money anywhere near this; the whole point of doing it in public is that the entries have to earn the capital before the capital exists.
Every trade, every fee, every stop — on the public Nova Trader dashboard.
Open Nova Trader →