Kalshi Election Predictions Scorecard: Florida's Misses

Updated September 3, 2026 with the certified count and the graded row. First published August 11, 2026 with the rules below, a week before the vote.

Anyone can quote election odds the week before the vote. The only version of that sentence that means anything is the one written down before the result, left untouched, and graded after. This Kalshi election predictions scorecard is the written-down version: a running record of what Kalshi's election markets said, frozen before each race was decided, then graded against certified results. Row 1 is the August 18, 2026 Florida Republican governor primary, and the name inside it that accounted for about 63% of every contract ever traded on the race's Kalshi winner board: James Fishback. The rules were published on this page on August 11, a week before the vote. The freeze happened on August 18. The state certified the count on August 27. Row 1 is now graded, and the interesting part is not that the market got the winner right. It is which contract it missed, and by how little.

The Quick Answer

On August 11, 2026, Kalshi priced James Fishback at 1.5¢ to win the Florida Republican governor primary and put the market's median for his vote share near 13%. By our freeze at 6:57 PM ET on August 18, the winner contract had slid to 0.5¢ and the median for his share had moved down to just under 10%. The certified count landed him at 10.5%. Graded under the rules published before the vote, the winner market went 9 for 9 and the 15 vote-share contracts went 14 for 15. The one miss is the deepest interior book on the whole ladder. The frozen prices, the certified numbers, and the arithmetic behind every grade are below.

Why This Page Exists

We already grade ourselves on the model verdict scoreboard, which tracks every prediction our AI panel has logged against how markets actually settled. This prediction market scorecard is its sister: it grades the market itself, because Kalshi's prices get quoted as forecasts everywhere, including in our daily Kalshi coverage hub, and Kalshi accuracy is only a checkable claim if the prices are written down at a fixed moment and scored against what certified. One fact about this specific market made a one-point grade possible at all: until August 10, 2026, Kalshi's Fishback ladder had no contracts between 5% and 10% or between 10% and 15%. The eight rungs the exchange added that day, the day before we committed the rules, are what let the certified share be scored at one-point resolution instead of five. The rules were committed on August 11, 2026, while the result was unknown to everyone, including the market, and they are reproduced below as published, with the only edits being two plain-English glosses on trading terms and one tie-break clause added today and disclosed where it sits.

What The Market Said A Week Out

The two tables in this section are the August 11, 2026 snapshot, as of 5:51 PM ET that day and captured from Kalshi's API, seven days before the primary. They are not the graded prices; the graded prices are the August 18 freeze further down. They are here because the week between them is part of the story. Kalshi's nominee event lists one yes/no contract per listed name; whichever listed name matched the certified Republican nominee for Florida governor settled YES.

Listed Name	Bid	Ask	Last Trade	Implied Probability (Mid)	Contracts Traded (Lifetime)
Byron Donalds	98.5¢	98.7¢	98.3¢	98.6%	3,385,767
James Fishback	1.4¢	1.5¢	1.5¢	1.5%	14,982,362
Jay Collins	0.2¢	0.4¢	0.3¢	0.3%	3,229,423
Paul Renner	0¢	0.1¢	0.1¢	<0.1%	1,280,924
Casey DeSantis	0¢	0.1¢	0.1¢	<0.1%	347,309
Charles Burkett	0¢	0.1¢	0.1¢	<0.1%	291,757
Matt Gaetz	0¢	0.1¢	0.1¢	<0.1%	124,607
Wilton Simpson	0¢	0.1¢	0.1¢	<0.1%	100,642
Jimmy Patronis	0¢	0.1¢	0.1¢	<0.1%	10,183

One caution before reading that table as a ballot: it is not one. Eleven Republicans qualified for the August 18 primary, and Kalshi's list matched only four of them: Donalds, Fishback, Collins, and Renner. The other five listed names, Casey DeSantis and Matt Gaetz among them, never qualified for this primary, so their contracts were always going to grade NO, and the seven qualified candidates without a contract sat outside the market entirely. The scorecard grades the listed contracts, nothing more.

The row worth staring at is the second one. A candidate the market priced at a penny and a half accounted for roughly 15 million of the event's 23.8 million lifetime traded contracts, about 63%, more than four times the volume on the favorite. One honest caveat before that number does too much work: contract count flatters cheap contracts, because a dollar of trading mints far more penny contracts than 98¢ ones, so counts measure argument, not dollars at risk. The concentration was still striking, and it did not fade. Hold that 63% figure; it comes back at the freeze. Our six-model panel's read on this same primary is a separate page, graded by the same result, and the most-traded-candidate page tracks the volume story on its own.

The Fishback Ladder A Week Out

Because a 1.5¢ winner contract was almost certain to settle NO, the market that actually measured Fishback's night was Kalshi's James Fishback vote percent event, what traders call a vote-share ladder: 15 separate yes/no contracts, each asking whether he receives at least X% of the certified primary vote. New to how these contracts price and settle? The mechanics are covered in our Kalshi guide and in how election prediction markets work. Here is the full ladder from the same 5:51 PM ET August 11, 2026 snapshot, quoted as bid and ask, meaning the best price a buyer was offering and the best price a seller was asking:

Rung	Bid	Ask	Last Trade	Contracts Traded (Lifetime)	Listed Since
At Least 1%	99.9¢	100¢	99.9¢	919,735	April 1, 2026
At Least 3%	98.2¢	99.8¢	99.4¢	157,515	April 1, 2026
At Least 5%	94.2¢	94.8¢	94.0¢	104,716	April 1, 2026
At Least 6%	92.0¢	93.0¢	92.0¢	11,972	August 10, 2026
At Least 7%	87.0¢	91.0¢	87.0¢	8,512	August 10, 2026
At Least 8%	80.0¢	87.0¢	79.0¢	602	August 10, 2026
At Least 9%	75.0¢	78.0¢	74.0¢	3,713	August 10, 2026
At Least 10%	63.0¢	65.0¢	63.0¢	311,461	April 1, 2026
At Least 11%	58.0¢	64.0¢	65.0¢	6,289	August 10, 2026
At Least 12%	56.0¢	61.0¢	56.0¢	1,806	August 10, 2026
At Least 13%	52.0¢	60.0¢	45.0¢	1,985	August 10, 2026
At Least 14%	31.0¢	34.0¢	33.0¢	713	August 10, 2026
At Least 15%	30.0¢	32.0¢	31.0¢	173,996	July 9, 2026
At Least 20%	13.0¢	14.0¢	13.0¢	700,034	April 1, 2026
At Least 30%	2.8¢	3.3¢	3.4¢	682,696	April 1, 2026

Eight of the 15 rungs, 6% through 9% and 11% through 14%, were listed by Kalshi on August 10, 2026, the day before this snapshot; the original rungs date to April 1 and the 15% rung to July 9. So the interior books were not thin because interest died. They were one day old. And the expansion itself was information: the exchange built its new rungs exactly where the action had concentrated, in the 6-14% range. (Last prints go stale in thin books, which is why a couple of last-trade cells sit out of order; the bid and ask are the live numbers.)

Read from the liquid rungs, the 10% contract, with over 300,000 contracts traded, sat near 64¢ at the mid, a solid favorite to reach double digits; the 15% rung, with 174,000 traded, quoted around 31¢, a real underdog outcome. The thin interior rungs flipped from favorite to underdog between 13% (56¢ mid, on a wide 52/60 quote) and 14% (32.5¢ mid), which is why we described the market's median as near 13% a week out, with the liquid books putting the flip anywhere from 13% to 15% and the day-old interior quotes pinning it to the bottom of that range. That number is the setup for everything that follows. Keep it next to the certified result when you get there.

The Grading Rules, Frozen Before The Vote

Here is the pre-commitment as it was published on August 11, 2026, stated as boringly as we could manage.

1. What gets frozen. Every contract on the winner market (ticker KXGOVFLNOMR-26, nine listed names in the August 11 snapshot) and every rung of the Fishback vote percent ladder (ticker KXVOTEPRIMARY-GOVFLNOMR26JFIS, 15 rungs in the same snapshot), each recorded as best bid, best ask, and last trade. If Kalshi lists or delists contracts before the freeze, whatever is listed at the freeze moment is what freezes.

2. When. One final snapshot at 6:59 PM ET on Tuesday, August 18, 2026, one minute before the first Florida polls close at 7:00 PM ET. Panhandle counties on Central time close an hour later, so the snapshot precedes every closed poll and every reported vote. Nothing frozen after that moment counts.

3. What grades it. Certified official results as published by the Florida Division of Elections. Kalshi's rulebook for the vote-percent ladder settles the same way; it reads certified or official results only, never preliminary ones. If Kalshi settles the winner market earlier off the projected result, that changes nothing here: our grade waits for certification either way, and election-night preliminary counts grade nothing.

4. How each contract grades. The listed contract matching the certified nominee grades YES; every other listed contract grades NO, including the names that never qualified for the ballot. Qualified candidates Kalshi never listed sit outside the scorecard. Each ladder rung grades YES if Fishback's certified share of the valid primary vote is at least that rung's number, endpoints inclusive, matching the ladder's own written rules. No judgment calls, no partial credit.

The one-sentence version: everything freezes at 6:59 PM ET on August 18, 2026, certified results grade every contract mechanically, and a miss is recorded exactly like a hit.

5. How the market grades. The graded probability for every contract is the midpoint between its frozen bid and ask, the best buying price and the best selling price on the book; bid, ask, and last are all recorded, but the midpoint is the number that gets scored. Three lines then go in the scorecard row, and all three are mechanical. First, sides: a frozen midpoint above 50¢ puts the market on the YES side of that contract, below 50¢ on the NO side, and any contract that settles against its side is recorded as a miss, printed at the same size as a hit. Second, the Fishback number: his certified share, set against the frozen median (13-14% on the August 11 quotes) and the liquid-book band of 10-15%. Third, two Brier scores, reported separately: one across the winner market's contracts and one across the ladder's rungs. The Brier score is the standard accuracy measure where 0.250 is a permanent coin flip and lower is better; the rungs restate a single quantity 15 ways, so they never blend into the winner market's number. Two fallbacks, committed now: a contract with no bid grades at half its ask, and every midpoint clamps to the 0.1-99.9¢ range.

6. When Kalshi misses, the row says so. A scorecard that only remembers wins is marketing. If Fishback clears 20%, then today's 13¢, or whatever the 20% rung freezes at on August 18, stays on this page next to the result. If he lands at 6%, the frozen price on the 10% rung stays too. The frozen numbers are never restated, reweighted, or quietly improved; every update appends.

One disclosure against rule 2, because the rules above are the contract: the final snapshot ran at 6:57 PM ET, two minutes earlier than the 6:59 PM ET the rule named. Every poll in Florida was still open, so the freeze still precedes every closed poll and every reported vote, which was the point of the rule. But it is not the minute we wrote down, and a receipt page that rounded that off would be the kind of receipt page nobody should trust. One clause is also added to rule 5 today, disclosed here rather than slipped in: a midpoint of exactly 50¢ is recorded as no side and scored for Brier only. It did not bind in Row 1.

The Freeze: 6:57 PM ET, August 18, 2026

Here is what froze. Both markets, as of 6:57 PM ET on August 18, 2026 and captured from Kalshi's API, before the first Florida polls closed at 7:00 PM ET. The freeze tables report Kalshi's fractional-contract count and the August 11 tables report whole contracts, so read the levels with that in mind; the 63% share below is a ratio within each snapshot, so it compares cleanly either way.

Listed Name	Frozen Bid	Frozen Ask	Frozen Mid	Contracts Traded (Lifetime)	Settled
Byron Donalds	98.9¢	99.4¢	99.15¢	9,490,083	YES
James Fishback	0.4¢	0.6¢	0.5¢	41,029,798	NO
Jay Collins	0.3¢	0.4¢	0.35¢	9,910,627	NO
Paul Renner	0¢	0.1¢	0.1¢	2,973,602	NO
Charles Burkett	0¢	0.1¢	0.1¢	941,098	NO
Casey DeSantis	0¢	0.1¢	0.1¢	557,975	NO
Matt Gaetz	0¢	0.1¢	0.1¢	148,918	NO
Wilton Simpson	0¢	0.1¢	0.1¢	128,091	NO
Jimmy Patronis	0¢	0.1¢	0.1¢	35,354	NO

The no-bid contracts show 0.1¢ because rule 5 says a contract with no bid grades at half its ask, and half of 0.1¢ clamps up to the 0.1¢ floor. Now the callback. At the freeze, the winner market had traded about 65.2 million contracts over its lifetime, and Fishback's contract accounted for 41.0 million of them: 62.9%, essentially the same 63% share he held a week earlier, at a price that had fallen from 1.5¢ to 0.5¢. The crowd's attention never moved. The price did.

Rung	Frozen Bid	Frozen Ask	Frozen Mid	Contracts Traded (Lifetime)
At Least 1%	99.7¢	99.8¢	99.75¢	1,384,641
At Least 3%	98.8¢	99.1¢	98.95¢	412,756
At Least 5%	93.6¢	94.0¢	93.8¢	259,192
At Least 6%	86.0¢	88.0¢	87.0¢	92,004
At Least 7%	70.0¢	78.0¢	74.0¢	76,039
At Least 8%	64.0¢	69.0¢	66.5¢	48,642
At Least 9%	54.0¢	62.0¢	58.0¢	75,439
At Least 10%	44.0¢	49.0¢	46.5¢	531,076
At Least 11%	41.0¢	48.0¢	44.5¢	82,054
At Least 12%	30.0¢	36.0¢	33.0¢	78,184
At Least 13%	27.0¢	32.0¢	29.5¢	50,671
At Least 14%	22.0¢	25.0¢	23.5¢	50,884
At Least 15%	17.0¢	19.0¢	18.0¢	463,515
At Least 20%	8.5¢	8.9¢	8.7¢	1,597,967
At Least 30%	1.7¢	2.5¢	2.1¢	1,864,846

Compare this ladder to the one a week earlier and the whole closing week is visible in one column. The 13% rung went from 56¢ to 29.5¢. The 12% rung went from 58.5¢ to 33¢. The 10% rung, the deepest interior book with 531,000 contracts traded, went from 64¢ to 46.5¢, which means the market crossed from favorite to underdog on double digits in the final week. Reading the frozen mids straight down, the ladder flips below 50¢ between the 9% rung (58¢) and the 10% rung (46.5¢), and a straight-line interpolation between those two puts the market's closing median at about 9.7%. A week out, the median was near 13%. The market moved a little over three points toward the low end in seven days, and then, as the next section shows, it moved slightly too far.

The Certified Result

The Florida Elections Canvassing Commission certified the August 18 primary on Thursday, August 27, 2026, and the Division of Elections publishes the official count on Florida Election Watch. Byron Donalds is the Republican nominee for governor.

Candidate	Certified Votes	Certified Share
Byron Donalds	811,131	47.79%
Jay Collins	427,014	25.16%
James Fishback	177,809	10.48%
Paul Renner	145,063	8.55%
Seven Other Qualified Candidates (Combined)	136,313	8.03%
Total Votes	1,697,330	100%

Fishback's exact certified share is 177,809 divided by 1,697,330, or 10.476%. Under the ladder's own rules, which read the certified percentage against each rung's threshold with both endpoints inclusive, that makes every rung from 1% through 10% settle YES and every rung from 11% up settle NO. Kalshi's settlement matches. Per the settlement timestamps in the market records on Kalshi's public API, which you can check against the race's market page, the exchange finalized the winner market at 8:12 PM ET on August 18 off the projected result, exactly the early settlement rule 3 anticipated, and finalized the ladder at 9:40 PM ET on August 26, with the 10% rung YES and the 11% rung NO. That is the evening before the state commission met; Kalshi has not published its reasoning for the timing, and we are not going to guess at it. Our row waited the extra day for the commission, as rule 3 required, and the certified count and Kalshi's settlement agree on every contract.

Row 1, Graded

The winner market first, because it is the easy half. Nine contracts, nine sides, nine hits: Donalds at a 99.15¢ frozen mid settled YES, and the other eight, none quoted above 0.5¢, settled NO. Its Brier score is 0.000013, which rounds to 0.0000 at four places. It is the least informative number on this page. A nominee market where the favorite closes at 99¢ is supposed to look like this, and a single row proves nothing about calibration either way, a point the accuracy explainer makes at length.

The ladder is where the grade lives:

Rung	Frozen Mid	Market Side	Settled	Graded	Brier Term
At Least 1%	99.75¢	YES	YES	Hit	0.0000
At Least 3%	98.95¢	YES	YES	Hit	0.0001
At Least 5%	93.8¢	YES	YES	Hit	0.0038
At Least 6%	87.0¢	YES	YES	Hit	0.0169
At Least 7%	74.0¢	YES	YES	Hit	0.0676
At Least 8%	66.5¢	YES	YES	Hit	0.1122
At Least 9%	58.0¢	YES	YES	Hit	0.1764
At Least 10%	46.5¢	NO	YES	Miss	0.2862
At Least 11%	44.5¢	NO	NO	Hit	0.1980
At Least 12%	33.0¢	NO	NO	Hit	0.1089
At Least 13%	29.5¢	NO	NO	Hit	0.0870
At Least 14%	23.5¢	NO	NO	Hit	0.0552
At Least 15%	18.0¢	NO	NO	Hit	0.0324
At Least 20%	8.7¢	NO	NO	Hit	0.0076
At Least 30%	2.1¢	NO	NO	Hit	0.0004

Fourteen hits, one miss, and a ladder Brier score of 0.077. The miss is the eighth row, and it is the one I keep coming back to, because of where it sits. The 10% rung was not one of the thin, day-old interior books we spent the August 11 section warning about. It was the deepest interior contract on the ladder, 531,000 contracts traded, and at the freeze it quoted 44¢ bid against 49¢ ask. The market was a 46.5¢ underdog on Fishback reaching double digits, and he reached them by 0.476 of a point. The whole miss is that: not a longshot landing, not a thin book printing a stale number, but the most-traded interior rung leaning the wrong way by three and a half cents on the eve of the vote.

The rung right below it tells you how close the call was. The 9% contract froze at 58¢, a YES side, and hit. The 10% contract froze at 46.5¢, a NO side, and missed. Together they say the market had Fishback's share pinned to a one-point window whose even-money crossing, 9.70%, sat about eight-tenths of a point below where he landed, just outside the top of that window. Under rule 5 there is no partial credit for that, and there should not be: the row records a miss at the same size as the 14 hits. But it is worth knowing what kind of miss it was, and that is what a Brier score is for. The 0.2862 term on the 10% rung is worse than a coin flip's 0.25, while the ladder as a whole scores 0.077, less than a third of a coin flip. Both numbers are true at the same time.

A Worked Example: The Miss, Line By Line

To make the arithmetic reproducible, here is the 10% rung from freeze to grade. The frozen quote was 44.0¢ bid, 49.0¢ ask. Rule 5 grades the midpoint: (44.0 + 49.0) divided by 2 is 46.5¢, or a probability of 0.465. That is below 50¢, so the market's side is NO. The certified share is 10.476%, which is at least 10, so the contract settles YES. Side NO against settlement YES is a miss. The Brier term is the squared distance between the probability and the outcome: (0.465 minus 1) squared, which is 0.2862. Sum that term with the 14 others in the table, divide by 15, and the ladder's score is 0.077.

Now run the same arithmetic on the August 11 quotes, purely for context, since the rules grade only the freeze. A week out, the 10% rung's 64¢ mid would have been a hit, but the 11% rung (61¢ mid), the 12% rung (58.5¢), and the 13% rung (56¢) were all quoted above 50¢ and all settled NO: three misses instead of one, and a ladder Brier score of about 0.099. Seven days of trading moved the market from three misses on the high side to one miss on the low side and cut its Brier score by roughly a fifth. The closing line was better than the week-out line. It was still not right, and the row says so.

The frozen ladder also shows exactly how the miss happened, if you read it as a distribution. The gap between adjacent rungs is the market's price on Fishback landing in that one-point bucket: 58¢ minus 46.5¢ put 11.5% on a 9-to-10 finish, 44.5¢ minus 33¢ put 11.5% on 11-to-12, and 46.5¢ minus 44.5¢ put just 2% on 10-to-11, the bucket he actually landed in. The market did not merely lean wrong; it priced the exact outcome as nearly impossible, and the shape says why: the 11% rung, with 82,054 contracts traded against 531,076 on the 10% rung, was quoted almost on top of its deeper neighbor. A thin rung that fails to leave room between itself and the liquid one next door is the kind of artifact a frozen bid and ask lets you see, and a last-trade table would hide.

What Sat Outside The Row

Rule 1 named two markets, and the row grades two markets. But the same 6:57 PM ET freeze file captured two more Florida boards, and leaving out misses that sat in the same snapshot would be the curve-grading this page promised not to do. So here they are, outside the row and outside its Brier scores, graded under rule 5 exactly as if they were in it.

Kalshi's Byron Donalds vote percent ladder had 11 rungs at the freeze. Graded against his certified 47.79%, it went 8 for 11:

Donalds Rung	Frozen Bid	Frozen Ask	Frozen Mid	Market Side	Settled	Graded
At Least 45%	94.0¢	96.0¢	95.0¢	YES	YES	Hit
At Least 50%	83.0¢	87.0¢	85.0¢	YES	NO	Miss
At Least 53%	75.0¢	82.0¢	78.5¢	YES	NO	Miss
At Least 55%	64.0¢	71.0¢	67.5¢	YES	NO	Miss
At Least 58%	43.0¢	49.0¢	46.0¢	NO	NO	Hit
At Least 60%	38.0¢	42.0¢	40.0¢	NO	NO	Hit
At Least 63%	19.0¢	26.0¢	22.5¢	NO	NO	Hit
At Least 65%	17.0¢	20.0¢	18.5¢	NO	NO	Hit
At Least 70%	3.0¢	18.0¢	10.5¢	NO	NO	Hit
At Least 75%	1.0¢	5.0¢	3.0¢	NO	NO	Hit
At Least 80%	0¢	3.0¢	1.5¢	NO	NO	Hit

Three misses, and all three were favorites: an 85¢ contract, a 78.5¢ contract, and a 67.5¢ contract on Donalds clearing 50%, 53%, and 55%, when he certified at 47.79%. Reading the frozen mids straight down, the ladder crossed 50¢ between the 55% rung (67.5¢) and the 58% rung (46¢), which interpolates to an implied median of about 57.4%. The market had the winner's share roughly 9.6 points too high, in the same snapshot, on the same race where it had the third-place share within half a point. That is a bigger measurement error than anything on the Fishback ladder, and if we had named the Donalds ladder in rule 1 it would be in the row. We did not, so it is here instead, in the same font. Kalshi's second-place market, the last board in the file, froze Jay Collins at 78¢ bid and 79¢ ask and Fishback at 21¢ bid and 22¢ ask; Collins finished second, a hit. The lesson for the next row is written into the rules below: name every market on the race at freeze time, not just the two the story is about.

How The Scorecard Grows

Every 2026 primary our prediction-market coverage takes a position on gets a row under these same terms: freeze at the last pre-close snapshot, grade on certification, and, starting with Row 2, freeze every Kalshi market listed on the race rather than the two the story happens to be about. The queue, honestly stated as of September 3, 2026: exactly one race has a pre-close freeze file of this kind on disk, and it is this one. We covered the September 1, 2026 Massachusetts primaries, but no pre-close freeze of those boards was taken, so they get no row; a row reconstructed after the result from remembered prices is not a receipt. The races that resolve on November 3, 2026, including the Senate and governor markets our 2026 midterm election odds board prices week over week, are the next freezes on the calendar, and their rows grade when those results certify. The cumulative record carries its own falsifiable claim: if these markets are well calibrated, frozen favorites priced at 90¢ or better should win at least nine times in ten, and 9¢ longshots should land about once in 11. Row 1 cannot test that on its own. It contributed one 99¢ favorite that won and, outside the row, three Donalds-ladder favorites between 67.5¢ and 85¢ that lost, which is a sample of four. By the time the November rows grade, there will be enough of them to start checking.

The Scorecard

#	Race	The Market's Line (Frozen)	Certified Result	Grade
1	Florida Republican governor primary, August 18, 2026	Frozen 6:57 PM ET, August 18, 2026: Donalds 99.15¢ mid to win; Fishback winner contract 0.5¢; Fishback share median about 9.7% (10% rung 46.5¢, 9% rung 58¢)	Certified August 27, 2026: Donalds nominee at 47.8%; Fishback 10.5% (177,809 of 1,697,330)	Winner market 9/9 (5 of the 9 listed names never qualified for the ballot, so 4 live contracts), Brier 0.0000. Ladder 14/15, Brier 0.077. Miss: the 10% rung, 46.5¢ NO side, settled YES. Outside the row, same freeze: Donalds ladder 8/11

One row, graded, with its miss in the same cell as its hits. The August 11 prices that used to sit in this table have moved up the page where they belong, as the week-out snapshot; the frozen set is what grades, and the frozen set is what stays.

The Record, For Completeness

One documented dispute touched this race's ballot, and it belongs on the page because it bears on what the ladder was measuring. In June 2026, Lt. Gov. Jay Collins filed a lawsuit to remove Fishback from the primary ballot under the Florida Constitution's seven-year residency requirement, citing a 2020 vote Fishback cast in Washington, D.C. and mortgage paperwork on a condominium there; in a ruling reported July 27, 2026, Leon County Circuit Judge David Frank dismissed the challenge, finding Collins had not overcome Florida's presumptions in favor of continued residency and ballot access, per WGCU and Florida Politics. The dismissal of that lawsuit is why the residency question never reached the grade: Fishback stayed on the ballot, the certified count includes his 177,809 votes, and every rung settled on that number. This page records the court's finding and nothing beyond it.

What We Are Not Claiming

This page grades market accuracy and nothing else: no endorsement, no characterization of anyone's voters, no trade recommendation. Fishback's 0.5¢ winner contract settling NO tells you almost nothing, because half-cent shots are supposed to lose about 199 times in 200; the 10% rung missing by half a point tells you a little more, but one interior rung on one ladder is not a calibration study. That question needs many rows, which is why our accuracy explainer leans on academic calibration data for now, and why the documented tendency of longshots to be slightly overpriced, the favorite-longshot bias, is one of the patterns this scorecard will test with its own frozen numbers once the November rows land. The Fishback-specific boards, including the concession market and how each scored, are graded on the James Fishback Kalshi odds page; the full certified field against the closing prices on both venues is on the Florida governor primary results page.

The Bottom Line

The Florida board held two numbers that could not both be the story, and now we know which one mattered. Forty-one million contracts traded on James Fishback, 63% of the winner market from the first snapshot to the last, and a price that fell from 1.5¢ to 0.5¢ while the attention never moved. The price was the story. The winner market graded 9 for 9. But the receipt worth keeping is the ladder, because it is where the market did the actual work of measuring a candidate's night and got it almost right: a median near 13% a week out, about 9.7% at the freeze, 10.5% certified, with the one miss landing on the deepest interior book by less than half a point. Kalshi election predictions were sharper at the close than a week before, and still a shade low on the third-place share, while the same freeze had the winner's share nearly ten points too high. That sentence is only worth anything because the rules that produced it were written down before anyone voted. Row 2 gets the same treatment, with every market on the race frozen this time. For the day-to-day coverage that feeds those rows, our full prediction-markets hub is the place to browse.

Trade prediction markets on Polymarket too.

Most readers here already have Kalshi — Polymarket is the other major venue, and code OS4 adds a $20 trading bonus on a $10 deposit: Get the Polymarket bonus →

Affiliate disclosure: OddsShopper earns a commission on signups through this link. Terms as stated by Polymarket; 18+, availability varies by state.

More on this: Brazil Election 2026: Trump's Tariff Is Boosting Lula · Israel Election 2026: Bennett Calls Netanyahu an 'Anarchist · Germany's AfD Is One Vote From Its First State Government · 2026 Midterm Election Odds: Every Race On One Board · How Election Prediction Markets Work & What Prices Mean