The models’ projections of the 2018 B12 season

Everyone loves models these days. Not the kind that post pictures of themselves on Instagram and somehow derive a career from that enterprise, but the kind that forecast upcoming events in complex systems.

For the most part, the good models are awesome and super useful for creating a picture of likely outcomes. My man Bill Connelly at SB Nation has a model built out of S&P+ and recruiting rankings that I’ve typically found to be very interesting and solid. Here’s how his model did the final 1-130 ranking.

Here’s how the Big 12 shook out:

1. Oklahoma (9th)

2. Oklahoma State (19th

3. TCU (22nd)

4. Texas (27th)

5. West Virginia (43rd)

6. Iowa State (46th)

7. Texas Tech (47th)

8. Baylor (50th)

9. Kansas State (61st)

10. Kansas (103rd)

As you can see, this more or less mirrors most other projections out there with the top seven or eight clustered within striking distance of each other before things fall off with Baylor and Kansas.

The big exceptions are Kansas State (always) and Bill’s model actually likes Iowa State more than most other folks out there. His model looks at S&P+ performance over rolling periods of time combined with recruiting rankings.

Here’s what the league looks like in terms of the composite recruiting average over the last five years along with the win-loss record in league play over that time:

From this starting point, which most everyone has in the back of their minds if nothing else, you’d tend to rank things more or less as Bill’s model did. Oklahoma is the clear no. 1 followed by TCU and Oklahoma State in some order and then Texas matched with West Virginia followed in some order by K-State, Texas Tech, and Iowa State.

Obviously Baylor would appear to be a solid looking team but their roster has been remade by various forms of attrition and addition and their numbers and experience level are not where the other .83 to .86 rated rosters are at.

Then there’s more recent events, which can both cloud and sharpen your perspective on this. For instance, both of Matt Campbell’s classes were rated higher than any of the previous three and Campbell is 7-11 in B12 play after going 5-4 in year two. Alternatively, Oklahoma State is 20-7 over the last three years but accomplished that with Mason Rudolph, James Washington, and Marcel Ateman on the field (except in 2016 when Ateman was out) and all three are now in the NFL.

Other factors that you’d want to take into account to get a glimpse into how this season could go:

-Recruiting ranking for the actual, prospective two-deep. 

-Returning an accomplished starter at QB?

-Number of returning starters?

-How many prospective starters have been in the program for three years or more? (Overall XP level)

The good models use something like W-L record though because then you can see how well the coaches are developing players, plugging holes, and deploying them on the field. Since Connelly uses S&P+ he’s got even a leg up there because he’s measuring output more precisely.

Measuring the levels of talent that come into a program is hard although service rankings offer a solid start. I think there’s a lot of problems with the rankings and the reasoning that the best teams are assembled out of grabbing the most NFL-looking talents at every position. If I were trying to build a model for every team though? I’m sure I’d use the composite rankings as well.

Another approach

The goal of a model is to try and isolate the main factors that can point to big picture trends. Like anyone else with half a brain, I tend to at least glance at these models that are trying to filter out the pure data of a complex system.

But it’s still that, a complex system. Something as small as the health or development of a QB can have a huge ripple effect across the system and it’s hard to get that in the raw numbers. Some schools regularly have good QB play, others are at the mercy of violent swings in caliber of play. Greg Davis had QB play pretty well figured out at Texas, then Mack told him to change up the system and it all collapsed.

I like to try and get a feel for what each team will feature as their main identity and how likely it is that they’ll find winning formulas on either side of the ball in order to win. That requires trying to process a ton of information, much of which becomes negligible after a few weeks because a key player or two in particular teams goes down with injury and they have to go into scramble mode to find new formulas for winning team-ball.

The big challenges with prognostication in the B12 are that…

There’s not much separation between the middle five teams.

After Texas and Oklahoma and ahead of the Kansas schools and maybe Iowa State, most everyone recruits at a comparable level. What’s more, they all have very effective coaching.

Gary Patterson, Mike Gundy, Bill Snyder, Matt Campbell, and Dana Holgorsen are all basically institutions at their respective universities. Holgorsen lags behind the others since WVU had their greatest run under RichRod but he’s pretty well established these days. Campbell hasn’t been around long but has already outperformed the last few decades of Iowa State coaches. Art Briles was the greatest (by on field results) coach in Baylor history and it’s highly unlikely that will change anytime soon.

Kliff Kingsbury and David Beatty don’t measure up and consequently both will likely be gone. Texas and Oklahoma are in a different class and are hoping they’ve have nailed down their leadership for the next decade.

Beyond that, the middle six all recruit the same turf with very little unique recruiting ground. Oklahoma State tends to hit San Antonio, Tulsa, and OKC more, Texas Tech theoretically has better access to West Texas, everyone wants to build connections in Louisiana. Beyond that, it’s Houston, DFW, East Texas, and Central Texas in that order for everyone.

The difference between Texas/Oklahoma, and the middle five has been negligible this decade

I know everyone will say, “hasn’t looked negligible with Oklahoma winning the last three titles!” but that ignores the struggles the Sooners had earlier in the decade with Baylor, OSU, and TCU as well as the fact that if Baker Mayfield hadn’t quit Tech and walked on it’s possible that none of that would have happened.

Texas has continued to recruit the highest rated classes but a lack of stability in the leadership either of the program or even within the team has prevented that advantage from ever materializing.

In a year where either OU doesn’t “have it” or someone in the middle five (or even the Midwest trio) is lucky with injuries and has veteran playmaking at QB and/or can play D…the race is on.

There’s usually iffy depth at every program

Texas and Oklahoma are the only programs that recruit well enough to expect to regularly have underclassmen that can compete for starting spots and be impact players. Most everyone else has to snag developmental prospects and develop them over time. That means that in a year where you don’t have a high hit rate, or otherwise have some kind of attrition, you can be vulnerable.

Most every team in the Big 12 has a few position groups where if they take an injury or two, their level of play stands to go over a cliff. This has been true of Texas and Oklahoma as well of late, particularly on defense for Oklahoma and offense for Texas.

Since it’s hard to prognosticate injuries, that means that the wild swings that occur over a season can wreck any projections.

The talent in the Midwest is not flashy and is consequently underrated regularly

This is where I think everyone is missing their marks in prognosticating the league race. The recruiting rankings are geared largely around the known talents at obvious positions like RB, QB, WR, etc. The guys that run at SPARQ events, compete in 7on7 events, show up at camps, and who have flashy HUDL film. The services obviously take line play seriously and will give big ratings to the big men who show major talent, but it’s a hard position to scout or be aware of and you can tell that there’s little market or incentive to due diligence in Wichita suburbs.

Kansas doesn’t seem to know what’s up and hired David Beatty to benefit from his recruiting connections in all of the normal hot spots. Meanwhile Iowa State is hoping to make a breakthrough in run blocking while starting four Iowans on the line this year the other guy is from Wisconsin. Go look at the better K-State OL of the Snyder era, such as the 2003 unit or this current squad, and guess what you’ll find?

You’ll find guys from Kansas, from the local JUCOs, from Colorado, and from Texas. Then their FB/TE/LB infrastructure positions are also typically stocked with locals.

What’s the difference between a team with solid line play and 4-star skill talent and a team with really strong line play and 3-star skill talent? Generally the latter wins and wins big.

What’s the difference between a team with good skill talent and speed on defense and the team with smart, stout LBs and Ss that consistently show up in the right spots with the right temperament? The latter wins and wins big.

The Midwest teams are often better at fielding consistently stout, tough teams with the kinds of kids they find locally. From there they can and need to bring in speed from Texas, California, or wherever they can find it. This model was well established by Tom Osborne at Nebraska and it’s been successful at multiple other places now as well.

So, when you look at the question marks with this Kansas State or Iowa State team, you tend to find some telltale signs.

They each have veteran play-making at QB.

Their both reasonably well stocked at the positions where you need speed and athleticism to see impact. K-State has three starters back in the secondary with a fourth guy that played a lot and played well in 2017 (A.J. Parker), they have some talent at RB, and they might be back to having pass-rush at DE with Reggie Walker now older and healthier.

Iowa State has tons of premier talent at WR and RB even after losing Allen Lazard and their questions on D regard positions like run support safety, strongside DE, and middle linebacker that Midwest teams typically don’t struggle to find. The question is literally, “will this Midwest team be able to field tough, cohesive OL and find an inside-backer for this season?” They’ve not done the former yet, but it’s hardly a stretch to imagine they could manage it in year three.

All that points to a season in which Iowa State and Kansas State are more likely than not to meet the criteria of teams that are on a level field with the middle five. When you consider the lack of veteran playmaking at QB across that swathe of teams combined with the relative weakness of OU and TX, you actually have a firm case that these two teams are as legitimate of contenders as anyone else in the league.

That’s how I see things heading into the year. The two perennially forgotten teams of the Big 12 are well poised to seize on the vacuum left by all the departing QBs and show better than five year recruiting or results would suggest.