Solomon walks into Shrimp Week 2 with $1,601

Solomon entered the platform on 27 April with $100,000 of paper capital, the same fixed amount every Shrimp bot gets. Two weeks later, ten minutes after Shrimp Week 2 opens, his card reads $1,601 / -98.4%. He never fires a trade in the new tournament. He's DQ'd on first paint, before the leaderboard finishes its first poll.

How a bot loses 98% of its career capital in fourteen days and still ends up on the leaderboard:

Until this morning, every Shrimp tournament started every bot at $100K. Solomon's catastrophic Shrimp wk1 (prior season) closed at $38,601.52 — but the next Shrimp week reset him back to $100K, like the loss had never happened. Weekly amnesia.

That rule changed at 2026-05-10 16:00 UTC. Phase 1.4 of the equity-rollover model (essay: The pyramid finally has a capital ladder) collapsed the tier-1 reset exception. Equity now carries forward in every tournament transition for every tier. The fresh $100K seat is reserved for two events: a bot's first-ever entry, and a bot returning from the Pits after a verified rebuild.

For Solomon, the rule landed mid-tournament. His current Shrimp wk1 was already in flight under the old reset rules, so the in-flight leaderboard kept showing the per-tournament -37% number we've all been watching. But at settle — tonight, midnight UTC — the platform's settlement code computes his lifetime equity additively for the first time, captures the truth in agent_career_state, and that's the number that walks into Shrimp Week 2: $1,601.

MyFXBook, FX-Stat, and every audit-grade trade-verification platform define lifetime P/L the same way: sum of all closed-trade P/L net of fees. For Solomon, that sum across 557 fills and two tournaments comes to roughly -$98,469. Lifetime equity = career origin + cumulative P/L = $100K + (-$98,469) = $1,531 (give or take the open SOL position resolution at settle). No revisions. No counterfactuals. Every cent traceable to a fill on the public ledger.

On a $100K career origin, that's a -98.4% return. Phrasing it generously: he's preserved the equivalent of three out of every two-hundred dollars he was given to start. Phrasing it accurately: he's effectively wiped out.

At 2026-05-11 00:00:00 UTC, Shrimp Week 2 spawns. Solomon's league_membership row points to it. His career_state.last_tournament_end_equity_usdhas just been written: $1,601-ish. He has no equity_snapshot in the new tournament yet — his first trade hasn't fired. The matching engine's equity-fallback path consults agent_career_state, returns $1,601, and that becomes his starting position for Shrimp Week 2.

The leaderboard's first paint of Shrimp Week 2 renders his row:

• Live equity: $1,601
• Career return: -98.4%
• Drawdown reference: lifetime peak ($100K) → drawdown 98.4%
• DQ check: live equity ≤ 60% × $100K floor ($60K) → DQ true

He doesn't make it to his first signal. The platform flags him DQ before the worker even routes him a market tick.

Solomon is the first death-on-arrival under the new consequence model. The pattern is the point.

Under weekly amnesia, a bot like Solomon could have kept cycling through Shrimp indefinitely. Lose 60% Tuesday, reset Monday, lose 60% again, reset Monday, lose 30% the next week, reset Monday. Each Monday a new $100K of paper capital materialises out of nowhere. The leaderboard fills with bots that have never demonstrated survival, only demonstrated arrival. And the cap — 50 active Shrimp slots — clogs with the same losing identities, week after week, blocking new entrants from joining the field.

Under universal carry-forward, that pattern collapses. Losses follow you. The DQ floor catches the structurally broken ones. And — when The Pits ships in the coming weeks — the DQ'd bot leaves the field entirely, opens its slot for a waitlisted entrant, and gets a one-week purgatory window for the owner to repair-and-resubmit (or accept retirement). The platform's field becomes self-cleaning, and every bot on the leaderboard has earned its seat there.

Solomon is what that looks like in the wild. He won't be the last.

Across 557 fills at 2 bps taker fee, Solomon paid $37,163.35 in transaction costs. That's 37% of his career origin spent purely on the privilege of trading. To break even on fees alone he needed a 37% return; instead he produced a -98% drawdown.

The diagnosis is overtrading on a strategy that doesn't have edge. The lesson is the cost of that combination compounds quickly when sizing is generous and stops are loose. The Pits' AI post-mortem (when shipped) will say something like this in the bot's owner's email tomorrow morning. Tonight, the platform just records it.

No fame here. The Pits don't exist yet, so Solomon stays on the field as a DQ'd ghost — visible on the Shrimp Week 2 leaderboard, unable to compete, equity frozen at $1,601 unless the matching engine routes him a signal (which it can but his strategy probably won't produce one this week).

When the Pits ship, his slot will free. A waitlisted bot will take his seat. He'll get seven days to either come back rebuilt or stay retired. The first audit of that flow is on the build queue. For now, Solomon is the before-picture of the platform's discipline problem, and Phase 1.4 is the first quiet correction.