Can you use 0xinsider for Kalshi or DraftKings? Yes: 81 games priced on all three

0xinsider tracks Polymarket. It grades every wallet on its settled profit and loss, shows which wallets hold each side of a game, and lists large trades as they fill. It does not track Kalshi accounts or DraftKings customers, because neither venue publishes them.

You can still use it for both. A Chiefs game on Kalshi, on DraftKings, and on Polymarket is one game with one result, and the three venues price it almost identically. We measured that on September 18, 2026, and the rest of this page is that measurement. The read comes from Polymarket, where the wallets are public. You act on it wherever you hold an account.

0xinsider is a read-only analytics site. It needs no Polymarket account, takes no deposits and places no orders, so using it does not depend on which venue you trade on.

The capture ran once, at 16:01 UTC on September 18, 2026, and took 8 seconds. It read three public endpoints: Polymarket's market list, Kalshi's event list, and ESPN's scoreboard, which publishes the DraftKings line for each game. It covered games starting September 18 to 21, 2026 in 4 leagues: the NFL, college football, MLB, and the WNBA.

ESPN listed 150 scheduled games, and 88 of them carried a DraftKings moneyline with a number on both sides. DraftKings posts most baseball and WNBA moneylines on the day of the game, so only 15 of 48 MLB games had one. Of those 88, Polymarket priced 83 and Kalshi priced 81. All three priced 81, and those 81 games are the sample.

Every price below is the home team's chance of winning, in percent. On Polymarket and Kalshi it is the midpoint between the best bid and the best ask. On DraftKings it is the moneyline converted to a probability with the margin removed, because a sportsbook's two sides add up to more than 100%. The chart draws each game twice, once for each exchange, against the DraftKings number. Points on the dashed line are games where the venue and DraftKings agree exactly.

On 57 of 81 games the two exchanges had the same midpoint. 84.0% of games were within 1 point, 98.8% were within 2, and the widest gap in the sample was 2.00 points. The mean gap, ignoring direction, was 0.25 points, and neither venue ran consistently higher: the mean signed difference was -0.02.

Both books were tight. The median gap between bid and ask was 1.00¢ on Polymarket and 1.00¢ on Kalshi, and no game on either venue was wider than 2.00¢. A difference of one point between the two venues is a difference of one tick.

For a Kalshi reader this is the whole case. A price you read on a 0xinsider game page is, to within 1¢, the price on your Kalshi screen. A large Polymarket trade that moves the price there moves it on Kalshi too, or a trader takes the difference between them.

The table lists every NFL game still to be played at the capture. Across the 15 games, Polymarket and Kalshi differed by a mean of 0.20 points and by no more than 1.00. DraftKings, with its margin removed, differed from Polymarket by a median of 1.81 points, the widest of the four leagues. MLB was the closest: a median of 0.36 points across 14 games, none further than 1.33.

The widest gap in the whole sample was Miami Dolphins at San Francisco 49ers. DraftKings had the favorite at -900 and the underdog at +600. Removing the margin proportionally gives the favorite 86.3%. Polymarket and Kalshi both had it at 89.5%. The -900 price alone implies 90.0%, so DraftKings was charging close to the exchange price for the favorite and holding nearly all of its margin on the underdog. The section after next measures that pattern across every game.

Across all 81 games, the DraftKings fair price sat a median of 0.99 points from Polymarket and a mean of 1.15. 50.6% of games were within 1 point, 84.0% within 2 and 97.5% within 3. The widest was 3.20 points. The correlation between the two sets of prices was 0.9993; between Polymarket and Kalshi it was 0.9998.

All three venues named the same favorite in 80 of 81 games. The exception was Wyoming Cowboys at Central Michigan Chippewas, a pick'em: DraftKings' fair price had the home team at 49.6%, Polymarket at 48.5% and Kalshi at 50.5%.

Part of the gap to DraftKings comes from the method used to remove the margin. The usual method divides each side by the two sides' sum, which takes the margin off in proportion to the price. The other common method subtracts the same number of cents from each side. Done that way, DraftKings sat a median of 0.66 points from Polymarket, 70.4% of games were within 1 point, and 95.1% were within 2. To compare your DraftKings line with an exchange price, take half the overround off each side.

Read on the DraftKings favorite, Polymarket's price was higher than the proportional fair price in 64 of 81 games, by a mean of 0.86 points. The difference grows with the favorite. In games where the favorite's fair price was 50 to 60%, it was 0.20 points. Where the favorite was 90% and up, it was 1.66 points, and Polymarket was the higher of the two in 13 of 14 games.

That is what a margin charged in roughly equal cents per side looks like. 2¢ on a 90¢ favorite is about 2% of the price; 2¢ on a 10¢ underdog is about 20% of it. In this sample the gap between the DraftKings price and Polymarket's price with the taker fee included was a median of 0.99% of the DraftKings price on the favorite and 2.27% on the underdog.

For a DraftKings reader the practical point is about underdogs. A plus-money price that looks generous still carries the larger share of the margin, so compare it against the exchange price before you take it.

DraftKings' two sides added up to a median of 4.25 points more than 100%, and never less than 3.11 or more than 4.83. That overround is what a customer pays to hold both sides of one game: 4.25¢ for every $1 of payout.

On Polymarket the same position costs the gap between bid and ask plus the taker fee. Polymarket's published sports fee is 0.05 times the price times one minus the price per share, so it is largest at 50¢ and falls toward either end. Buying both sides at the ask and paying the fee cost a median of 3.16¢ over $1, from 0.92 to 4.45, and it came in under the DraftKings overround in 80 of 81 games. A resting order that fills pays no taker fee, which leaves the 1.00¢ spread.

One side at a time, the favorite cost a median of 1.55¢ more per $1 of payout on DraftKings than at Polymarket's ask, and 0.74¢ more once Polymarket's taker fee is added. Polymarket was the cheaper of the two on the favorite in 62 of 81 games and on the underdog in 61.

Kalshi's quoted spread matched Polymarket's, and DraftKings' favorite cost a median of 1.55¢ more than Kalshi's ask. Kalshi also charges a trading fee. We could not read its current schedule from an official source when we ran the capture, so no Kalshi fee is in these numbers. Read Kalshi's fee schedule before you compare the two exchanges on cost.

A Polymarket or Kalshi price is a probability in cents: 62¢ is a 62% chance. A DraftKings moneyline is the same probability written as a payout, with the margin on top. To compare them, convert the exchange price to American odds. At 50% or more, the fair American price is minus 100 times the chance, divided by 100 minus the chance, so 62¢ is about -163. Under 50%, it is plus 100 times 100 minus the chance, divided by the chance, so 38¢ is about +163.

The table carries the common prices, along with what a book holding this sample's median overround would post for the same chance. If your DraftKings price is better than the fair number in the middle column, the book is offering more than the exchanges think the side is worth. The calculator at 0xinsider.com/tools/odds-calculator converts any price in either direction.

The price is the part of a game that every venue shows. Who is behind it is not. Kalshi clears trades through a regulated exchange with no public account identity, and a sportsbook publishes a line and nothing about its customers. Polymarket settles on a public ledger, so every trade carries a wallet, and every wallet has a settled record anyone can add up. 0xinsider grades each wallet from S to F on that record.

The grade carries information in sports. Across 155,832 graded sports buys from June 1 to September 11, 2026, wallets graded S, A, or B won 65.9% of the time against an average price of 64.7¢, +1.25 points, with an interval from +0.20 to +2.31. Wallets graded D or F came in -1.21 points. The full study is at 0xinsider.com/research/sharp-money-polymarket-sports.

Where that edge sits matters for anyone acting on another venue. Before kickoff the S, A and B cohort was +0.27 points, which is close to nothing. In-play it was +1.95 and the D and F cohort was -2.56. The pre-game signal with an interval that clears zero is a different one: when 90% or more of the D and F money on a market sat on one side before kickoff, that side was priced at 67.3¢ and won 63.6% of the time across 844 markets from June 1 to September 12, 2026, -3.68 points, with an interval from -6.71 to -0.69. That study is at 0xinsider.com/research/fading-the-crowd-polymarket-sports.

Open the game on 0xinsider. The sports boards at 0xinsider.com/sports list every game Polymarket prices, and each game page shows the moneyline, spread and total families with the volume on each. The Top holders tab lists the wallets on each side with the grade each has earned. The feed at 0xinsider.com/whale-alerts shows large trades, live with Pro and a day behind without it.

Then find the same market on Kalshi. Moneylines map one to one: both venues list a market on each team to win, and this capture found their prices within a tick of each other. Check the price on Kalshi against the price on the game page before you trade. If they differ by more than 1¢ or 2¢, one of them has moved and the other has not caught up yet.

Spreads and totals take one more step. Polymarket lists a ladder of lines on each game, and Kalshi lists its own. Match the line as well as the side, because a spread of 2.5 and a spread of 3.5 are different markets with different prices. The companion study at 0xinsider.com/research/polymarket-spreads-totals-vs-draftkings measures how Polymarket's ladders line up with a posted line.

Start the same way: open the game on 0xinsider and read which wallets are on each side and what they paid. Then convert. Take the Polymarket price from the game page, turn it into American odds with the table above, and compare it with the moneyline on your DraftKings screen. The DraftKings number will be worse than the fair number on both sides, by about 2 points each at the median. It is a candidate only when the gap is smaller than that, or reversed.

DraftKings moves its line on its own schedule. A large trade can move Polymarket in a second, and a book may take longer to follow or may not follow at all if its own customers are on the other side. This capture measured agreement at one moment before the games. It did not measure how quickly either venue follows the other.

DraftKings also offers bets that Polymarket does not list, such as same-game parlays and many player props. 0xinsider has nothing to say about those. Its read covers the markets Polymarket lists: the moneyline, spreads, totals, and the props and futures on the game page.

Every game in the capture has now been played. This section scores the three prices against the results. The sample is the one the capture fixed: the same 81 games, each final read back from ESPN's scoreboard by the event id the capture itself recorded, with none added and none dropped once the results were known. The home team won 48 of 81.

Two scores, both on the home team's chance of winning. Brier is the mean squared distance between the price and what happened. Log loss is the mean of minus the log of the probability the price put on the side that won, and it punishes a confident price that was wrong much harder than a cautious one. Lower is better on both.

Polymarket scored 0.1502, Kalshi 0.1504, and the DraftKings moneyline with its margin removed 0.1510. A price that gave every game the sample's own home rate of 59.3% would have scored 0.2414, and a coin 0.2500. All three venues carried real information about these games, and close to the same amount of it.

Take each difference game by game and average it. Polymarket's Brier came in 0.0008 below DraftKings', with a game-resampled interval from -0.0027 to +0.0011. Kalshi against DraftKings was 0.0006 lower, from -0.0025 to +0.0014. Between the two exchanges it was 0.0002. Every interval spans zero, on both scores.

The ordering holds across resamples without being decided by them. Polymarket beat DraftKings on Brier in 80.7% of 20,000 resamples and on log loss in 94.9%. Kalshi beat DraftKings in 74.0% and 93.2%. Polymarket against Kalshi is 68.3% and 60.8%, which is a coin with a lean on it.

What is missing is games, not a closer reading of these. Holding the per-game spread where this sample put it, a Polymarket-to-DraftKings gap this size stops spanning zero at around 425 games; between the two exchanges it takes around 1,658. One Friday-to-Monday across four leagues is 81. A claim that one of these venues prices sports better than another, made on a sample this size, is reading noise.

The DraftKings favorite won 65 of the 81 games. Banded by what each venue charged for the favorite, the prices sat close to the record everywhere. On DraftKings, favorites priced 50 to 60% won 17 of 25, 12.6 points above the 55.4% the price implied, which is 1.27 standard errors on a band of 25 games. Favorites priced 90% and up won 14 of 14.

Every venue's near-even band came in above its price and every venue's heaviest band went undefeated. Both are what 81 games look like rather than a pattern: the largest band here holds 25 games, where one standard error on the win rate is about 9.9 points.

Read across the venues rather than down one. Each venue's games are banded by its own price, so a game close to a boundary can sit in one band on DraftKings and the next one on Polymarket, which is why the counts move by a game or two from row to row.

The obvious test is the games the venues disagreed on most. Take the 10 games where Polymarket sat furthest from the DraftKings fair price, 2.25 to 3.20 points apart. The side Polymarket leaned toward won all 10. Its Brier on those games was 0.0281 against DraftKings' 0.0358.

That is a fact about the margin, not about Polymarket. Removing a sportsbook's margin proportionally leaves the exchange price above DraftKings on the favorite in 64 of 81 games, and the favorite won 65 of 81. All 10 of those leans pointed at the favorite, so they were right about four times in five before any venue showed judgment.

Take the margin off the other way and the test reverses. Subtract the same number of cents from each side, which is the method this page recommends for comparing a DraftKings line with an exchange, and Polymarket's lean points at the favorite in 40 of 81 games instead of 64. It then named the winning side in 36 of 81, worse than a coin, and in the ten widest disagreements on that measure it was right in 6 of 10: 6 of 6 where it leaned to the favorite and 0 of 4 where it leaned to the underdog.

So the answer to which price was right is: all three, about equally. The one game where the venues did not name the same favorite was Wyoming Cowboys at Central Michigan Chippewas. Kalshi had the home side at 50.5%, DraftKings' fair price at 49.6% and Polymarket at 48.5%. The home team won 24-10. One game decides nothing either, which is the same point.

It does not show that any one venue is more accurate. The addendum above scored all 81 games and every pairwise interval spans zero, so this sample separates none of the three. For a larger read on how Polymarket's sports prices compare with results, the calibration study at 0xinsider.com/research/favorite-longshot-bias-polymarket-sports scored 411,770 buys against the outcome.

It is one moment. Every price was read within 8 seconds at noon ET on a Friday, a day or more before most of the games. It covers moneylines only, in four US leagues, before the start. It says nothing about in-play prices, where Polymarket's graded wallets show most of their edge and where a sportsbook's live line is slowest to compare.

It does not show that Polymarket is the larger market on these games. At the capture, Polymarket's moneylines on the 81 games had traded $1,119,880 and Kalshi's game markets 14,800,286 contracts, each paying $1. On US league games before the start, Kalshi's book was the bigger one, and the prices matched anyway. What Polymarket has that Kalshi does not is the public wallet.

The DraftKings line is the one ESPN's scoreboard publishes, read from ESPN and not from DraftKings. A line can differ by state and can move between ESPN's refresh and your screen. And the rules for settling a game can differ between venues in rare cases such as a postponed or shortened game, so read the rules on the venue where your money is.

All three sources are public and need no key. The capture script, the fields it saved exactly as each provider returned them, the analysis script, its full output and a row for each of the 81 games are published at github.com/0xinsider/research under three-venue-prices. Running the analysis script on the saved capture reproduces every figure on this page with no database and no account.

The addendum adds two scripts beside them. The first reads each game's final from ESPN's scoreboard by the event id the capture recorded and writes one row per game; the second joins that file to the capture and prints every score, interval and band on this page. It refuses to run if the two files do not hold the same 81 games, which is the check that the sample did not move once the results were known. The finals on this page were read on September 23, 2026.

Running the capture script again gives a new sample, because these are live prices. Expect the same result at a different moment: two exchanges a tick apart and a sportsbook line a point or two away once its margin is out. If you find a moment where that fails, we would like to see it.