The Reflex Map

The finding

News is supposed to move markets, but most of it does not, and which markets move is not the story you would guess. We paired more than a hundred thousand news items to the prediction markets that carried them and logged what each market did next. News about crypto, soccer or geopolitics repriced a market five cents or more about one time in ten. News about economy, politics or tech did so about one time in thirty to one in forty-five. Same pipeline, same threshold, so a high-reactivity headline is roughly three to five times likelier to move a market than a low-reactivity one.

That between-group gap is the finding, and it is the only comparison here that survives scrutiny. The order inside a group is not: crypto’s lead is a join artifact that falls to third on the wider join, and crypto, geopolitics, politics and tech each take four or more rank positions across 28 specifications. Read the two tiers, not the six bars.

How often a market moves is not how often the news moved it.

The Reflex Map · Vera Research

The method

The study spans 567,033 news-to-market reaction measurements across 3,707 markets and 123,424 news items, from April 29 to July 13, 2026. A reaction is one news item paired to one market, counted as a move when the market’s peak absolute odds change reached five cents inside a two-hour window. Every rate is a frequency, the share of reactions that moved, never an accuracy, and direction is not measured.

The rates are float-corrected: a true five-cent move was often stored just under five cents and excluded, holding every rate down. Recomputing on the rounded magnitude lifts each by one to four tenths of a point and leaves every whole-percent claim intact.

Reflex is not response

There is a trap in the table on the first page, and it is worth naming because a careless reading falls straight into it. A market that reprices often after news is not the same as a market that reprices because of news. Crypto is the clear case. The median crypto market in our baseline panel moved five cents in about ten percent of its windows that carried no matched news at all. Their high reflex is mostly their own ambient volatility, not a response to the headline that happened to run alongside it.

To separate the two, judge each market against its own quiet baseline: how much more often does it move after news than it moves anyway? On that measure the map reorders. Soccer is the most news-reactive subject we track, first in 27 of 28 specifications and never below second. Economy, a low-reactivity subject on the table above, is second, in the top three in 27 of 28. Economy pairs a low move-rate with a high response: a market that sits quiet until information arrives, then moves on it. That is the opposite of crypto.

What stands, and what does not

Three things follow. First, nothing in the published table is withdrawn. Every rate in it is a true conditional frequency: given a news match, this subject moved five cents that share of the time. Second, the honest correction is to the interpretation, not the number. The ordering conflates how much a market moves anyway with how much it moves because of news, and it should carry that companion sentence.

Third, once ambient movement is separated out, crypto, tech, politics and geopolitics cannot be ordered against each other at all. Each occupies at least four different rank positions across the specifications we tried, and the earlier reading that crypto or tech showed no news effect was itself an artifact of a mismatched baseline: matched on price level, both carry a real, modest lift. Soccer and economy stand apart. The other four are a tie the data cannot break.

The takeaway

1. News repriced crypto, soccer and geopolitics markets five cents or more about 1 in 10 times; economy, politics and tech about 1 in 30 to 1 in 45. Float-corrected: crypto 11.31%, soccer 11.16%, geopolitics 10.14%, economy 3.49%, politics 3.22%, tech 2.25%.
2. The gap between the two tiers, roughly 3 to 5 times, is the only comparison the data supports. The order inside a tier is not stable: crypto’s lead is an artifact of which markets are joined and falls to third on the wider join, behind soccer and geopolitics, and across 28 specifications crypto, geopolitics, politics and tech each take at least four different rank positions.
3. How often a market moves is not how often the news moved it. The median crypto market in the baseline panel moved five cents in about 10% of its windows with no matched news at all, so most of that movement is ambient, not a response.
4. Judged against each market’s own quiet baseline, soccer is the most news-reactive subject (1st in 27 of 28 specifications) and economy is second (top-3 in 27 of 28). Economy is quiet until news arrives, then moves on it.
5. On that news-response measure, crypto, tech, politics and geopolitics cannot be ordered against each other. Nothing in the published table is withdrawn; the correction is to its interpretation.

The published table is a true record of how often each subject’s markets moved on news. This note adds the companion sentence it needs: a naive reading of that order conflates how much a market moves anyway with how much it moves because of news. Separate the two and soccer and economy lead, while the middle of the table cannot be ranked.

Method and data

Built on the Vera dataset: 567,033 news-to-market reaction measurements across 3,707 markets and 123,424 news items, April 29 to July 13, 2026. A reaction counts once per news-item-and-market pair; a move is a peak absolute odds change of 5 cents or more within a two-hour window. Rates are float-corrected. The own-baseline (news-response) figures come from an adversarial review across 28 specifications.

Odds shown are each market’s own price at the time Vera flagged the story, stated as fact, not advice. Nothing here is a recommendation to buy, sell, or trade any market or asset. Vera and Crypto Briefing are not registered investment advisers.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.