How Do The Computers Have Maryland Football Finishing

facebooktwitterreddit

As Maryland heads into Big Ten play, there are a lot of questions about how well they’re going to fare in what is historically perceived as a tougher conference than the ACC. While Maryland has had a solid season up until this point, their best win was against a 2-2 Syracuse team, and while they did hang in there with West Virginia, it is still a loss.

We can make all the eyeball predictions we want, but it does help to have numbers confirm our observations about a team. Before the season began, I felt the Terps were a 7- or 8- win team, and after watching them through four weeks that feels comfortable. But do the computers still have Maryland in line with being a dark horse Big Ten contender? Let’s see what a few have to say.

(For all predictors, you can find them here)

Massey Rating with Monte Carlo Simulation (remaining games):

The Massey rating happens to be one of the computers that the BCS used to use to gauge teams competency on the field. In the past it’s been both scarily accurate and somewhat wrong, but by and large it is a decent indicator of probabilities. Massey has Maryland ranked #54th nationally, and against Indiana they’re considered an underdog, so it’s safe to say their assessment is bullish.

As for remaining wins, using the Monte Carlo simulation:

WinsPercentCDF
02.9802.980
112.53715.517
223.43538.952
326.55065.502
419.88385.385
510.26095.645
63.50299.147
70.78199.928
80.072100.000

Final record: 6.8 – 5.2

The most likely scenario for this simulation is that the Terrapins finish with two or three more wins. That’s a pretty low number given how many games are remaining, but this isn’t opinion; it’s just simulated numbers. Presumably, this would make the Iowa, Indiana, and Rutgers games of the utmost importance should Maryland want to go bowling the easy way.

ESPN Football Power Index:

The ESPN Football Power Index is also controversial, but has accurately guessed the outcomes of the national championship game (and Maryland’s win total) the past two seasons. It also adjusts week-by-week. In terms of accuracy, it hovers around 75%, which for a prediction machine isn’t too bad.

For the Terps, the final record and their overall stats are below:

Most Likely Final Record: 6.9-5.2

Odds To Win Conference: 1.8

Remaining Strength of Schedule: 24th

Payne College Football Ratings:

Payne College Football Ratings System is similar to Massey in that it just throws a ton of variables and numbers into a machine to get its final computation. As a personal preference, it’s not as accurate as Massey (or even TeamRankings.com), but it is absolutely worth its salt as a predictor ~75% of the time. With that said, Payne’s prediction model:

Win-Loss Record: 6.8-5.2

—————

You get the gist. Most every site that offers predictions pegs Maryland in the 6-7 win range, with very little variance. Coach Edsall is 6-20 in games played after October 1 during his time at Maryland, so there is a trend going that the Terps win early and fade a bit down the stretch. Last year Maryland defied that trend, but they also got some pretty easy games down the stretch in the ACC (a 7-5 Hokies team and a 3-9 Wolfpack team).

They won’t really be afforded that this season in the Big Ten. They have a top 25 schedule to finish the season, and are going to be underdogs in at least five (likely six) of their next seven games. That alone makes these predictors that are bullish on Maryland hold some weight. Personally? I think they’re a little too negative on the Terps. The defense isn’t nearly as bad as the yardage it has sacrificed suggests, and the offense is better than the computers will give it credit for.

All that said, Maryland needs to win this weekend if they want to take the easy (or easier) route to bowl eligibility.