Edge Labs EDGE LABS Market intelligence
CFB NFL
← The slate
CFB · Week 2 · Sat, Sep 12 · 7:00 PM ET · Elliott T. Bowers Stadium · FINAL Tulsa 23, Sam Houston 17

Tulsa at Sam Houston. Our picks: Sam Houston +13.5 (medium). Under 51.5 (low).

The picks come first on this page. The market's number, our number and the gap between them sit under them as MODEL CONTEXT: that is the input our analyst panel argued from, not a call of its own. Tiers are how decisively the evidence agreed, never a win probability.

WIN We had Sam Houston +13.5. It cashed.
ATS pick
Sam Houston +13.5
medium WIN
locked Thu 1:06pm ET
locked +13.5 · closed +12.5
Tulsa 23, Sam Houston 17
Total pick
Under 51.5
low WIN
panel 2-1

Panel reasoning

On the spread

Tulsa at Sam Houston, Tulsa -13.5. The number has firmed toward the favorite since the open: DraftKings moved from -10.5 to -13.5 and FanDuel from -12.5 to -13.5 between September 4 and 10, while BetMGM, Bovada and Pinnacle first posted -14 on September 6, right after Tulsa's Week 1 win, and Pinnacle is still there. No reliable current-game ticket or handle splits are available. Tulsa beat Oklahoma State 24-10 on September 5 on only 9.1 percent third-down conversion, helped by four takeaways. Sam Houston lost 24-21 at Troy the same day while outgaining Troy 373-342, with Landyn Locke throwing for 335 yards. One panel notes Tulsa's top cover corner has been out since preseason, though no update confirms his status this week. Our own model's raw read favored Sam Houston before calibration flipped it to Tulsa by 8.1, and the 5.5-point gap to market sits under our lean floor, so it offers no edge. One panel took Tulsa anyway, citing Tulsa's defense and Sam Houston's weak run game. Take Sam Houston +13.5 (+14 at Pinnacle).

On the total

Under 51.5. Every book we track has bled this number down since the open, from the 52-53.5 range to 51-51.5, with the sharpest book now at 51. Both Week 1 games finished well under comparable totals, and Tulsa converted only 9.1 percent of third downs in its opener, capping drives. The forecast calls for a hot evening with a modest chance of storms and no meaningful wind. Our own totals model sits just 0.8 points under market, well inside its own noise threshold. The counter-case: one panel leans over, arguing missed field goals and failed fourth-down tries left points on the board and better drive-finishing could lift scoring. Take Under 51.5 (51 at Pinnacle).

Model context

The numbers the panel argued from. Our models price a game, the panel makes the call, and the picks above are the call. Nothing in this block is a pick.

Locked line
Tulsa -13.5 · Sam Houston +13.5
the spread we locked at, it does not move
Our model number
Tulsa by 8.1
the calibrated EL CFB rating number, home field included
Spread gap
4.4 points
the distance between our model number and the market. Context for the panel, not a call of its own.
Market total
51.5
consensus of 5 books at the model run, Sep 12, 12:38 PM ET
Our total
50.5
EL CFB Totals v1
Total gap
-1.0
the distance between our model total and the market total. Context for the panel, not a call of its own.

Context, not a call. The spread number is our rating difference plus home field. The totals number is a registered candidate that loses to the closing total: across 2,264 held out games it missed the combined score by 12.862 points on average and the closing total missed by 12.635, ahead of us in all 3 of those seasons, so it never posts to the board and never enters the record on its own.

Over and under splits

Descriptive context on this game, every line with the games it counts. None of it is a reason a pick wins and none of it enters a tier.

  • Early-season total of 51.5 fits the validated under window, unders 55% historically (n=943)
    total · under · moderate · n=943
    H64 (validated): Aug through mid-Sep games with a closing total of 47+ landed under 55% of 943 games, 2015-2025. Recent seasons: 2022: 58% (n=92) · 2023: 53% (n=81) · 2024: 57% (n=103) · 2025: 52% (n=119). 2025 ran soft at 52.1%, below the -110 breakeven, and the 2026 forward test is pre-committed on the early-totals page. Early-season totals are shaded up by public over-action; the market corrects as real form data arrives.

The panel

Panel A

Sam Houston +13.5

Over 51.5

Take Sam Houston +13.5 at the freshly verified FanDuel/BetMGM number. EL makes Tulsa an 8.1-point favorite, a 5.4-point disagreement, but its unsuccessful ATS validation prevents treating that gap as established value. The stronger argument is underneath last Saturday's scores: Tulsa won by 14 despite being outgained 368-361, benefiting from four takeaways and repeated fourth-down stops. Sam Houston outgained Troy 373-342, with Landyn Locke producing 335 passing yards. That supplies a plausible route to staying competitive or scoring a late cover. The dossier's DraftKings move from Tulsa -10.5 to -13.5 already rewards Tulsa's opener; other books opening at -14 show this was not uniform market steam. Tulsa's efficient passing and Sam Houston's weak rushing output are legitimate reasons to favor the visitors. Nevertheless, nearly two touchdowns asks too much given the underlying Week 1 performance. Unverified availability and losing the key +14 keep this below a strong position.

Total. A small over lean despite EL's 50.57 projection. The under case is credible: both openers finished below this total, Sam Houston managed only 38 rushing yards, and the dossier shows totals falling from 52-53.5 at several books. Tulsa's defensive stops could again stall possessions. The over case is that those final scores conceal scoring opportunities: Tulsa missed two field goals, Oklahoma State repeatedly failed on fourth down, and Locke's 46 attempts produced 335 yards. Hayes also generated 227 yards on just 24 attempts. More ordinary drive conversion could lift scoring without requiring faster tempo, which I could not verify. Heat and uncertain shower timing provide no firm weather edge. The reduced market total makes over marginally preferable, but this is close to a coin flip.

Where our model looks wrong

The margin-only refit rewards Tulsa's 14-point win without recognizing its near-even yardage and turnover dependence, while Sam Houston's three-point loss conceals a yardage advantage and improved passing production. Both effects could exaggerate Tulsa's superiority, although they do not validate EL's exact -8.1. The totals documentation also contradicts itself: it claims no weather input while applying a -1.751-point wind adjustment.

Panel B

Sam Houston +13.5

Under 51.5

Our number is Tulsa -8.1; the market is -13.5 retail and -14 at Pinnacle. The 5.5-point gap sits just under our own lean floor, and I do not take it at face value: preseason SP+ implies roughly Tulsa -15 after home field, so our rating is the outlier, not the market. What moves me to the dog is the spot. Tulsa is a 4-8 team from 2025 that beat Oklahoma State 24-10 as a 13.5-point home underdog on four takeaways while converting 9.1% on third down (133rd), scoring on 50% of red-zone trips (120th), missing two field goals and taking 102 penalty yards (130th). Tre Lamb himself said it should have been 37-10 and that they must clean things up in a hurry before a road game. That offense must now cover two touchdowns on the road. Sam Houston covered +16.5 at Troy, outgained the Trojans 373-342 and lost by three, with Landyn Locke throwing for 335 and three scores in Longo's Air Raid, against a Tulsa secondary missing first-team all-conference corner Elijah Green. It is also the first true home game at renovated Bowers Stadium since 2024. Retail 13.5 is a half-point better than the sharp number.

Total. Every book we track has bled this down: BetMGM 52.5 to 51.5, DraftKings 53.5 to 51.5, Bovada 52 to 51.5, Pinnacle 52 to 51 between 9/4 and 9/10. That one-directional drift is the cleanest signal in the game, and the Week 1 tape agrees. Tulsa scored 24 with four takeaways handing it short fields while going 9.1% on third down, 50% in the red zone and missing two field goals; Sam Houston turned 373 yards into 21 points. Tulsa's defense held Oklahoma State to 10; Sam Houston's produced four sacks and eight tackles for loss. Both Week 1 games landed well under (34 vs 60.5, 45 vs 54.5). The over case is real, Longo's Air Raid threw 46 times, Tulsa is missing its best corner, and there is no wind, so this is a lean, not a bet.

Where our model looks wrong

Our number (Tulsa -8.1) is the outlier, not the market. Preseason SP+ had Tulsa -10 and Sam Houston -27.8, which is Tulsa by about 15 after home field; the market at 13.5-14 sits in that neighborhood, and Pinnacle posted 14 after Week 1 and has not moved off it. Our EL rating is still mostly the preseason prior after one game and it liked Sam Houston far more than SP+ did to begin with, so the 5.5-point gap is largely a disagreement between two priors, not new information. Separately, the flat 2.5 home field is wrong in both directions here: Sam Houston played no true home games in 2025, so anything the rating learned from that season carries no Bowers Stadium signal at all, and this is the venue's reopening. The rating also cannot see that Tulsa is without its best cover corner against an Air Raid that just threw 46 times. Net: I like the dog, but I would size it off the spot and the news, not off the model's implied 5.5 points.

Panel C

Tulsa -13.5

Under 51.5

Our model reads Sam Houston as only an 8.1-point dog after calibration (the raw margin actually favored Sam Houston by 6.3 before the calibration map stretched it); the market sits at Tulsa -13.6, a 5.5-point gap under our own 6.5-point lean floor, so the model itself declines to take a side here. Independent handicapping leans Tulsa anyway. Tulsa's defense held a Big 12 opponent (Oklahoma State) to 10 points and ranks top-40 nationally in scoring defense, while Sam Houston's defense allowed 37.8 ppg in 2025 and already gave up 24 to Troy in week 1. Sam Houston's rush offense (38 ypg, ~131st nationally) makes it one-dimensional, and its -1 turnover margin (89th) against Tulsa's +3 (6th) raises the odds Tulsa gets short fields. Covers shows no confirmed injuries for either side as of 9/10; an AI-picks site's injury claims could not be verified and are excluded from the read. The line has tightened toward -13.5/-14 since opening, not moved off the favorite. Taking Tulsa -13.5 at moderate conviction, not full size, given the size of the number and our own model's lack of edge here.

Total. Our totals model (50.57) sits just under the market (51.4), a 0.8-point gap far below its own 12.86-point signal threshold -- essentially no read either way. Independently, both teams' week 1 games landed under this number (Tulsa-Oklahoma State totaled 34, Sam Houston-Troy totaled 45), and the market total has already drifted down from 52-53.5 at open to 51-51.5 now, with sharpest-book Pinnacle sitting at 51. A fresh NWS forecast (updated today, 9/10/26) shows a hot evening with a real 29% chance of showers/thunderstorms in Huntsville around and after kickoff. Tulsa also converted just 9.1% of third downs in week 1 (133rd nationally), which caps sustained scoring even against a leaky Sam Houston defense. Taking under 51.5 at modest conviction; a lopsided Tulsa win with backups playing extended garbage-time snaps could still push the total past this number either way, so this is a lean, not a strong position.

Where our model looks wrong

The model's own gap (5.5 points) sits under its 6.5-point lean floor, so it produces no actionable signal on this game at all -- and its disclosed weakness list says the calibration map stretches raw margins for big favorites, which is exactly the situation here (raw margin favored Sam Houston by 6.3, calibrated number flips to Tulsa by 8.1). That swing should be treated with real suspicion rather than as insight. Separately, neither the rating nor the totals model has any way to see that Sam Houston's defense, which allowed 37.8 ppg in 2025, opened 2026 by giving up 24 more to a Sun Belt team -- a continuation of a real, ongoing weakness the model can't read until enough games accumulate.

The write-up

The case

Tulsa at Sam Houston, Tulsa -13.5. The number has firmed toward the favorite since the open: DraftKings moved from -10.5 to -13.5 and FanDuel from -12.5 to -13.5 between September 4 and 10, while BetMGM, Bovada and Pinnacle first posted -14 on September 6, right after Tulsa's Week 1 win, and Pinnacle is still there. No reliable current-game ticket or handle splits are available. Tulsa beat Oklahoma State 24-10 on September 5 on only 9.1 percent third-down conversion, helped by four takeaways. Sam Houston lost 24-21 at Troy the same day while outgaining Troy 373-342, with Landyn Locke throwing for 335 yards. One panel notes Tulsa's top cover corner has been out since preseason, though no update confirms his status this week. Our own model's raw read favored Sam Houston before calibration flipped it to Tulsa by 8.1, and the 5.5-point gap to market sits under our lean floor, so it offers no edge. One panel took Tulsa anyway, citing Tulsa's defense and Sam Houston's weak run game. Take Sam Houston +13.5 (+14 at Pinnacle).

Confidence: medium

Three independent analyst panels (A, B and C) worked the same evidence pack and the week's news independently, each returned a side, and each then ranked its five strongest picks for the week. Two of the three panels were on this side, and the third did not rank the other side among its strongest. Medium is a decisiveness label for how firmly the panel agreed, never a win probability: high means every panel was on this side and ranked it among the week's best; only high picks count on the record and go free.

The total: Under 51.5 (low)

Under 51.5. Every book we track has bled this number down since the open, from the 52-53.5 range to 51-51.5, with the sharpest book now at 51. Both Week 1 games finished well under comparable totals, and Tulsa converted only 9.1 percent of third downs in its opener, capping drives. The forecast calls for a hot evening with a modest chance of storms and no meaningful wind. Our own totals model sits just 0.8 points under market, well inside its own noise threshold. The counter-case: one panel leans over, arguing missed field goals and failed fourth-down tries left points on the board and better drive-finishing could lift scoring. Take Under 51.5 (51 at Pinnacle).

Our number

edgelabs EL rating (preseason v1 until a team has played, then the in-season v1 Elo-style refit, Sun 23:00 and Tue 02:00 PT) run through the 2026-09-07 CFB calibration map; home field 2.5. Home perspective: negative = home favored. Ours: Sam Houston +8.1; market at synthesis: +13.6. Held-out 2024-2025 reconstructions (9,927 FBS games): 46-50% against the spread in every gap bucket, calibrated margin error 12.7 points. No against-the-spread edge at any gap size. Treat the number as context, not as a priced edge.

written before kickoff, frozen at kickoff

Trends

Descriptive context, every line with the games it counts. Filters like these did not hold up as predictors in our testing, so none of this is in the confidence read and none of it is a reason a pick wins. Situational splits look back 3 seasons and need at least 8 decided games to be shown at all. Records against the number and on the total join these once our closing-line history is restored; the straight-up splits are live now.

Tulsa
  • won 3 of their last 5 (n=5, 2025 to 2026)
  • 1-10 straight up in divisional games after a straight up loss (n=11, 2024 to 2025)
Sam Houston
  • won 2 of their last 5 (n=5, 2025 to 2026)
  • lost 3 straight
  • 3-8 straight up in night games after a straight up loss (n=11, 2024 to 2026)
The rest of the splits we can compute on this game
Tulsa
  • won 4 of their last 10 (n=10, 2025 to 2026)
  • 2-14 straight up in divisional games (n=16, 2024 to 2025)
  • 1-7 straight up in home divisional games (n=8, 2024 to 2025)
  • 1-7 straight up in road divisional games (n=8, 2024 to 2025)
  • 2-9 straight up in day divisional games (n=11, 2024 to 2025)
  • 3-9 straight up on the road (n=12, 2024 to 2025)
  • 4-10 straight up in day games (n=14, 2024 to 2026)
  • 3-7 straight up in day games after a straight up loss (n=10, 2024 to 2026)
  • 3-6 straight up in home day games (n=9, 2024 to 2026)
  • 6-11 straight up after a straight up loss (n=17, 2024 to 2026)
  • 4-7 straight up in night games (n=11, 2024 to 2025)
  • 5-8 straight up at home (n=13, 2024 to 2026)
  • 4-6 straight up at home after a straight up loss (n=10, 2024 to 2026)
Sam Houston
  • won 2 of their last 10 (n=10, 2025 to 2026)
  • 3-7 straight up in road night games (n=10, 2024 to 2026)
  • 4-9 straight up after a straight up loss (n=13, 2024 to 2026)
  • 4-9 straight up on the road (n=13, 2024 to 2026)
  • 7-9 straight up in divisional games (n=16, 2024 to 2025)
  • 9-11 straight up in night games (n=20, 2024 to 2026)
  • 5-6 straight up in night divisional games (n=11, 2024 to 2025)

The signals and the work

Members

3 independent signals on this game

The signal tally per side, the totals read, every angle with its sample size, and the line movement from open to now. 3 picks are released free on the slate every week. $24/mo, founder rate.

The two numbers underneath

TeamEL ratingConference
Tulsa -6.886 American Athletic
Sam Houston -15.658 Conference USA

Rating difference plus 2.5 points of home field is the rating read above. Full board: the power ratings.

Conditions at kickoff

Partially cloudy. 92F · wind 9 mph · 15% chance of precipitation

Provenance: schedule and finals from the Edge Labs database (2026 season); the line is the latest capture for this game with the book named; our number is the EL rating read (in-season once a team has played, preseason before that) plus 2.5 home field, none at neutral sites, and it is context on this page rather than the call. Signals come from the trend engine, each with its real sample; a tier is how decisively the evidence agreed and is never a win probability. Once the game kicks off this page stops computing and reads our pick lock ledger instead: the pick and the spread we were locked at, written once at kickoff, never updated, graded against that same number. Line movement comes from our permanent capture log, one book across both ends and never a capture taken at or after kickoff, so the second number is the last pregame line and after kickoff it is the close; a game the log has captured only once shows no movement rather than an invented one. Weather is the captured forecast for the venue, and the chip appears only when it is worth saying (wind at 12 mph or more, a 50 percent or better chance of rain, or 35F or colder), never on an indoor venue. Injury counts are our latest daily scan, real report rows only, and a team the scan does not cover is left out rather than shown as zero. The write-up is assembled from those same stored rows, never written around them: the case comes from the lock row, the signals from the trend engine with their own samples, the series from our game database from 2013 forward, and the scoring profiles from completed games only, each with the number of games it averages. A section with no data behind it is left out instead of filled in, it refreshes while the game is pregame, and it can never be edited once the game has kicked off. Trends are descriptive only, computed by the same shared module the write-up uses so the two cannot disagree: straight-up splits come from completed games in our database, situational splits look back three seasons and need at least eight decided games, records against the number and on the total arrive with the closing-line restore, and every line carries the games it counts. Filters like these did not hold up as predictors in our testing, so none of them enter the confidence read. The totals block is a read and not a pick: our number is EL CFB Totals v1 (edgelabs.el_cfb_totals, season 2026, model_version v1, docs/el-cfb-totals-v1.md), a registered candidate that loses to the closing total, so it never posts to the board or the record; the market total beside it is the same capture the spread comes from before kickoff and the frozen consensus the model priced against after it, and the gap is measured against whichever number is printed. The totals trends under it are the trend engine's own rows for this game, deduped on the headline, each carrying its sample. On the games the analyst panel works each week (the games our model prices off the market plus the ranked and Power Four games, at most twenty), the pick and the total call are the panel's synthesis: three independent analyst panels, the same evidence pack, their own research, one side each, tiered by how firmly they agreed, never a win probability. A panel majority on a side is what makes a pick; without one the game is a published no-pick, printed as No pick rather than left blank. Research and context, never a guarantee, 21+.