Advanced Metrics Still Like Oklahoma State Football In 2016

This is a guest post by friend of the blog Justin Packard. It’s a good reminder that not all was lost last weekend in Stillwater. 

Oklahoma State entered this season ranked eighth in ESPN’s Football Power Index (FPI) preseason rankings, its highest preseason rank since ESPN expanded the FPI to the preseason.  Wondering if the FPI really just likes the Cowboys, or if we should be expecting another big season, I decided to look into how accurate the FPI preseason rankings have been over the past few seasons.

A little bit of background on the FPI rankings. The FPI is a predictor measurement comparing the relative strength of a given team to an “average” team.  In this year’s preseason rankings, OSU has an FPI score of 19.3 meaning that the Cowboys would win by 19.3 points when playing an “average” team. The FPI is calculated using four metrics: previous performance (with the most recent season counting the most), returning starters, recruiting and coaching tenure.

Let’s get to it.

Oklahoma State started the 2014 season at 27th in the FPI rankings.  FPI predicted the Cowboys would go 7.1-4.9.  The Cowboys finished that year at 6-6 (FPI does not predict bowl games, so all records in this post include only regular season and conference championship games).  An interesting note from 2014: in the preseason, FPI had TCU and Texas Tech being equally good, both at a 12.6.

The 2014 preseason FPI expected OSU, TCU and TTU to all have similar seasons. (Photo courtesy: ESPN)

This shows one of the major pitfalls of the preseason FPI. Because the major contributor in the calculation is previous performance, preseason rankings cannot take injuries into account or, in the case of TCU, Heisman finalist performances.  TCU ended 2014 ranked third in the FPI and first among Big 12 teams.

Oklahoma State dropped to 48th by the end of the 2014 season, with an FPI of 6.5.  Although a drop of 21 spots seems extreme, the FPI overestimated the Big 12 in the preseason.  Six Big 12 teams dropped in the FPI rankings, five of which dropped by at least 15 spots.  Among Power 5 teams, the Big 12 had the highest average ranking change, losing an average of 5.8 spots.

This shows the average rank change for the Power 5 conferences from the preseason rankings to the end of the season. The Big 12 dropped nearly six spots on average in 2014, but gained more than seven in 2015, the biggest change of any Power 5 conference.

Contrast that with the 2015 season, where Big 12 teams gained an average of 7.2 spots. With the Big 12 underperforming in 2014 according to FPI, it made it easier for the teams to get looked over in the FPI in 2015. Only two teams from the Big 12 fell in the rankings that season, OU and WVU. It all goes back to the first of the four metrics. The previous seasons performance affects FPI the most in the preseason rankings.

Oklahoma State started 2015 ranked 24th and that is right where it ended up at seasons end.  FPI predicted the Cowboys to go 7.5-4.5.  It wasn’t until week 9 that FPI finally projected the Cowboys to get to 10 wins. OSU finished the regular season at 10-2.

With that kind of performance, it means the Cowboys were bound to end up ranked higher to start this season compared with the last two. With a large number of returning starters, the Cowboys got boost and a consistent coaching staff helped gain a top ten FPI ranking to start the 2016 campaign.  

So, can we expect the Cowboys to end the season ranked in the top ten at the end of the season? The FPI preseason projections get the W-L record to within 1.5 games around 60 percent of the time.  That would put the Cowboys anywhere from 7.8 to 10.8 wins.  

The top 10 of the preseason FPI rankings. (Photo courtesy: ESPN)

The game against Central Michigan gave us a good indicator of the Cowboys in the FPI because the Chippewas started the season close to average, scoring a 0.4. The loss dropped the Cowboys to an FPI of 16.0 and out of the top 10, coming in at 15th. The projected W-L moved to 7.9-4.1.

FPI projects the Cowboys to get eight wins after the performance against Central Michigan and has OSU ranked as the number two team in the Big 12.

But remember the FPI only takes what it knows. As long as the Cowboys don’t have any major injuries, FPI will continue to favor them. After all, OSU dropped only six spots after last weekend’s debacle, keeping them in the top 15 whereas the AP and Coaches Poll both have the Cowboys unranked.  

The next three weeks will be crucial for the Cowboys.  Pittsburgh is currently ranked 33rd with an FPI of 10.3. For comparison, Central Michigan is ranked 46th after beating the Cowboys, with an FPI of 6.1. Baylor is ranked 26th and Texas is 21st.  Strong performances to end the non-conference part of the schedule and open up Big 12 play will go a long way in helping OSU get back into the FPI top ten.