Every few years Texas Democrats announce that the state is about to turn blue, and each and every time, since 1994, it has not. In 2018, Beto O'Rourke toured all 254 counties, broke fundraising records and still lost. Demographers have been promising a Hispanic electorate that votes like California's for two decades. The state kept voting red.

This year the prophecy looks less foolish. Polymarket puts Democrat James Talarico's chance of winning Texas's Senate seat at 66%. A former schoolteacher and seminary student who quotes the Gospels on the fly, Mr Talarico is the candidate Texas Democrats have long said they needed and never quite found: young, fluent in faith and hard to paint as a coastal liberal.

Yet the more interesting story is not really about him. It is about the man he is running against, and about what traders expect to happen everywhere else on the ballot.

Ken Paxton, the state's attorney-general, won the Republican nomination by humiliating John Cornyn, a four-term senator, with Donald Trump's blessing. Primary voters got what they wanted: a fighter, unbowed by an impeachment in 2023, of which he was acquitted, and by a divorce his wife sought on biblical grounds. General-election voters, including the suburban moderates who used to vote for Mr Cornyn, are less keen. The polls have drifted Mr Talarico's way since the summer, and Mr Paxton has begun to sound like a man who has noticed. 'I don't control message or what we're doing,' he told donors last month, in a recording obtained by Politico.

Republicans are trying to fix that with money, on a scale Texas has never seen; one super PAC spent more than $131m in 17 days. That may yet work. Texas polls have flattered Democrats before, and the Republican floor in the state is high.

But look down the ballot and the idea of a blue Texas falls apart. The same traders who favour Mr Talarico make Greg Abbott, the Republican governor, a 72% favourite for a fourth term, and expect Republicans to keep almost every other statewide office and the state Senate. Their most likely outcome for the two big races is a split: a Republican governor and a Democratic senator.

That is not what a realignment looks like. Realignments move whole tickets, because voters who change their minds about a party change them about all its candidates. What the markets describe is older and more Texan: voters willing to cross over for one candidate against one opponent, and to go straight back afterwards. In 2024 Tarrant County, around Fort Worth, voted for Mr Trump and for the Democratic Senate candidate on the same day. Hispanic voters who swung hard to Mr Trump that year have cooled on him since. Whether they have warmed to the Democratic Party, rather than to Mr Talarico, is another matter.

The 66% also deserves some scepticism. Traders bet seven times as much on the Republican primary as they have on the general election, and a thin market tells you mostly what its few traders think. Texas has disappointed people who bet on Democrats before.

So will Texas turn blue? It may well elect a blue senator, which would be a genuine shock. But that would be a verdict on Mr Paxton, not on Texas. Democrats have spent three decades waiting for demography to hand them the state. What they have been handed instead is an opponent.

Early voting begins on October 19th. Democrats are hoping for a blue wave. The prices describe something more modest: a purple patch, painted by a Republican.

Writing off Binyamin Netanyahu has lost money as a trading rule for seven years. Indicted in 2019, ousted in 2021, blamed for October 7th: each time the obituaries were filed, each time the markets agreed, and each time he was back in the chair or had never left it.

He does not need 61 of the Knesset's 120 seats. He needs nobody else to have them: until a new government is sworn in, the old prime minister stays.

Israel votes in sixteen days, and the market has Gadi Eisenkot, a former army chief whose Yashar party is level with Likud, as favourite to be sworn in next, at 43%. The polls say something narrower.

In the five published when party lists closed last month, neither bloc reached 61, and Eisenkot's gets there only with Arab parties most of his allies have ruled out. Traders half know it: a thin side market puts a second election before any government at 35%.

Our view is that the board is pricing polling day and not the months after it. In 2019 and 2020 it took three elections and thirteen months to swear anyone in, and the man sworn in was the caretaker. The one we would put it on: Netanyahu as the next sworn-in prime minister, 36% in a $44 million market that runs to the end of 2027, long enough for a rerun.

In an interview published on Tuesday, Samuel Alito said he is staying on the Supreme Court. Bizarrely, the market heard something else. The chances of him announcing a retirement by 31 December traded at 6% before the interview and 13% after it. After he had publicly stated he won’t go.

The justice, who is 76, told Bloomberg Law that he intends to “stick around through this term”, which runs to June, and that he decided so months ago.

Our analysis puts it around 4-5%.

The trade against him is the Ginsburg lesson: stay too long and the other side fills your seat.

Democrats are 63.5% to win the Senate, the new one sits on 3 January, and Amy Coney Barrett was confirmed inside a month, so a lame-duck exit would work. It would also mean a justice eating his words within weeks, and doing what none has done since William Douglas in November 1975, after a stroke: announce between October and December. A week of book-tour interviews gave traders headlines and no new information.

The odder price is next door. The same board has him announcing by 30 June 2027 at 39%, unchanged in a month in which Democrats' Senate price rose 11%. A man who will not go while Republicans can replace him is unlikely to go once Democrats can block the replacement.

What changes things: Democrats win the Senate and Alito finds the Breyer route, announcing now, leaving in June, with a successor confirmed by this Senate before it goes. Then fair value sits above 30%.

Traders have stopped worrying about GTA VI. Polymarket gives a third delay just 3.9%, down from 54.5% in January. For a game that has already slipped three times, has been leaked by hackers making demands and, with under six weeks to go, has still not said it has gone gold, that is far too cheap. Our analysis believes there could be an edge here. Wishful thinking over operational realities.

Start with the record. Rockstar has delayed GTA VI twice in public, and this spring Strauss Zelnick, Take-Two's boss, let slip a third, internal delay: the game is landing 18 months after its original date. Red Dead Redemption 2 slipped twice too. In January Bloomberg's Jason Schreier reported it was still not content complete. Since then Rockstar has said little, beyond opening pre-orders and putting a 26-minute preview on Netflix after a group calling itself CyberLeek posted clips from a playable build.

Next, the money, which everyone reads the wrong way. Take-Two's forecast of $8bn or more is for the year to March, not for 19 November. A slip of a few weeks, or even to February, still lands inside it. Cyberpunk 2077 shows how late such slips come: due on 19 November 2020, it had already gone gold when it was pushed back three weeks on 27 October. Take-Two reports its quarterly results in early November, the occasion it chose for last year's delay. And as recently as September, Kalshi gave about an 11% chance that the game would not be out by December. Even Mr Zelnick hedges: “We haven't sold one unit yet.”

A real $10,000 book on Polymarket - @thebasementview. Every position is live and public.

1. No on Bolsonaro, Eisenkot, Dems - 350 shares at 67%. Cost: $234.50 plus a $3.10 fee.

2. No on Bolsonaro, Netanyahu, Dems - 350 shares at 73%. Cost: $255.50 plus a $2.76 fee.

Rationale - Polymarket's Autumn Elections combo pays out only if Brazil, Israel and the US Senate all go one way. These two combos trade at a combined 61.5%. Polymarket's own, far deeper markets on each election (Bolsonaro 86%, Democrats 63.5%, Eisenkot or Netanyahu 79%) multiply to 43%. The parlay is overpriced, so we are selling it.

Expected return - Total outlay $495.86, fees included. If Bolsonaro, the Democrats and either Israeli candidate all win (43%), we lose $145.86. Otherwise (57%) we make $204.14. Expected profit: about $53, roughly 11%.

Risks - This loses about four times in ten. It only loses money on average if the true chance of all three is above 58%, which would need the elections to move almost in lockstep. The money may also be tied up until Israel swears in a prime minister, unless Lula or the Republicans win first.

Nothing to sell yet.

Know one person who is fed up with traditional news? Forward them this. Been forwarded this? The button gets you your own copy every Sunday.