TX-15 Odds: Pulido And De La Cruz Trade The Lead
Updated August 12, 2026 · 10 min read · by Sam Smith
Texas's 15th congressional district gave Donald Trump 58.5% of its vote in 2024. It is held by a Republican who has won it twice. And for most of the past week the TX-15 odds on Kalshi made the Democrat the favorite to win it in November — until this afternoon, when the board changed its mind again. This is not a thin market nobody is watching: roughly 193,000 contracts of combined open interest sat across the two sides when we pulled the board, and more than 90,000 traded on the Democratic contract alone in the single hour the price turned. So we handed the race to our model panel, showed each seat every verified fact about the district and not a single market price, and asked it to price the seat cold.
The Quick Answer
Kalshi now prices Republican incumbent Monica De La Cruz near 56.5¢ and Democrat Bobby Pulido near 43.5¢ — a reversal from earlier the same day, when Pulido was the 60¢ favorite. Our price-blind panel blends to a 54% chance the Republican holds the seat, leaving the Democrat about 46%. The number worth keeping is not today's price. It is that this board has changed favorites twice in eight days, and that the panel's figure, set without sight of any price, sits close to where it has just landed.
- Market Read: Republican hold, about 56.5¢ implied, as of 17:08 UTC on August 12
- Model Read: Republican hold, about 54%
- The Gap: roughly 2 points today — but it was about 14 points, pointing the other way, two hours earlier
- The Story In One Line: the price of this seat has given back the entire move it made on an August 4 poll and kept going, with no rating change, further public poll or reported campaign development since to mark it.
The full board, the one poll number that survived every model's skepticism, and the two models that reversed themselves are below.
Free: The Weekly PM Market Brief — the 8-model panel's graded record, the week's biggest market-vs-model gaps, and what's spiking next. One email, Sundays. Sign up in the box at the end of this article.
New to CFTC-regulated event contracts and how they settle? Start here.
The Market At A Glance
| Market | Which party wins the 2026 U.S. House race in Texas's 15th district |
| Venue | Kalshi, a CFTC-regulated event-contract exchange (18+; availability varies by state, as of August 2026) |
| Ticker | HOUSETX15-26 (paired contracts on the Democratic and Republican outcomes) |
| Settles | On the party of whoever is sworn in for the term beginning January 2027, not the named candidate. Because it pays on the party of the member actually seated, a third-party or independent winner would resolve both contracts NO, and a candidate who switched to the other major party before being sworn in would resolve that party's contract YES. |
| Market Read | De La Cruz (R) 56-57¢ bid/ask (56.5¢ mid), Pulido (D) 43-44¢ bid/ask (43.5¢ mid); last trade on the Democratic side 44¢ |
| Panel Read | 54% Republican hold, 46% Democratic flip |
| Prices As Of | August 12, 2026, 17:08 UTC |
These are model estimates, not predictions of fact and not financial advice. Kalshi event contracts trade on a CFTC-regulated exchange; you must be 18 or older and located in an eligible state to participate, and availability varies by state and can change. Prices on this page are read from Kalshi and refresh as the race develops; the "as of" line above carries the timestamp. Every number in this piece gets graded in public once the market settles, on the full graded scoreboard.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $177M traded
- 2027 Pro Football Champion · $59M traded
- 2028 U.S. Presidential Election winner? · $59M traded
- 2028 Republican presidential nominee · $56M traded
- Pro Baseball Champion · $52M traded
- More tech layoffs in 2026 than in 2025? · $31M traded
Every market above links to our full AI model verdict; browse them all on the OddsShopper prediction markets hub, and see how every settled call actually scored on the full graded scoreboard.
The Board
| Outcome | Candidate | Market | Panel |
|---|---|---|---|
| Republican Hold | Monica De La Cruz | 56.5¢ | 54% |
| Democratic Flip | Bobby Pulido | 43.5¢ | 46% |
Every Seat's Number
Each model priced both outcomes in a single pass, so the two numbers always sum to 100. The figures below are the Republican-hold probability after the revision round, with each model's opening number beside it.
| Model | First Round | After Revision | Move |
|---|---|---|---|
| Fable | 54% | 58% | +4 |
| ChatGPT | 55% | 57% | +2 |
| Sonnet | 65% | 56% | −9 |
| Kimi | 45% | 56% | +11 |
| Opus | 65% | 54% | −11 |
| GLM | 54% | 52% | −2 |
| DeepSeek | 60% | 47% | −13 |
| Panel Blend | 56.9% | 54.3% | −2.6 |
Every seat on this panel is graded against real market settlements — records to date: GLM 81% on 2,355 graded calls · Kimi 83% on 2,333 graded calls · DeepSeek 80% on 2,380 graded calls. Recomputed daily; the full scoreboard is public.
Seven of the panel's eight models returned a number on this board; one did not run, and its absence is disclosed rather than papered over. These are model estimates, not predictions of fact and not financial advice.
More live boards from the same panel: Texas Senate: Paxton vs. Talarico — Talarico 53.5¢, Paxton 47.5¢ · House control 2026 — Democratic control 85.5¢ · Midterms balance of power — Senate control dead even at 49.5¢ a side. Prices fetched August 12, 2026.
What Just Happened To This Price
This has been one of the most violently repriced boards of the cycle, and the eight days behind this page are why the number at the top of it carries a caveat.
On July 21 the Democratic contract traded at 40¢. On August 4, House Majority PAC released the Normington Petts survey fielded July 23-27 that put Pulido ahead 51-45. That day the contract opened at 61¢, traded as high as 77¢, and closed at 67¢. It then held a 62¢ to 68¢ range for a week.
Then it stopped behaving like a market digesting news. On the evening of August 11 the Democratic contract fell from 62¢ to 40¢ and climbed back to 62¢ inside about three hours, on well over 100,000 contracts, finishing the night at 66¢ as though nothing had happened. Around midday on August 12 it fell again, from 61¢ to 44¢ in a single hour on more than 90,000 contracts, and open interest in the Democratic contract more than doubled.
Nothing on the record moved with it. No handicapper has changed its rating since Sabato's Crystal Ball on July 30, and no new public poll has appeared since August 4.
Two more details belong on the page before any of these prices is read as a consensus. The first is depth. The quote is well supported at the top of the book and thin underneath: at the time of writing a single resting bid of about 27,000 contracts sat at 43¢ on the Democratic side, with the next meaningful bid roughly 1,000 contracts three cents lower. A quote can be firm and shallow at once, and this one is both.
The second is that another venue disagrees. Polymarket runs its own TX-15 market, and on the same afternoon it sat close to even, the Republican side near 51% and the Democratic side near 50%, on roughly $10,400 of total volume. That is a far smaller book than Kalshi's and carries correspondingly less information. But two exchanges pricing the same seat several points apart is a fair warning against treating either as settled.
One last piece of timing, offered as method rather than a boast. The panel's numbers on this page were generated at 15:51 UTC, while the Democratic contract was still trading at 60¢, and they put the Republican at 54%. About an hour later the market arrived at 56.5%. A single board is not a track record, and the panel's actual record is graded market by market in public rather than claimed here. But it is the reason these runs are done price-blind.
The Two Numbers This Race Turns On
Almost every disagreement on this board traces to one question, and it is worth understanding before anything else.
The first number is the presidential one. In 2024, Trump carried this district 58.5% to 40.7%. By that measure the seat is not competitive.
The second number is the down-ballot one. Inside Elections scores the redrawn district at R+2 by Baseline, a composite of down-ballot statewide races rather than the presidential result. By that measure the seat is nearly even.
A gap that large between a district's presidential vote and its down-ballot vote is rare, and here it is not a statistical artifact. The same voters gave Beto O'Rourke 55.4% in the 2018 Senate race and held Ted Cruz to 53.5% in 2024, five points behind Trump's share on the same ballot. De La Cruz herself ran about 1.4 points behind Trump's district margin. This electorate splits its ticket at a rate almost no other district in the country does.
So the whole race reduces to a single judgment: in a midterm with no presidential candidate on the ballot, which number is the better guide? Every model on the panel answered the same way, and it is the reason none of them treated this as a safe Republican seat.
Why The Panel Still Landed On The Republican
The models sided with the down-ballot number, then declined to take it literally. That second step is where the market and the panel part company.
Their reasoning was that Baseline is a composite weighted toward the district's Democratic past, and that past is partly gone. This is one of the sharpest realignments in the country: the district voted for Hillary Clinton by 13.5 points in 2016, for Trump by 2.5 in 2020, and for Trump by 18 in 2024. The question is whether that move belongs to Trump personally or has hardened into something a generic Republican inherits. The panel found evidence for the second reading in De La Cruz's own 2024 result, when she ran only about 1.4 points behind him.
"I do not take R+2 literally: it is a composite weighted toward a pre-realignment past, and De La Cruz running only ~1.4 points behind Trump's 2024 margin shows at least partial transfer to a non-Trump Republican." — Kimi
From a structural anchor around R+4 to R+6, most models then passed roughly 55% to 65% of the national environment through to this district rather than all of it, on the grounds that the Rio Grande Valley is precisely where Republican gains ran deepest and is therefore the region least likely to snap back in full.
The single most original observation on the board came from Fable, and it is an awkward fact for the bullish case:
"The Applecart internal is the disciplining datum both ways: its in-district generic ballot of R+3, taken in May 2026 with the national environment already at roughly D+6, implies the district's underlying lean is still solidly red even in a bad Republican year." — Fable
That is the strongest argument against the market's price, and it comes from a poll paid for by a group supporting the Democrat.
The Case For The Flip
The models did not dismiss it. Four separate facts kept the Democratic number in the mid-40s rather than the 30s.
Roughly 80% of this district is new to the incumbent. Texas redrew its congressional map in 2025, and while the new lines barely moved TX-15's partisan lean, they changed who lives in it. A University of Texas at San Antonio political scientist quoted in Ballotpedia's race preview put it at about eighty percent territory De La Cruz has not represented in her two terms. The name recognition, casework and local relationships that normally protect an incumbent reach about one voter in five here. The panel credited her roughly one point of incumbency advantage rather than the usual three to four.
Pulido is not a generic Democrat. He is a Latin Grammy-winning Tejano musician with 73% name identification in the district and a net favorability of +8, figures almost no first-time candidate posts. The panel found one poll-based number it trusted, and it is this: in the same May survey, Pulido led 48-46 while a generic Democrat trailed 46-49. A candidate running five net points ahead of his own party label is a real effect, and because both halves of that comparison come from one instrument with one house effect, the sponsor's bias largely cancels out.
The national environment is bad for the incumbent's party. Across an average of six public aggregators updated August 5, the generic congressional ballot sits at Democrats +6.7; across ten approval aggregators updated August 7, the president is at 37.8% approve against 59.3% disapprove. The president's party has lost House seats in 18 of the last 20 midterms. The two exceptions, 1998 and 2002, both required approval far above water. Neither condition is present.
The party committee is treating this as winnable. The DCCC added Pulido to its Red to Blue program in May, a designation it reserves for top-tier challengers, and CHC BOLD PAC, Blue Dog PAC, Latino Victory Fund and WelcomePAC have all endorsed.
The Case The Panel Discounted
Two arguments carried the Democratic price to 67¢ earlier this month, and both deserve a fair hearing even now that it has come back down.
The first is the statewide Latino picture. Sabato's Crystal Ball, moving the race from Likely Republican to Leans Republican on July 30, cited a late-June New York Times/Siena poll showing Democrats leading by about 30 points among Latinos in the Texas Senate race. Sabato tied it directly to this seat, noting TX-15 "is wedged in between" the two heavily Latino South Texas districts Democrats are defending. Every model on the panel named this as its single largest uncertainty, and every model discounted it for the same reason: a statewide Latino crosstab is weighted toward urban Latino populations that look nothing like the Valley.
The second is that Texas itself is close. On the same exchange, the Texas Senate race is priced with Democrat James Talarico at 53.5¢ against Republican Ken Paxton at 47.5¢, and the independent Times/Siena poll had that race tied 47-47. A market that believes a Democrat can carry Texas statewide is being coherent, not careless, when it prices a Democrat competitively in a district whose down-ballot lean is R+2.
The Case Against The Flip
Every rater is on the Republican side, and none has crossed to Toss-up. Decision Desk HQ (Lean R, July 14), Cook Political Report (Lean R, July 16) and Sabato's Crystal Ball (Lean R, July 30) have all published under the new map, and Sabato's move that day was the second step of a one-way drift toward the Democrat that began in November 2025. Inside Elections and Race to the WH also sit on the Republican side of the line.
Every district poll was paid for by one side. The most recent, fielded July 23-27 by Normington Petts for House Majority PAC and released August 4, put Pulido ahead 51-45. It is also the poll that moved this market, and the move has since round-tripped. But there is not a single independent public survey of this race: all three were commissioned by Democratic committees, and sponsored polls are released selectively, since a sponsor that polls and finds bad numbers usually publishes nothing. The models discounted the toplines by three to four points as a result, which turns that six-point lead into something closer to a tie.
The incumbent has the money. Through June 30, De La Cruz had raised $5.27 million to Pulido's $2.96 million and held $2.46 million in cash against his $1.16 million, a 2.1-to-1 advantage entering the fall. In a district where 80% of voters are new to the incumbent, and all of them are new to a first-time challenger, the ability to define the race first is worth more than usual.
Where The Panel Changed Its Mind
After the first round, every model read the others' reasoning anonymously and could revise. Two models crossed the line in opposite directions, which is unusual and informative.
DeepSeek made the largest move on the board, dropping the Republican from 60% to 47% and switching its favorite:
"A midterm with an unpopular president, a strong challenger, and a hollowed-out incumbency flips a historically elastic district just enough." — DeepSeek
Kimi moved almost as far the other way, from 45% to 56%, and its stated reason was a methodological one about the polling:
"I discounted the sponsored toplines ~3-4 points for house effect plus selective release, which converts Pulido +6/+2 into a coin-flip, not a lead." — Kimi
ChatGPT barely moved at all, from 55% to 57%, and explained why the debate did not shift it:
"The R+2 down-ballot baseline dominates Trump +18, but durable South Texas realignment and GOP cash/incumbency keep R modestly favored." — ChatGPT
The net effect was convergence. The panel's spread narrowed from 20 points to 11, and the blend moved about two and a half points toward the Democrat, from 56.9% to 54.3%. The revision round pulled the panel slightly toward the market without closing the gap.
What Would Change The Panel's Mind
Four things, each checkable and each with a direction.
- The First Independent Poll Of The District. Every model discounted the existing polling for sponsorship. A non-partisan survey would carry several times the weight of the three on the record, in whichever direction it lands.
- The October FEC Filings. The quarterly report lands October 15 and the pre-general report on October 22. If Pulido closes the 2.1-to-1 cash gap in the quarter that matters, the panel's "she can define the new turf first" argument weakens and the number moves toward the Democrat.
- A Rater Crossing To Toss-Up. No handicapper has. Cook, Sabato or Decision Desk HQ moving would push the panel toward the Democrat; a move back toward Likely R would push the other way.
- Any Valley-Specific Evidence On Latino Voting. This was the named uncertainty for all seven models. District-level or Rio Grande Valley crosstabs showing the statewide snapback reaching this region would move the number materially toward Pulido; evidence that the 2020-2024 realignment is holding locally would move it the other way.
Settlement Timeline
| Date | What Happens |
|---|---|
| October 15, 2026 | October quarterly FEC report: July-September fundraising for both campaigns |
| October 22, 2026 | Pre-general FEC report, covering through October 14 |
| November 3, 2026 | General election |
| January 3, 2027 | New Congress sworn in; the contract's outcome is determined by the party of the member seated |
| January 4, 2027 | Expected expiration of the Kalshi contract |
There is no settlement date on election night. The contract resolves YES on the side matching the party of the member seated, and it settles once the representative is sworn in, which is why the expected settlement date sits in January rather than November. This page is re-scored when the story moves; prices and panel figures are current as of August 12, 2026.
How To Read These Prices
A price of 56.5¢ is a probability, not a prediction. It says the market thinks this outcome happens a little under six times in ten, which means it also expects it to fail four times in ten. The two contracts here quote at 56.5¢ and 43.5¢ at the timestamp above, summing to 100 and sitting a single cent apart. That tightness is not a fixed feature of the board: within half an hour, as the price moved again, the quoted spread on both sides had widened to several cents. A market can look tight in a still moment and cost noticeably more to trade in a moving one.
The panel's number is built differently. Each model saw the same fetched data card and no price at all, then priced both outcomes in one pass so its two numbers could not contradict each other. That is the point of running it price-blind: a model shown the market price tends to cluster near it, and you learn nothing.
Kalshi contracts settle at $1 or $0, and event contracts carry fees and can be illiquid. This is a market with real open interest on both sides, but the panel and the price landing in the same place today is a description of two methods agreeing, not an instruction.
FAQ
Who is favored to win TX-15 in 2026? Both candidates have been favored inside the last eight days, which is what makes this board unusual. As of the afternoon of August 12, 2026, Kalshi prices Republican incumbent Monica De La Cruz near 56.5 cents and Democrat Bobby Pulido near 43.5 cents. Eight days earlier the same board traded Pulido as high as 77 cents. Our panel of AI models, none of which saw a price, blends to a 54% chance the Republican holds the seat after a revision round. The five professional handicappers who rate the race all have it Lean or Likely Republican, and none has moved it to Toss-up.
When is the TX-15 election and when does the Kalshi market settle? The general election is November 3, 2026, but the market does not settle that night. The contract pays on the party of whoever is sworn in for the term beginning in January 2027, so it closes early once the representative is seated on January 3, 2027, with expected expiration January 4. Two more FEC disclosures land before the vote: the October quarterly report on October 15, covering July through September, and the pre-general report on October 22, which closes its books on October 14.
Why is TX-15 competitive at all in a district Trump won by 18 points? Because the presidential number and the down-ballot number disagree more here than almost anywhere else. Trump carried the redrawn district 58.5% to 40.7% in 2024, but Inside Elections scores it R+2 on Baseline, a composite of down-ballot statewide races. The same voters gave Beto O'Rourke 55.4% in 2018 and held Ted Cruz to 53.5% in 2024, five points behind Trump on the same ballot. In a midterm with no presidential candidate on the ballot, which of those two records is the better guide is the single question this race turns on.
How reliable is the polling in the TX-15 race? Treat it carefully. All three public polls of this district were commissioned by Democratic groups, and there is no independent survey on the record. Sponsored polls are also released selectively, since a sponsor who finds bad numbers usually publishes nothing. The one figure that largely survives that problem is an internal comparison: in the same May survey, Pulido led 48-46 while a generic Democrat trailed 46-49, because a house effect biases both halves of one instrument in the same direction.
The Bottom Line
The interesting thing about this board is no longer the disagreement. It is the motion. In eight days the market has priced Bobby Pulido at 77¢ and at 40¢, changed favorites twice on a sustained basis with a third brief flip overnight on August 11, and arrived this afternoon within about two points of a panel that was never shown a price. Apart from the poll that started the run on August 4, none of that motion was accompanied by a rating change, another public poll, or a reported development in the race.
The underlying question has not changed either. This is a district whose presidential record says one thing and whose down-ballot record says nearly the opposite, in a national environment currently favoring Democrats, with an unusually well-known challenger facing an incumbent who has never represented most of these voters and holds a two-to-one cash advantage. The panel's answer, at 54-46, is that this is a near-tossup leaning slightly Republican, and that the honest read of the price is a snapshot rather than a verdict. That call gets checked against the result on November 3 and graded in public when the contract settles in January.
