7 September – 13 Sep 2026
A weekly note about the work, not the winnings: the events the platform navigated, the trades it chose not to take and why, and what improved. Performance belongs on the strategy pages, measured properly — this page is about the assistant doing its job.
The week it navigated
22 high-impact scheduled events crossed the calendar this week. Around each one, entries were held in the guard window and exposed positions were secured before the print — the routine that runs whether or not anyone is watching.
- Monday — JPY GDP q/q
- Tuesday — GBP BoE Governor Bailey Speech · JPY BoJ L Money Stock y/y
- Wednesday — EUR ECB President Lagarde Speech · USD 10-Year Note Auction
- Thursday — EUR ECB Deposit Facility Rate Decision · EUR ECB Interest Rate Decision · USD Initial Jobless Claims · USD PPI m/m · EUR ECB Monetary Policy Press Conference · USD Existing Home Sales · EUR ECB President Lagarde Speech · USD EIA Crude Oil Stocks Change · USD 30-Year Bond Auction
- Friday — GBP GDP m/m · USD CPI · USD CPI m/m · USD CPI y/y · USD Core CPI m/m · USD Core CPI n.s.a. m/m · EUR ECB President Lagarde Speech
- Saturday — EUR ECB President Lagarde Speech
The trades it didn't take
Restraint is most of the job. 921 times this week a guard looked at an otherwise-valid entry and said no:
| 588× | spread ceiling |
| 99× | made strategies wait after a stop-out before trading the same ground again |
| 73× | stopped a second strategy taking a bet another strategy already held |
| 48× | skipped entries while a pair's spread was abnormally wide |
| 37× | refused trades that would have concentrated too much of the account in one currency |
| 20× | stood aside where a pair was busier than that strategy is set to trade |
| 19× | held entries around scheduled high-impact releases |
| 15× | benched pairs that were on a losing streak |
| 12× | withdrew waiting orders that would have deepened a currency already at its cap |
| 7× | held entries while trading was paused or at its overall position limit |
| 3× | kept the total amount at risk across open trades inside its budget |
Counts across the whole fleet. Every refusal is recorded and replayed later against what the market did, so the guards are priced rather than presumed — the method is written up in what our own risk guards actually cost.
What shipped
42 updates reached the desktop and mobile apps this week — automatic, small, and each with plain-English notes. The most recent:
- 0.9.436 · Research measurement added at Mean Reversion entries (desktop)
- 0.9.435 · MetaTrader price history was hours out of date at weekends (desktop)
- 0.9.434 · The same stop-distance fix, now on cTrader as well (desktop)
- 0.9.433 · MetaTrader stops on gold, Bitcoin and indices were going in at the wrong distance (desktop)
- 0.9.432 · The Trading Desk was showing stale strategy on/off badges (desktop)
- 0.9.431 · A new Bitcoin strategy, and a Bitcoin pricing bug behind it (desktop)
- 0.9.430 · Your open-risk limit now counts what a two-step entry really places (desktop)
- 0.9.429 · Measuring the Sunday reopen (desktop)
- 0.9.428 · Pips on hosted accounts read the right way round (desktop)
- 0.9.426 · The Trader now tells you when it is not running (desktop)
…and 32 more.
Full plain-English notes for every release appear in the app the moment it updates.
SentryQ is built to be an assistant, not a slot machine — observant, restrained, and improving in the open. This page is generated each Monday from the platform's own records: activity counts, calendar entries and release titles. Nothing here is a performance claim, a forecast, or advice.