The Colts Won Eight Games and Played Like Ten

Indianapolis finished 8-9 last season. That's the part everyone saw. Here's the part that matters more: the Colts outscored their opponents by 54 points over the course of the year, which — run through the points-per-win conversion our model is fitted to — projects to roughly 10 structural wins. The actual ledger came in two wins short of that. That gap is the luck delta, and a –2 luck delta is not a small number. It means something happened in the margins that the underlying play didn't deserve.

The margins, in this case, were close games. Indianapolis went 3-7 in one-score decisions — a coin flip gone wrong, basically, and sustained over ten opportunities. One-score results are the noisiest data in football; they correlate almost nothing with the next year's one-score record, but they correlate strongly with the next year's win total, because teams that played well enough to win ten games while going 3-7 in the close ones tend to find the middle of the distribution the following season. That's the single most predictive fact about this roster heading into 2026: it played like a ten-win team and got eight. Something like regression is the base case, not the optimistic one.

The EPA profile is quietly tidy. Indianapolis generated 0.09 expected points added per pass attempt on offense — positive, functional, nothing that makes a coordinator nervous — and 0.08 per carry on the ground. Neither number is elite, but both are in the black, and a team that's above zero in both run and pass phases on offense has its house in reasonable order. The situational numbers back that up: a 47.2% conversion rate on third and fourth downs, and 0.14 EPA per play in the red zone. You'd prefer more separation, but those aren't the numbers of a team that breaks down when the stakes go up.

The defense is where the profile gets genuinely interesting. Opponents produced 0.06 EPA per pass attempt against this unit — positive for them, but only modestly so. Against the run, the Colts held opposing backs to –0.06 EPA per carry. A negative rush EPA allowed is meaningful; it says Indianapolis was actively taking value away from opposing offenses on the ground, not just limiting damage. Combined, this is a defense that could hold a game together while the offense found its footing.

Head coach Shane Steichen ran a go-for-it rate of 25% on fourth down in 2026 situations where our decision model recommended going 40.9% of the time. His overall decision agreement rate was 77.3%, and the model pegs the conservatism on fourth down at about 0.285 expected wins left on the field. That's not a catastrophic number, but it's real — roughly a quarter of a win per season sitting in the gap between what Steichen called and what the numbers preferred. On a team that finished two wins below its structural expectation, that's a detail worth watching.

The departures define the offseason more than the arrivals do. Kwity Paye left for the Raiders on a deal worth $16M annually — a contract the market priced at 1.29 wins per year. Braden Smith went to the Texans at $10M annually (0.77 market wins), taking with him one of the better right tackles in the division. Nick Cross headed to Washington at $6.5M (0.46 wins), and Neville Gallimore signed with Chicago at $5M (0.33 wins). Add those up and Indianapolis shipped out 2.85 market wins worth of talent in the front seven and secondary.

What came back: Arden Key from Tennessee at $8M annually (0.59 market wins) and Micheal Clemons from the Jets at roughly $5.7M (0.39 market wins). Together, 0.98 market wins at edge rusher, replacing a group the market valued at closer to 1.62 at that position alone. The net offseason swing on a market-value basis is –1.87 wins. On a production basis — where we'd adjust for what those players actually delivered versus what their contracts implied — the number nets to zero, meaning the departed players underperformed their price tags enough to wash. Both readings are honest; neither is comfortable.

The contract verdicts are the part of the ledger that will be debated longest. Daniel Jones signed a two-year, $44M-per-year deal — 94th percentile in pay, 69th percentile in production, with an aging multiplier of 1.06 (meaning our curves still project modest growth at 29). The exit if it goes sideways is $35.2M in dead cap in 2027. The verdict from our underwriting model is OVERPAY. Alec Pierce got four years at $28.5M annually — 98th percentile in pay, 71st in production, aging multiplier of 0.75 at age 26, meaning the curves expect the gap between pay and production to widen rather than close. Exit in 2028 carries $15.6M dead. Also OVERPAY. The one clean verdict in the new contracts: Jonathan Taylor at $22M annually reads as a FAIR PRICE — 99th percentile in pay, 95th in production, aging multiplier of 0.92 at 28. For a running back at that age, you're buying the tail end of a production peak, and the price reflects it accurately.

The draft added 1.77 chart wins across eight picks. CJ Allen, a linebacker out of Georgia at pick 53, brings 0.8 chart wins — the clear headliner of the class. A.J. Haulcy, a safety from LSU at pick 78, adds 0.6. Jalen Farmer (guard, Kentucky, pick 113) and Bryce Boettcher (linebacker, Oregon, pick 135) contribute 0.37 and 0.34 respectively. The remaining picks — George Gumbs Jr., Caden Curry, Seth McGowan, and Deion Burks — total 0.71 across rounds five through seven. A pair of compensatory picks projected for 2027 (rounds five and six) adds another 0.4 in future value.

The honest answer is: hurt, modestly, at the position that matters most. The edge rusher group is thinner than it was. Paye's contract implied 1.29 wins of value; Key and Clemons together imply 0.98. That's a real step back in a division that will require the pass rush to work. The linebacker group is the most underspent unit on the roster by cap allocation — our spending z-score for that group is –1.84, meaning Indianapolis is investing significantly below the league average there — which makes Allen's arrival at pick 53 feel less like depth and more like triage. Whether a second-round rookie can close a gap that large is a reasonable question and an open one.

The corners are the inverse story. The cap z-score for the CB group is +1.69, the highest positive allocation on the team, which means Indianapolis has committed real money to that position. Charvarius Ward sits at 6.6% of the cap — the third-largest individual hit on the roster. Whether that investment pays off depends in part on the pass rush creating enough pressure that Ward's coverage doesn't have to be held long.

The roster's clearest structural asset remains Tyler Warren, who produced at the 93rd percentile among tight ends last season on a cap hit of just 1.6% — 56th percentile in pay. That's the kind of value that buys an offense time while the expensive pieces find their level. The aging flags are concentrated at receiver: Pierce's curve is decelerating at 0.75 even before the ink is dry on his extension, which is an unusual place to find a curve that shape on a 26-year-old. It suggests the model sees his peak production already in the rearview, which makes a four-year commitment at 98th-percentile pay an uncomfortable hold.

The running back group is, by z-score, the second-most-invested position group on the roster at +1.42. Taylor's fair-price verdict means that money is at least going somewhere the production warrants. But the overall cap picture — $445.4M in future proration already committed — is a reminder that flexibility isn't unlimited. The two overpay verdicts at quarterback and wide receiver will constrain what Indianapolis can do if either player underdelivers.

Here is the arithmetic, plainly stated. The 2025 point differential projects to a structural base of 10 wins. The offseason moves net to zero on a production basis — the market-value basis shows –1.87, but where we have production data to adjust, the gap closes to flat. The coaching staff is unchanged, so there's no decision-model delta to add or subtract. The structural wins projection for 2026 is 10.

That is an accounting projection from the ledger — not a betting line, not a prediction, not a pick. It is the number the inputs produce when you run them through the same methodology applied to all 32 teams. What would prove it wrong: if Indianapolis wins 12 or more, the bounce-back from the one-score record was real and probably reinforced by the Jones-Pierce combination outperforming their contracts. If they win seven or fewer, the –1.87 market-basis swing at edge rusher and the overpay exposure at quarterback will have been the story, and this read will have been too generous to the production-basis adjustment. The win total that makes this a miss in the optimistic direction is 12. The win total that makes it a miss in the pessimistic direction is seven.

The brief shows no vacated opportunity numbers to redistribute — departing players were concentrated on defense, where targets don't accumulate the same way. The fantasy story here is less about a target windfall than about the efficiency picture at the positions that do touch the ball.

Warren is the cleanest case on this roster. A 93rd-percentile producer at tight end paying 56th-percentile cap cost is exactly the gap that takes time to close in the market, and fantasy markets tend to follow the same lag. The efficiency was real last season; the price hasn't fully caught up. That's what a sleeper actually is — production the market hasn't priced in yet — and it describes Warren better than any manufactured narrative.

The aging flag on Pierce deserves a direct note. A 0.75 aging multiplier at 26 is low enough to treat as a genuine caution rather than background noise. Our curves project his production trend decelerating, not accelerating, which makes the upside case for him in fantasy dependent almost entirely on volume rather than efficiency improvement. Volume-only cases are fragile ones. Managers who paid up for Pierce based on the contract extension as a signal of organizational confidence are leaning on the right instinct — the Colts clearly believe in him — but the curves say that belief may be running ahead of where the production is headed. If you're building a season-long roster, the efficiency trajectory is worth respecting.

Everything above traces to a live instrument, and the receipts stay public: the GM boards carry Shane Steichen's decision card, the team's cap posture ($445.4M already charged to future caps), and every contract verdict with its falsifier and review date — including Daniel Jones (OVERPAY), Alec Pierce (OVERPAY), Jonathan Taylor (FAIR PRICE). When the season grades these reads, the grades post whether they flatter us or not.