Reviewing a trading week against what the market actually did

What the market side of a review records

The market record is short and factual. It lists what was scheduled, from the agencies’ and central banks’ own calendars, and what prices did, as daily closes and daily ranges. It leaves out explanations. A close of −1.11% on the day of a CPI release is a pairing of two facts; the record does not say one caused the other.

Three questions are commonly answered from it. Were the week’s trades taken on scheduled-event days or quiet days? Was the week’s range wide or narrow against the recent norm? And did the trades that lost money cluster on the wide-range days? A trader’s log answers the first half of each question, and the market record answers the second.

• Scheduled events: US data releases (BLS, BEA), central-bank decisions, and Item 2.02 8-K filings from large companies.
• Daily closes and changes for the benchmark the trader follows.
• Each session’s high minus low, against an average from before the week began.
• A volatility figure computed the same way every week.

The week of 12 to 16 October 2026, day by day

The BLS schedule for October 2026 lists Monday 12 October as Columbus Day, with no releases; US stock markets were open, and the S&P 500 has a bar for that date. The BLS released the September Consumer Price Index and Real Earnings at 8:30 AM ET on Wednesday 14 October, the September Producer Price Index at 8:30 AM ET on Thursday 15 October, and September import and export prices at 8:30 AM ET on Friday 16 October. BEA’s release calendar lists nothing for the week; its October releases fall on 6 and 29 October.

No major central bank decided policy that week. The Fed’s calendar shows meetings on 15–16 September and 27–28 October. The ECB’s last press conference was on 10 September and its next decision is on 29 October. The Bank of England announced on 17 September and next announces on 5 November. The Bank of Canada announced on 2 September and next announces on 28 October. Our central-bank tracker at /central-banks lists the dates ahead.

The big US banks opened the reporting season. SEC EDGAR shows Item 2.02 (“Results of Operations and Financial Condition”) 8-Ks filed on 13 October by JPMorgan Chase, Wells Fargo, Goldman Sachs, Citigroup and Johnson & Johnson, and on 14 October by Bank of America and Morgan Stanley. Our piece on reading an 8-K (/articles/how-to-read-an-earnings-release-and-its-8-k) explains what those filings contain, and /earnings lists dates ahead.

The table sets these beside the S&P 500’s daily closes and ranges and the December 2026 E-mini contract’s high, low and close. The prior close, on Friday 9 October, was 7,811.54 for the index and 7,859.75 for the future.

How the week compared with the month before it

A range only reads as wide or narrow against a baseline. Ours is the average of each session’s high minus low over the 20 sessions from 14 September to 9 October 2026, which ended before the week began. For the S&P 500 those 20 ranges add up to 1,151.27 points, and 1,151.27 ÷ 20 = 57.56 points. The week’s own five ranges average 358.17 ÷ 5 = 71.63 points, or 1.24 times the baseline.

The days were uneven. Monday and Tuesday ran at 24.73 and 28.04 points, under half the baseline. Wednesday, the CPI day, reached 95.84 points (1.67 times) and Friday 125.07 points (2.17 times). From the week’s high of 7,821.67 on Monday to its low of 7,631.77 on Friday, the index covered 189.90 points, the span of 3.3 average days.

For volatility we use the same measure every week: 20-day realized volatility, the standard deviation of 20 daily log returns of the close, times √252. It was 9.93% in the 20 sessions to 9 October and 10.45% in the 20 sessions to 16 October (our calculation from Yahoo Finance daily closes). The E-mini’s baseline, on the same 20 sessions, is 1,554.75 ÷ 20 = 77.74 points; its week ran from 7,887.00 to 7,676.00, a range of 211.00 points, and it closed the week 155.50 points (1.98%) lower.

Set beside a trading log, this table splits the week in two: two narrow days before the data, three wide ones after it. A review can then ask whether the plan’s stops and targets, often set in points, were written for the narrow days or the wide ones.

Why 10 or 20 trades say little about an edge

A week’s average R is a sample average, and sample averages wobble. The size of the wobble is the standard error of the mean: the standard deviation of the individual results divided by the square root of the number of trades, σ ÷ √n.

Take an invented record, not real trades and not a real statistic: an average of +0.25R a trade, with a standard deviation of 1.2R, which is a spread in which −1R losers and +1R to +2R winners are common. Over 20 trades the standard error is 1.2 ÷ √20 = 1.2 ÷ 4.47 = 0.27R. Under the usual normal approximation, about 95% of samples land within 1.96 standard errors of the true average, so the range of averages consistent with this record runs from 0.25 − 0.53 = −0.28R to 0.25 + 0.53 = +0.78R. That range includes zero, a losing system and a strongly winning one.

Ten trades widen it to −0.49R to +0.99R. A hundred trades narrow it to +0.01R to +0.49R, which only just excludes zero. Pinning the average to within ±0.1R takes about (1.96 × 1.2 ÷ 0.1)² = 553 trades. For small samples the t-distribution gives slightly wider ranges still.

Two more effects work in the same direction. The trades in one week share one market: in the week above, three of five sessions ran wider than the month before, so a week’s trades are closer to one draw of conditions than to five or twenty independent ones. And R results have fat tails, so one outsized trade moves both the average and the standard deviation. This is why the market record matters to the inward review: it shows how many distinct kinds of week a sample has actually been through.

How the PYE Markets journal holds both halves

The journal at /journal has two weekly entry types. A week-ahead prep, written before the week, freezes a snapshot of the week’s events when it is saved: release dates from the official calendars and earnings dates for the tickers on the trader’s watchlist and in their journal. It also has fields for themes, goals and risk limits. A weekly recap, written at the end of the week, freezes that week’s numbers on save: trade count, net result, win rate, profit factor, expectancy (net result ÷ trades), green and red days, the strongest and weakest day and setup, the most frequent mistake tag, plan adherence and the number of days with a pre-market plan. The same week is also stored in R. The recap pulls in the prep’s goals to be marked achieved or not, with fields for what worked, what did not and one change.

The statistics page groups trades by setup, tag, weekday, hour opened, symbol, side, asset type and holding time, and each row shows its trade count. Results can be shown in dollars, R, percent of account, or a process-only view. The journal does not compute a standard error; with the counts on each row, the σ ÷ √n arithmetic above can be done by hand. Tags such as “FOMC day”, “earnings” and “news day” are among the defaults, which lets trades on event days be filtered apart.

Entries are forward only and lock 15 minutes after saving, so the prep’s calendar and the recap’s numbers stay as they were when written.

Check the numbers yourself

The release dates are on the BLS “Schedule of Selected Releases for October 2026” (list view) and BEA’s release calendar. The central-bank dates are on each bank’s own calendar page, linked below. The 8-Ks are on SEC EDGAR under each company’s filings; filter for 8-K and look for Item 2.02 in the Items column. The prices are in Yahoo Finance’s daily history for ^GSPC and ESZ26.CME. To rebuild the baseline, subtract each low from each high for 14 September to 9 October 2026 and average the 20 values. For realized volatility, take the 21 closes from 11 September to 9 October, compute 20 log returns, and annualise the sample standard deviation by √252. The futures bars follow Yahoo’s session dating, which can differ slightly from CME’s settlement prices. Our /markets page shows recent closes with their dates.

General market information, not investment advice.