College Football Misleading Box Scores: Unexpected Results from Maryland vs. SMU, Kansas vs. Houston, More Week 3 Games

If you’ve ever watched any sport, particularly college football, then you’ll understand that the best team isn’t always victorious. That proved to be true in Week 4 as well.

Once you look at the metrics behind the game, you can better understand how the game actually played out rather than just looking at the final score.

This weekly column looks deeper than the basic box score statistics to point out games from the previous week that didn’t end as expected. This information can help us become more informed sports bettors as we get a better idea of who these teams really are.

The must-have app for college football bettors

The best NCAAF betting scoreboard

Free picks from proven pros

Live win probabilities for your bets

DOWNLOAD NOW

Indiana 33, Western Kentucky 30 (OT)

Indiana barely escaped this weekend with an overtime win over Western Kentucky. However, Indiana probably should not have won this game.

Western Kentucky outgained the Hoosiers, 545-484. The Hilltoppers also had a yards-per-play output in the 88th percentile compared to the 48th percentile for Indiana.

Western Kentucky left a lot of points on the board that nullified its advantages. In this game, WKU had the following:

• Turnover on downs at IU 33-yard line.
• Field goal from IU 16-yard line.
• Interception from the IU 9-yard line.
• Field goal from Indiana’s 2-yard line.
• Field goal from Indiana’s 6-yard line.
• Missed field goal from IU’s 26-yard line to send the game to OT.
• Blocked 38-yard field goal in OT.

There were plenty of chances for Western Kentucky to score more than it did, but these long drives ending in turnovers or settling for field goals in the red zone killed the Hilltoppers.

By ESPN's Bill Connelly’s post-game win expectancy metric, this was expected to be a 10.7-point Western Kentucky victory, and Indiana had just an 11% post-game win expectancy.

Instead, Indiana prevailed with a successful field goal in overtime.

Maryland 34, SMU 27

It seems like every metric from this game points to an easy Maryland win. Luckily, this odd result didn’t have any impact on the spread, but it may have repercussions for how the market views SMU or Maryland going forward.

Maryland dominated SMU pretty thoroughly. The Terrapins had a 95th percentile Success Rate and 84th percentile EPA per play compared to 29th and 13th in those two metrics for SMU, respectively.

The Mustangs were outplayed in almost every facet.

Connelly’s post-game adjusted margin for this one came out to Maryland by 21.2, but it just barely held on for a seven-point victory.

Looking at this net Success Rate as compared to other results of the week, as Parker Fleming visualized on Twitter, this game was in the same range as other blowouts like Baylor (35 points), Washington State (31 points) and Iowa (27 points).

Penalties killed Maryland. It lost 7.59 EPA on penalties, while the Mustangs gained 9.61 EPA on penalties. 

As Maryland travels to Ann Arbor in Week 4, this may not be the time to do it. But keep in mind going forward that Maryland has looked strong against solid opponents and may be underrated by the market.

Kansas 48, Houston 30

Kansas is 3-0 for the first time since 2009.  It appears that Lance Leipold has the program headed in the right direction. There rightfully should be a ton of excitement around Kansas, but there's a little cause for concern moving forward.

Kansas won this game by 18 points, but it was actually expected to be a 9.7-point win, per Connelly's metrics. Two Houston turnovers led to 14 points from the Jayhawks and caused the box score to look a bit more lopsided than it should have.

By net Success Rate, this would have been expected to be a close Houston win.

Both teams were very similar in most metrics. They had virtually similar Havoc rates and third-down efficiency metrics. Houston had the edge in Success Rate (49%-45%), while Kansas had a slight edge in yards per play (6.57-5.69).

I’m not saying Kansas shouldn’t have won this game, but it was a much closer contest than the final score would indicate.

It seems the market has already overreacted to this game. Kansas is currently a nine-point favorite against Duke, while we make it a 7.2-point favorite and SP+ projects it to win by 7.3. Houston is favored by 17.5 over Rice, but we make it 22.2-point favorites and SP+ has it favored by 20.1.

I’m not making a bet on either of these sides yet based on just this, but it would be something to monitor as the week goes on.

Bowling Green  34, Marshall 31 (OT)

After its shocking upset of Notre Dame, Marshall losing to Bowling Green as a 17-point underdog this week was almost equally as surprising. The game went to overtime and appears to have been close, but Marshall should have won handily.

The Thundering Herd outgained Bowling Green, 547-377. Marshall averaged 7.29 yards per play to just 5.09 on the Falcons’ side.

It lost this game because of the turnover battle. The Herd turned it over three times as opposed to zero for Bowling Green. Each of these costly mistakes occurred in Falcons’ territory, including a lost fumble at the 4-yard line.

This was a three-point Bowling Green win, but it had just a 14% post-game win expectancy. The expected margin for this game indicated a 9.2-point advantage for Marshall, per Connelly.

Rice 33, Louisiana 21

Looking at the advanced box score of this game, you would have expected a low-scoring game and the under 52 to cash easily. Both teams posted Success Rates at or below the fourth percentile and EPA per play figures below the second percentile. 

I’m honestly not sure how this happened.

Almost every score in this game came from a long touchdown drive as opposed to anyone settling for field goals. ULL did also score on a pick-six to push this one over.

If you were on the under in this one, you probably had the right side. But unfortunately for your ticket, things did not break your way.

The ultimate CFB betting cheat code

Best bets for every game

Collin Wilson's biggest weekly edges

Profitable data-driven system picks

TRY FOR FREE