Kalshi's protected NFL combos traded at higher prices in every exact pair with usable nearby execution comparisons in an MSR Decode sample. The median difference across those 28 pairs was 0.24 cents per contract. Public settlement records show what that extra protection can do: a player who appears before halftime and never plays afterward can leave a losing selection with a positive fractional payout. They also show its limit: another losing selection can still leave the entire combo worth zero. [2][4][5][7]
Cam Skattebo's October 18, 2026, game supplies a concrete example. He recorded no touches, and the NFL reported that he did not play in the second half. His protected touchdown market settled at 43 cents while its ordinary counterpart settled at zero; a three-leg NFL combo containing that protected selection also settled at 43 cents. The participation rule does not require an injury. This case study follows those prices and payouts to explain what protection changes and where its benefit ends. [1] [2] [3]
A player can miss his target and still leave a payout
Skattebo entered the New York Giants' October 18 game against New Orleans but finished without a carry or reception. The league's account says he appeared before halftime and was absent afterward. The official gamebook's unofficial playtime table records six offensive snaps and none on defense or special teams. Those appearances matter because protection depends on participation, rather than whether the player received the ball. [1]
A combo combines several selections into one contract. In the protected player markets examined here, a player who participates in the first half but in no play afterward, including overtime, can have the contract settle at its pre-kickoff fair price. If the target was reached, the actual statistic controls. Any participation after halftime also sends the contract to actual statistics. The returned terms expressly say that an injury or out designation is unnecessary. [2][11]
The three Skattebo markets examined here show how that condition can change a recorded payout. Each row compares the same player, game and statistical threshold, with and without protection. [2]
Protection changed three Skattebo payouts
Recorded settlement per $1-face-value contract
| Skattebo selection | Protected payout | Ordinary payout |
|---|---|---|
| At least one touchdown | $0.43 | $0.00 |
| At least 30 rushing yards | $0.72 | $0.00 |
| At least 50 rushing yards | $0.41 | $0.00 |
The protection extended into an observed combo. Its three selections were touchdowns by Colston Loveland, Matthew Golden and Skattebo in three separate NFL games. Loveland's and Golden's protected markets each settled at $1; Skattebo's settled at $0.43. The combo itself settled at $0.43, consistent with multiplying those three values. [3]
Skattebo's case also shows why an injury description alone is too narrow. His account and coach John Harbaugh's explanation concerned his usage: Skattebo described punishment, while Harbaugh denied deliberately benching him and described the changing game situation. The reviewed NFL account cites no injury explanation. The documented finding is the combination of a first-half appearance, reported absence afterward and positive protected settlements; it does not establish the reason for every personnel decision. [1][2]
What the matched executions show
A higher protected payout in one game raises a separate question: how did the protected and ordinary versions trade? To examine prices, MSR matched ordinary and protected combo contracts with the same selections, statistical thresholds and selected YES or NO sides. The underlying player markets were checked against Kalshi's official protected-product designations rather than identified from similar titles alone. This produced 55 exact combo pairs. [4]
The acquisition frame used 42 five-second trade windows placed 60, 15 and five minutes before the 14 distinct kickoff anchors for 29 NFL games on October 11-22, 2026. Because the windows sampled the global trade feed, they could include other games and sports. For each protected execution, MSR used the latest strictly earlier ordinary execution on the same trading side within that same window. [4][5]
The qualifying comparisons produced two distinct findings:
Price direction. All 28 exact pairs with usable comparisons had a positive median protected-minus-ordinary execution-price difference. The remaining 27 pairs had no eligible comparison. Every eligible protected execution was on the YES side. [5]
Magnitude. Giving each eligible pair equal weight, the median difference was $0.0024 per contract, just under one-quarter of a cent. The result was unchanged when multiple ordinary trades sharing the latest timestamp were first combined into one median for each protected execution. [5]
These are observed trade-price differences for closely matched contracts. The comparison does not reconstruct a simultaneous menu of quotes offered to one customer: trade sizes differed, and the median separation between the paired execution times was about one second. The result measures an execution-price difference; it does not establish the extra cost paid by a particular customer. [5]
Protection can change one selection while the combo still loses
The same 55 exact pairs allow a direct comparison of recorded final payouts. The results separate cases in which protection changed the outcome from cases in which both versions ended at the same value. [6]
Most matched combos had the same final payout
Recorded settlement per $1-face-value contract
| Final payout comparison | Pairs | Protected | Ordinary |
|---|---|---|---|
| Both settled at zero | 48 | $0.00 | $0.00 |
| Both settled at full value | 6 | $1.00 | $1.00 |
| Protected settled higher | 1 | $0.31 | $0.00 |
The unequal pair combined touchdown selections for George Kittle and Terry McLaurin in the October 19 Washington-San Francisco game. Kittle's protected touchdown selection settled at $1 and McLaurin's at 31 cents. Their protected combo settled at 31 cents; the exact ordinary combo settled at zero. This gives a direct ordinary/protected combo comparison, whereas no exact ordinary twin was recovered for the Skattebo combo. [6]
The broader sample explains a different outcome: protection can leave a constituent with a fractional value while another selection defeats the combo. Among 484 sampled combos containing at least one protected selection, 50 contained a protected selection that settled between zero and $1. Thirty-nine of those combos settled at zero, and every one had another observed losing selection. The remaining 11 settled at fractional values. [7]
The distinction is between protecting a player's participation condition and guaranteeing a return on the entire combination. The zero-paying combos show that a protected selection is not a promise that the customer's purchase will be refunded whenever that player leaves early. The sample's positive fractional payouts show that protection can also preserve value when the other selections succeed. [2][3][6][7]
What the case establishes
Protection changes how a selected player market can settle when participation stops at halftime. In the Skattebo example, that change preserved 43 cents of face value in a combo whose other selections succeeded. In the broader sample, another losing selection left 39 combos at zero despite a fractional protected constituent. These outcomes explain the product’s conditional reach. [3] [7]
The distinction is also public in Kalshi’s help materials, which describe participation conditions, higher protected prices and the effect of another losing selection. MSR’s contribution here is to connect those mechanics to exact contract pairs, nearby executions and recorded settlements. This is an explanatory case study, with evidence that readers can inspect, rather than a finding of undisclosed conduct. [11]
What the evidence does not settle
The observations do not establish overpricing, deceptive sales, a customer loss rate or negative expected value. A benefit can be uncommon and still be fairly priced. Evaluating value would require comparable prices and a probability model; calculating a customer’s result would require the accepted fill, quantity, fees and actual settlement credit. Public contract values supply none of that complete account history.
The price frame covers 210 seconds of global trading, not all NFL activity. Its 55 pairs, broader 484 combos and individual examples overlap and must not be added as independent samples. Eighteen of the 55 pairs were already in an earlier 39-pair index; all 37 newly added pairs had equal final payouts. The protected and ordinary executions were sequential and could differ in size. These counts describe the recovered records, not all customers or all protected combos. [4] [5] [6]
A separate full returned catalog for the October 22 New England–Chicago game supplies a useful equal-outcome check. All 349 protected individual markets had exact ordinary twins: 237 pairs paid zero and 112 paid $1, with no fractional or unequal outcome. Another 50 ordinary-only markets were retained. These are individual player markets, not 349 additional combo pairs. The game was date-selected before acquiring its full catalog, but it was already known in the research. [9]
Participation reporting and the gamebook support the Skattebo reconstruction; they do not establish every snap by half, his medical condition or deliberate punishment. The terms in the captured API records are later returned terms, not a historical order screen. A separate rules question also remains open: older linked combo terms specify a product rounded down to the cent, while a filed amendment uses Exchange Fair Price rounded down to $0.0001 for nonbinary settlements, and a product rounded to the nearest $0.0001 only when the Exchange does not determine Fair Price. The records do not assign a governing version to each legacy contract. A payout compatible with a product calculation therefore does not prove which rule governed, or establish a wrongful payment. [1] [2] [10]