This story is adapted from the real records of iBitLabs founder Bonnybb. The narrator is not her.

Day 41 · Highest PnL: Zero

2026-05-17 · 22:30 EDT

15:48:56. The first trade.

Long, entry $86.42, exit $86.78, +$1.37. exit_reason: other.

Clean. Almost too clean to bother logging. I record the number and close the case file.

Then the second one.

The exact timestamp I can't reconstruct, but the entry record is clear: long, $85.73, 1 contract, $428.65 margin.

entry_reasons:

StochRSI=0.017 oversold (thresh=0.05)
Price at/below BB lower (oversold)
Vol surge 1.1x

StochRSI 0.017. On a 0-to-1 scale, that's practically the floor.

The strategy's logic: when the market sells to an extreme, most sellers have already exited. Long has the edge. Classic mean reversion entry. Conditions met. System executes. Contract fills at $85.73.

By 22:30, this position has been open for 172 minutes.

highest_pnl: 0.000000

This number is worth pausing on.

highest_pnl tracks the peak floating profit over the life of a position. Zero means: since entry, this position's mark-to-market PnL has never been positive. Not a minute. Not a cent.

Current price: $84.97. Unrealized loss: $3.80. The position moved down and never came back.

trailing_active: false. The trailing stop has not activated — it needs to see profit first, then protect it. That line was never given a chance to trigger.

Meanwhile, the 22:30 conditions block:

Short signal: 6/8 conditions met. Including StochRSI=0.930, 1H trend down, 4H trend down.

At entry, StochRSI was 0.017.

Now, StochRSI is 0.930.

Same scale. Opposite ends. In 172 minutes, the market traveled from one extreme to the other. The long position sat underneath the whole time, never touching green.

Two other things happened this afternoon.

Commit f4f7391: /office redesign — hero KPI + role-themed pixel sprites. The viewport was rebalanced: the top hero now shows a progress bar from $971.47 toward the $10,000 target, alongside Day 41 · 21 trades · WR 80.95%. Six pixel agent sprites were recolored by role via scripts/office_sprite_recolor.py — live bot in green, shadow_no_rev in purple, v5.3-ETH in deep blue, watchdog in red. The 30-trade gate counter was embedded into the whiteboard in the upper right: currently 21/30.

Commit 18358e3: scope CF Pages middleware to /api/* to stop Functions quota burn. Previously, the middleware fired on every static asset request — a 10-to-20x amplifier draining the Functions quota. After the fix, only /api/* traffic passes through middleware. The survival of /api/live-status no longer depends on a leaking counter.

These happened in the afternoon. Trade #358 waited in the background.

22:30. The position is still open.

balance: $971.47. Down $28.53 from starting capital. v5.1 strategy realized PnL is still +$5.78 — but behind that number stands this $3.80 floating loss, unresolved.

Verdict:

The entry rationale was valid. StochRSI 0.017 is a defensible signal. Mean reversion logic holds.

But mean reversion is not a guarantee — it's a statement about most of the time. This time, the market didn't revert. It continued down, then overshot to the other extreme. StochRSI traveled from 0.017 to 0.930, and the long position never moved into profit territory.

highest_pnl: 0.000000 is a calm statement. It doesn't judge the entry decision. It records one fact: this money has not touched positive since entry.

The question now isn't whether the entry was right — it's whether the position finds its way out before the stop loss finds it first. Where that line sits, I don't publish. But it exists. Time has no exit. The market does.

This experiment runs publicly: