Army at Temple. Lean: Army -3. Lean: Under 48.5.

Spread.The strongest explanation for Army -3/-3.5 is its preparation advantage: Army comes off a bye while Temple faces a short turnaround after an overtime road loss. That matters against an option offense when Temple has reported assignment problems. Our Army -1.8 offers only a small disagreement, not a demonstrated betting edge. The incremental Temple case is better than the injury feed suggests: Sheppard is the announced starter, and Sam Brown has returned productively. Temple also began preparing for Army in spring, tempering the short-week argument. Those developments support a Temple lean, but they do not establish that bookmakers have missed them. The captured spread has remained steady, and FanDuel charges -120 for the valuable +3.5. I would require +3.5 at -110 or better; +3 at the available prices does not compensate adequately for Army's preparation advantage and Temple's defensive uncertainty. Pass, with LOW conviction; confidence is ordinal, not a probability.

Total.The strongest explanation for 47.5-48.5 is limited possessions from two rushing-oriented offenses, Temple's personnel uncertainty, and breezy conditions. Every captured book has lowered its total by one to three points. The over response is substantive: Sheppard generated four touchdowns at Toledo, Brown returned effectively, and Temple's defense allowed 536 yards. However, that 97-point final included overtime and an extraordinary late comeback, making it a poor baseline. Army's 59 against Bryant is another inflated comparison. Our 51.36 supports the direction but lacks a proven edge and does not model this possession matchup. I lean over after the decline, yet require 47.5 at -110 or better. DraftKings offers that number only at -118; the cheaper alternatives require 48 or 48.5. Pass with LOW ordinal conviction.

Where our number looks wrong.The spread model cannot distinguish Temple's new Sheppard/Brown offensive configuration from its earlier personnel, or account for the unequal preparation windows. Its flat 2.5-point home advantage is also unvalidated for this setting. I cannot establish that Army -1.8 is more accurate than the market. The totals projection similarly misses possession structure and personnel changes; its supplied implementation already deducts 2.597 points for wind despite boilerplate claiming it sees no weather. Applying another full wind adjustment would double-count.

• Covers lists Smolik, Hardy, Stocksdale and Logan as questionable, with September 20-21 designations.
• The accessible 2025 American policy specifies availability reports for conference games by two hours before kickoff; it does not establish this game's current availability.
• An indexed September 14 report says Bousum left the USF game with a foot injury; no current designation was established.
• Sheppard produced 164 passing yards, three passing touchdowns, 66 rushing yards and one rushing touchdown against Toledo.
• Toledo accumulated 536 yards; the teams reached overtime tied 42-42 following an onside-kick-assisted comeback.
• Friday's daily forecast is mostly cloudy and breezy, with a 73°F high and 1% precipitation probability; this is not an hourly stadium forecast.

Spread.Market first: Army -3 has not moved from the open at four of five books (FanDuel -3.5), so the market is comfortable pricing Army's ground game, 45th in EPA per rush and over 7 yards a carry, against a Temple run defense that SI has 118th in EPA per rush, offset by Temple's 18th-ranked pass offense. The news since then leans Army. Temple starts backup Ajani Sheppard with Jaxon Smolik nursing a foot injury (Inquirer 9/22), plays on six days' rest after blowing a 14-point fourth-quarter lead in a 49-48 overtime loss at Toledo, and faces an Army team coming off an idle week. The dampener is that Sheppard threw three touchdowns in relief at Toledo, so the quarterback downgrade may be small, and Army itself blew a 10-point lead in the final 3:13 against USF. The situational case is real but -3 is the key number and I would not lay a unit on it. Lean Army at -3 flat, pass.

Total.The total has fallen from 49 to 51.5 at open to 47.5 to 48.5, so the market already expects a slower game. The under case: NWS forecasts north wind around 15 mph with gusts to 25 for a 4pm kickoff, Army will try to grind a 118th-ranked run defense and shorten the game, and Temple is on a backup quarterback. The over case is just as concrete: Temple's last game produced 97 points, Army scored 59 on Bryant and 24 on USF, Army's pass defense collapsed late against USF, and our totals model sits at 51.4 even after its wind penalty. With the market already three points lower and the model three points higher, I have only a slight lean under at 48 and no bet.

Where our number looks wrong.The rating has Army with only two games and still near its preseason prior, and Temple's rating does not know its starting quarterback is out. It also ignores rest: Army has 13 days, Temple 6. If anything the model understates Army here.

• Friday: north wind around 15 mph, gusts as high as 25 mph, high near 70
• Prediction Temple 26, Army 23; line Army -3.5, total 49.5

Spread.Army is favored by 3 to 3.5 on the road at Temple. Army's roster continuity is extreme by FBS standards, 80.2% returning production versus Temple's 37%, and its disciplined, run-first option attack (only 150 passing yards through two games) has produced a 59-3 blowout of Bryant and a competitive 24-28 road loss to South Florida. Temple, by contrast, needed overtime to lose to Toledo 48-49 and was blown out by Penn State 9-27, and now carries real QB uncertainty: starter Jaxon Smolik is questionable with a foot injury (team log, 9/21) while ESPN's listed season passing leader for Temple, Ajani Sheppard, already has meaningful volume (297 yards, 3 TD), suggesting a timeshare or a recent change I can't fully resolve. The line has been stable at Army -3 to -3.5 across every tracked book since open, so the market isn't reacting to fresh news, but the returning-production gap plus Temple's QB uncertainty argue for taking the favorite at a standard stake rather than passing.

Total.All five books have cut this total 2-3 points since open (from a 49-51.5 range down to 47.5-48.5). There's no confirmed weather driver: ESPN's own note for Philadelphia is mild (73F) and there are no active Atlantic tropical cyclones as of 9/23, so I can't point to a storm. The more plausible driver is Temple's QB uncertainty: Smolik (foot, questionable as of 9/21) sits alongside three questionable Temple receivers, and Army's option offense is built to shorten games (150 passing yards through two games, ball-control style) rather than push pace. Both offenses lean toward fewer possessions and lower efficiency than the preseason total implied, and I don't have a specific data point pointing back toward the over at this price.

• ESPN lists the current total at 47.5 (DraftKings, over -118/under -102) with mild forecast conditions (73F) at Lincoln Financial Field
• Temple QB Jaxon Smolik is questionable (foot) as of 9/21, with three additional Temple wide receivers also questionable