Booming? A pattern? A record? Checking three of this week’s stats

Each item is a question to check, not an accusation. “Our arithmetic” marks sums we did ourselves.

(a) Are Saudi oil exports “booming”?

What was claimed

Rigzone ran a Bloomberg story on 28 October under the headline “Saudi Oil Exports Booming Despite Attacks” (Rigzone); Bloomberg’s own headline says exports “hold near wartime highs”. The story’s first line says the same, not “booming”: exports are holding near wartime highs, based on preliminary ship-tracking data (estimates built from tankers’ position signals).

What the data shows (fact)

The article’s own figures are October-to-date daily averages, in b/d (barrels per day):

• Vortexa: 6.5m b/d so far in October, against 6.6m b/d for the whole of September. Red Sea loadings rose; Persian Gulf loadings fell.
• Kpler: 5.4m b/d, lower than September.
• Tankertrackers.com: 6.99m b/d, described as prewar levels; no September figure, so no comparison.

So two of three trackers show October below September: Vortexa by about 1.5% (our arithmetic), Kpler by an amount the article doesn’t give. They also differ by about 1.6m b/d (5.4m to 6.99m), which the article puts down to tankers switching off position beacons.

And a part month is set against a full one. Bloomberg doesn’t say what date “so far” runs to, and the latest days are usually the least complete: Kpler, whose own count for 1–25 October was 5.52m b/d, told Asharq Business it expects the full-month figure to rise to about 5.8m as more cargoes are recorded (English summary: Egypt Oil & Gas). Official figures from JODI (a database of government-reported oil statistics) cover only August 2026 so far (JODI, 21 October). For context on the Gulf route, Kpler’s 30 October media briefing puts oil flows through the Strait of Hormuz, excluding Iranian cargoes, at a seven-day average of about 8.5m b/d (data to 28 October), about 49% of its prewar baseline of about 17.1m b/d (the March 2025–February 2026 average). Kpler treats its latest readings as a floor: they have typically ended up 1.5 to 2 times higher once confirmed.

Verdict: Missing context. The first line holds up: exports have stayed near wartime highs (on Vortexa’s figures, with a shift from Gulf ports to the Red Sea). “Booming” isn’t supported; the figures point to flat or slightly lower.

Why it matters. “Booming” implies more Saudi oil reaching the market; “roughly flat” tells a different supply story.

What would settle it. Full-month October tracker data (early November) and JODI figures for September–October (around late December).

(b) Has the S&P 500 jumped after every Warsh Fed meeting?

What was claimed

CNBC’s Fed live blog on 28 October, citing Nicholas Colas, a market commentator and co-founder of DataTrek Research, says the S&P 500 (an index of 500 large US companies) “has rallied at least 1% on the three post-meeting days of Warsh’s tenure”. It adds that since Kevin Warsh became Fed chair in May, stocks have generally fallen on Fed meeting days (when US rates are set).

What the data shows (fact)

FRED closes and the Fed’s calendar (changes are our arithmetic):

• 18 June 2026: +1.08%
• 30 July 2026: +1.66%
• 17 September 2026: +1.14%

The claim checks out. But on 28 October, this month’s meeting day, the index rose 0.28%, so the “falls on meeting day” half failed. The fourth test of the up-day half came on 29 October: the index fell 0.30%, to 7,720.39 (Reuters), so the streak ended at three.

For scale: from 2 January to 28 October 2026, the index rose 1% or more on 23 of 207 trading days, about 11% (our arithmetic). That is the base rate: how often something happens anyway. If each post-meeting day were an ordinary, independent trading day, the chance of three out of three would be roughly 1 in 729 (our arithmetic: 23/207 is exactly 1 in 9, and 1/9 × 1/9 × 1/9 = 1/729). That overstates the rarity: post-meeting days carry fresh news, and patterns spotted after the fact turn up more often than such odds suggest.

Verdict: Fair. Correct and striking, but three days are too few to call it a pattern.

Why it matters. Streaks like this get repeated as trading rules; three days said little about the next, as 29 October showed.

What would settle it. Many more meetings.

(c) Israel’s “most-voted for” election

Provisional: based on the Central Elections Committee’s full count, ahead of the official results due on 4 November. We will update this item then.

What was claimed

The Guardian’s live blog on 28 October called it “the most-voted for election in Israel’s political history”, with “nearly 80% of IDF soldiers voting”.

What the data shows (fact)

• Votes cast: about 5.72 million (5,716,665 in the full count released overnight on 29–30 October by Israel’s Central Elections Committee (CEC), via ynet). The previous record was 4,794,593 in 2022 (Israel Democracy Institute, CEC final figures), so this is about 19% more (our arithmetic).
• Turnout (the share of eligible voters who voted): 77.9% in the full count, per the CEC (Israel National News, 30 October), of about 7.33 million eligible voters (CEC, via Al Jazeera, 27 October). That is the highest since 1999 (78.7%), not a record rate: turnout ranged from 77.4% to 83% in 1955–1999 (IDI), and was above 77.9% in all but two of those elections (1981 and 1992) (Wikipedia, secondary); 1949 reached 86.9% (Wikipedia, secondary source).
• Soldiers: 79.9% is the Israel Defense Forces’ own figure (Times of Israel); the army publishes no headcounts and the CEC has not confirmed it.

The electorate grew about 8% since 2022 (our arithmetic), which makes a record count easier than a record rate.

These are the CEC’s full-count figures, which include absentee (“double-envelope”) ballots, cast mainly by soldiers, prisoners and diplomats and counted only after checks that no one voted twice (Times of Israel). The CEC says they are still subject to minor corrections; official results go to the President on 4 November. Turnout moved from 77.7% at poll close to 77.9%.

Verdict: Missing context. Both figures hold up as worded, but “most-voted for” is a record number of votes, not a record share: turnout was the highest since 1999, on a bigger electorate.

Why it matters. “More votes” and “a higher share voting” differ; only the second shows whether engagement rose.

What would settle it. The official results, due on 4 November.

How to read a stat like this

• Part month or full month? Early figures for a month in progress are incomplete; expect revisions.
• How many cases? Three true examples can still tell you little. Ask how often it happens anyway.
• Count or rate? A bigger total can just mean a bigger population. Check the share too.

What to watch next: full-month October tanker data (early November) and JODI for September–October; Israel’s official results (4 November).

In summary

• Saudi oil (Missing context): near wartime highs, but October looks flat or lower on part-month data.
• ”Warsh bounce” (Fair): three 1%+ days after three meetings, then a 0.30% fall after the fourth.
• Israel (Missing context, provisional): record votes; 77.9% turnout is a high since 1999, not a record. Full count; official results 4 November.

This is not financial advice.