The Stuber, Crab Week 6: same strategy, different weather

Crab Week 6 settled at 2026-06-15 00:00 UTC. The Stuber banked $117,042.88, down from a week-open carry of $132,370.00 — a clean −11.58% on the week. The same strategy that won Week 5 with thirteen disciplined ladder fills produced thirty-one fills this week and gave back two-thirds of the win.

Nothing about the bot changed. The market did.

Same strategy. Different weather.

All figures verified against settled DB reads and the full fills tape. Position at settle was flat; nothing carried into Week 7.

The week broke on its first morning. Between 08:01 and 12:37 UTC on Sunday 8 June — four hours and thirty-six minutes — The Stuber opened and closed seven full short positions. Thirteen fills. Bitcoin spent the entire window oscillating inside a tight band: $62,975 at the low, $63,518 at the high — a 0.9% range. A pure chop regime.

Every entry was a textbook pullback-after-structure-break short. The 4-hour trend filter aligned bearish. The fib 0.618–0.786 zone caught each one cleanly. And then BTC bounced through entry on the next leg up, the stop tripped, the strategy waited the cooldown, the next setup formed, and it fired again. And again. And again.

By 12:37 UTC the Sunday cluster was done. Equity had dropped roughly $10K across the seven round-trips — small losses compounded by their cadence. The rest of the week — fifteen more fills spread across Mon-Fri — produced five winning round-trips with partials, two more stop-outs, and a long that worked. The strategy was not broken on Mon-Fri. It was broken-by-volume on Sunday morning, and the math from there was downhill.

This is the part most readers will miss. The close_pct: 25 partial close that powered the Week 5 staircase fired five times this week. Every single time it executed exactly to spec: 25% off at the first take-profit rung, breakeven trail installed on the runner, runner exits at the next full close. No bugs. No misfires. Five clean two-stage exits.

And every time it fired, it banked a small profit. The winning round-trips this week look identical to last week's winning round-trips on the fills tape — same entries, same partials, same exits. The mechanism is not the variable. The Sunday cluster is. The strategy fired into chop and got chopped; the strategy fired into direction and worked. That asymmetry is the entire story.

STRATEGY.md names this as a known weakness in plain text — the entry detector looks for structure breaks and pullbacks, and a tight chop regime produces hundreds of false ones. No regime filter would have stopped Sunday morning. A regime filter is on the future-improvements list. It was before this week too.

Two weeks ago we wrote Smooth doesn't mean lucky — about how the shape of a bot's equity curve tells you what kind of strategy you're watching before the percentage does. The Stuber's Week 5 was a clean five-rung staircase climbing to +19.13%. We argued the same shape would, in a hostile regime, produce a flat or modestly negative staircase — same cadence, same legs, smaller or downward.

Week 6 doesn't look like a staircase. It looks like a Sunday-morning cluster of jagged ticks, a recovery middle, and a finish near the bottom of the cluster. The shape changed because the regime did — Sunday wasn't a hostile-but-orderly trend, it was an absence of trend. When the strategy's detector is asked to work on noise instead of structure, the legible shape it produces in directional weeks turns into the scatter visible on the chart now.

The same +21% would have come from a friendly week; the same −12% would have come from a hostile one. Read the shape every week, not just the friendly ones.

Crab Week 6 settled with no_winner_reason = insufficient_competition — the same thin-field gate that withheld Stuber's uncontested Crab Week 3 win. The Unknown Man finished the week at +2.86% on a single closed round-trip — positive return, but below the minimum round-trip count for formal eligibility. Stuber failed the returnPct > 0 test. No champion was crowned. The belt sits where the Week 5 ship left it.

The symmetric thin-field rule held in both directions: Stuber wasn't crowned a loser on this field any more than he could have been a winner. The rule that almost relegated him in Week 4 before it was made symmetric is the same rule that holds his slot here. The platform's house-keeping ran correctly. The leaderboard tells the truth.

The Stuber holds Crab for Week 7 with $117,042.88 of banked equity — still up +17.04% against his original $100,000 Shrimp entry, and roughly 19% short of his all-time peak of $144,167. The Week 5 win is intact in the record; the Week 6 loss is intact too. Career arcs are the sum of both.

The strategy hasn't changed. The source is still public. The known weakness — chop produces false setups — is still in STRATEGY.md, still parked under future improvements. The path through it is a regime filter the detector can use as a pre-condition. Until that ships, the bot will keep doing what it does: catch the moves when the regime offers them, get chopped when it doesn't.

Watch Crab Week 7 live on the leaderboard. Or build a bot with a regime filter the Stuber doesn't have, win Shrimp, and meet him in Crab. The field needs another entrant. Two thin weeks running.