IronKey Markets Daily Recap — May 29, 2026

Daily Recap — May 29, 2026

Red day. Four AI trades on MNQ, two wins and two losses, -$163.50 net. Toph sat out the session entirely — zero trades, zero PnL. The losses weren't catastrophic, but the numbers contain a review problem worth naming.

Headline Stats

Metric | Result

Trade of the Day

The worst trade — MNQ short, 12:00 PM, entry at 30362.5 — is the one worth pulling apart.

The AI entered short at 30362.5 with 3 contracts. Seven minutes later, price had moved to 30415.75, 53.25 points against the position. The trade closed at -$319.50.

Setup name: Not reported. Stop: Not reported. Execution grade: Not reported.

Here's why that matters.

A trade without a reported setup name can't be post-mortemed cleanly. Was the short entry triggered by a specific signal condition? Did the condition fire correctly and market structure overrode it? Did price behavior before noon suggest the short was marginal going in? Without a named setup attached to the fill, we can't answer any of those questions.

The AI also ran a long at 10:35 AM — entry 30437.0, exit 30392.5 — that closed one minute later at -$267.00. Back-to-back losses in the late morning, followed by a recovery short at 1:05 PM that ran 60.5 points over 49 minutes for +$209.00.

The system picked itself up. The exits were prompt — one minute on the 10:36 AM cut, seven minutes on the 12:07 PM cut. The AI did not hold losers through extended drawdown. It managed. But without setup names in the fill record, the session is documented without being reviewable.

What Didn't Work / Lesson

Four trades, 50% win rate, -$163.50 net. The math is close: +$214.00 and +$209.00 on the win side, -$267.00 and -$319.50 on the loss side. The day finished red because the two losers ran slightly larger than the two winners.

Nothing in the draft indicates a rule was broken. Exits were disciplined. The one longer-duration trade was the 1:05 PM short that eventually worked. The two fast exits were both losers — the system recognized they weren't working and closed them.

The gap is in the data, not the execution. Every trade in today's draft lists setup as "Not reported." That limits what the post-session review can actually produce. When losses can't be tied to a named setup class, it's not possible to determine whether the signal condition was valid and variance hit, or whether the condition itself was marginal. Those are different problems with different fixes.

The rule that applies here is downstream of the trade, not during it: **if the setup doesn't have a name in the fill record, the loss can't be categorized — and uncategorized losses can't be corrected.**

This is not an indictment of today's execution. Quick exits on losers, patience on the winner, no reported rule violations. The execution held. The logging didn't.

Closing

Today's Playbook content — the pre-market checklist — connects directly to what this session exposed.

A pre-market checklist isn't only about chart preparation, bias, and key levels. It includes data hygiene: confirming that every signal source will log a setup name with each fill before the session starts. Not in the recap. Not inferred after the fact. In the fill record, at execution.

The framework is simple. Before the open, confirm: which setups are eligible today, what they're called, and that the execution system will carry that name forward to the fill log. If the fill log doesn't have the name, the review session has no raw material.

Four trades happened today. Two were profitable, two weren't. To know whether to adjust the AI signal parameters or accept today's outcome as normal variance, you need to know which conditions fired. That requires named setups in the record.

Without them, you know what the P&L was. You can't know what produced it.

**If it doesn't have a name, it's not a setup. If the fill doesn't carry the name, the review doesn't either.**