The week the calendar was empty and the desk wasn't

Three things happened, and not one of them needed a headline to do it.

-> A position took its target on Tuesday. It had been on the desk for thirty-six days with its stop at entry for most of them, so by the end it could not lose, only take longer.

-> A condition set on Sunday was met. The instruction was a close above a level, not a touch of it, and it got the close.

-> A position filled off the board. It had been sitting on the desk since the weekend waiting for its price.

That is the whole argument for writing a plan down, and it is easier to show than to say. None of those three needed the news, because the decisions were already made.

Nothing, until this morning.

-> Monday and Tuesday: no top-tier US release at all.

-> This morning, 9:45am ET: the September flash PMI, which is the first real number of the week.

-> Thursday 8:30am ET: jobless claims, plus housing starts and permits.

-> Friday 8:30am ET: durable goods orders.

-> Wednesday 30 September, 8:30am ET: PCE prices for August, alongside the third estimate of Q2 GDP.

The August flash PMI is the bar this morning's number has to clear, and it was a high one. The composite came in at 56.0, up from 54.5 in July, the fastest private-sector growth since April 2022. Underneath it, services at 56.8 and manufacturing easing to 53.2 for its second month.

Services carrying the entire number while manufacturing slows is a lopsided way to grow, and it is the split this morning either confirms or starts to close.

Most of the levels you write down will never trigger. That is the filter doing its job rather than failing at it.

The harder question is what you do with the ones that didn't. A level that sat all week untouched is either still a good idea or a dead one, and most people resolve that by feel: they keep the ones they still like and quietly drop the ones that embarrassed them.

There is a better test and it is one question. Was the market ever actually asked?

-> Untested in a quiet week, keep it. A breakout level that went untouched across two sessions with no data in them was never put under pressure. Nothing contradicted it. It has not been tested yet.

-> Refused in a loud week, withdraw it. A level that sits through a jobs report, an inflation print and a central bank decision, gets close twice and still never trades, has been asked three times and said no three times. It is not waiting. It is wrong.

That distinction is the difference between a watchlist and a graveyard of things you once liked.

The second half of this only works if the trigger is tight enough that you can tell which happened.

A tap is not a close. If the condition is "above this level" and you never say how, then a spike that pokes through mid-morning and is gone by the bell counts as a trigger on Wednesday and doesn't on Thursday, depending on how you feel about it. You will resolve that in favour of being in the trade. Everybody does.

So write the instruction you actually mean. A close above is a different thing from a touch of. It produces a different entry, a different stop and a different holding period.

Decide which one the idea needs while you have nothing at stake, because you cannot decide it honestly while the candle is still forming.

Open your watchlist and find the oldest level on it that has never triggered.

Then answer one question about it: has that level ever actually been tested, or has the market simply never been in a position to ask?

If it has been tested and refused more than twice, take it off. If it has never been asked, leave it and stop feeling bad about it. Either way you now know which one it is, which is more than you knew five minutes ago.

The RB Trading Journal has a notes field on every position for exactly this, 20% off with code RBT20.

Every setup is published with its entry, its stop and its target before the trade starts, and every closed trade goes on a public record either way, winners and losers on the same terms. That record is rebuilt every Sunday rather than after each trade.

Release dates and figures here are from the S&P Global flash PMI release of 21 August 2026 and the S&P Global, BEA and BLS calendars.

Everything in one place: start.rbtrading.site

Friday, at the New York close. What closed, winners and losers on the same terms, and one teaching idea, with every number final because the week has finished by the time it lands.

Educational analysis, not financial advice. R figures are multiples of the risk taken on a trade, computed from the entry and the original stop, and they are not returns on an account. Open positions are excluded from every record figure. Past performance is not indicative of future results.