Weekly gold outlook

Gold forecast scorecard: every week, misses included

Every archived week of the weekly gold forecast with its bias call, expected range, what gold actually did and whether the call was a hit or a miss, plus the running hit rate. Education, not signals.

Read this first: the archive is young. It started on 24 October 2026 and holds 9 weeks. 8 of them were back-filled from price data that day with the same rules, so they show how the rules would have scored, not calls we published before the week began. The first week published in advance is the week of 26 October 2026; it is scored automatically once it ends. A handful of weeks is far too few to judge any method, and we will not pad this page: it grows by one row a week.

Weeks in the archive98 back-filled · 1 published in advance · 1 pending

Hit rate, all scored weeks75%6 hits of 8 scored weeks

Hit rate, published in advance onlyNone yetno week published in advance has finished yet

Band held, all scored weeks12%high and low both inside the expected range

Week of	Bias call	Expected range	Actual close	Actual range	Hit or miss	Band held	Running hit rate
26 Oct 2026 published in advance	Range	$3,906 – $4,110	Pending: scored automatically once the week has ended in the data
19 Oct 2026 back-filled	Range	$4,081 – $4,281	$4,007.9open $4,185.0	$3,989 – $4,211	Miss	No	6/8 (75%)
12 Oct 2026 back-filled	Range	$4,114 – $4,318	$4,181.1open $4,220.8	$4,095 – $4,242	Hit	No	6/7 (86%)
5 Oct 2026 back-filled	Range	$4,060 – $4,265	$4,216.3open $4,169.4	$4,091 – $4,234	Hit	Yes	5/6 (83%)
28 Sep 2026 back-filled	Range	$4,219 – $4,424	$4,162.3open $4,315.0	$4,143 – $4,316	Miss	No	4/5 (80%)
21 Sep 2026 back-filled	Range	$4,320 – $4,530	$4,321.2open $4,413.0	$4,278 – $4,422	Hit	No	4/4 (100%)
14 Sep 2026 back-filled	Range	$4,300 – $4,518	$4,424.9open $4,375.0	$4,273 – $4,440	Hit	No	3/3 (100%)
7 Sep 2026 back-filled	Range	$4,366 – $4,587	$4,408.9open $4,466.5	$4,333 – $4,489	Hit	No	2/2 (100%)
31 Aug 2026 back-filled	Range	$4,415 – $4,645	$4,476.6open $4,483.2	$4,329 – $4,558	Hit	No	1/1 (100%)

Hit or miss uses the rule printed on the forecast hub: Lean higher is a hit on a close above the week's open, Lean lower on a close below it, Range on a close inside the expected range. "Band held" is a separate, stricter test. Prices are COMEX gold futures daily data, which can differ from a spot XAUUSD quote by a few dollars.

This week's forecast First archived week

How the method works

Every week the same recipe runs on the daily gold data up to the Friday close. The bias comes from one rule: Lean higher when last week's close is above both the 20-day and 50-day averages and each of the last four weekly closes is higher than the one before. Lean lower is the mirror: close below both averages and four falling weekly closes. Anything else is Range. The expected range is last week's close plus and minus half of the average weekly range of the previous 12 weeks. Nothing is adjusted by hand, and the numbers are frozen in the archive file the first time they are published.

Once the week has ended, the build adds what gold actually did: open, high, low and close. The page then scores the call with the rules above and updates the running hit rate in date order, so you can see how the record developed rather than a single headline number.

How to read the numbers

Treat the hit rate as a description of the rules, not a promise. A few weeks of results can look very good or very bad by chance, and back-filled weeks are a test of the rules on past data, not a track record. The band-held rate is usually much lower than the hit rate, because gold often travels further than an ordinary week, especially around a Fed decision, CPI or the jobs report. That is why the forecast lists events and conditional scenarios instead of a target.

The useful habit is the one traders build inside the community: mark the levels on Sunday, write down what each scenario would mean for your own plan, and review on Friday which levels actually mattered. Our posts on how a trading pod works and the FOMC day gold playbook show how that review is done.

Education, not signals. Past weeks do not predict future prices. Nothing here is a buy or sell recommendation.

Frequently asked questions

How is a weekly gold forecast scored as a hit or a miss?

By fixed rules, applied automatically. A Lean higher call is a hit when the week closes above its own open, a Lean lower call when it closes below its open, and a Range call when the week closes inside the published expected range. Separately, the band is marked as held when the week’s high and low both stay inside it.

Why are some weeks marked back-filled?

When the archive was created, the weeks before it were calculated from price data with exactly the same rules, using only the daily bars before each week. They were not published in advance, so they show how the rules would have scored, not a record of published calls. They are labelled on every row.

Are misses ever removed or edited?

No. Every archived week stays on the page, hit or miss. Published numbers are written once and never changed; only the outcome is added after the week has ended.

Why is a Range call easier to hit than the band holding?

A Range call is scored on the close alone, while the band only holds if the whole week stays inside it. A week can spike outside the band and still close inside it. Read the two numbers together: a high Range hit rate with a low band-held rate means gold travelled further than an ordinary week but came back.

Is this a trading signal service?

No. The weekly forecast is education, not signals: a rule-based bias label, a statistical band and conditional scenarios. The scorecard exists so you can judge the method honestly before you use any of its levels in your own plan.

Generated automatically from the forecast archive. Education, not signals. Not financial advice. See our risk disclaimer.