Loyal Cougars
Home / Football / By the Numbers: 2016 Season Preview

By the Numbers: 2016 Season Preview

As we prepare for the kickoff of BYU’s 2016 season, there are more unknowns surrounding the program than any year in recent memory. For starters, the Cougars are implementing new systems on both sides of the ball. The amount of staff turnover is staggering: of the 22 coaches and support staff involved most directly with BYU football, only four had worked in the building prior to 2016 (Russell Tialavea, Steve Kaufusi, and GAs Jan Jorgensen & JD Falslev). Next, we have the shift in defensive scheme with its attendant changes in personnel groups. Moving from a 3-4 base to a 4-3 has unsurprisingly left the Cougars with excellent depth at Linebacker, but a shortage of bodies along the defensive front (compounded by the uncertainty surrounding the timing of Travis Tuiloma’s return from Lisfranc injury).

With so many question marks, it’s not difficult to understand why the prognosticators are having a difficult time reaching a consensus in their predictions for BYU in 2016. ESPN’s Football Power Index predicts a 6.4 win regular season. Brian Fremeau is a bit more bullish on the 2016 Cougars, predicting 8.4 wins on the season. Looking at the differences between these two prediction models reveals a few striking differences, particularly in the September games. Whereas the ESPN model gives BYU a 38% chance of winning the opener in Arizona, Fremeau puts BYU’s odds at 75% and predicts a 33-23 win for the Cougars. That’s quite the variation: 37%. Fremeau and ESPN also see the Cougars’ chances against UCLA differently with ESPN giving BYU a 36% chance to win and Fremeau forseeing a statistical toss-up with BYU at 50%. And the two models vary even more in their predictions for BYU’s neutral site game against West Virginia: ESPN gives the Cougars a pessimistic 26% chance of victory, while Fremeau predicts a 64% likelihood of a BYU win.

In putting together my game-by-game probabilities for the 2016 season, I found that while I have a slightly rosier view than the ESPN index, I’m not quite as optimistic as Fremeau’s predictions. I tend to fall somewhere between the two indices. And my predicted win-loss record falls between their two extremes as well. I’m sure we could nitpick that some of my game-by-game odds could be higher or lower, but for the most part, this is where I would put BYU’s chances of winning each games based on what I’ve heard and read this offseason and based on what I’ve observed at BYU practices:

RecordCalculator

At first glance, it may seem that based on these probabilities, I would predict a 7-5 regular season for Kalani Sitake’s debut season. However, while a 7-5 season is the most likely individual outcome (25.6% chance based on my game-by-game odds), the model actually leans toward an 8-4 season. And before you ask, yes, I’m a statistics nerd and I built my own database that runs all of the scenarios to spit out the odds of each game being won or lost and then provide the total chances for each win total. Based on my model, here are the chances of BYU finishing the regular season with a given number of wins and the total odds that BYU finishes with at least that many wins:

RecordCalculatorWins

So if you’re banking on an undefeated regular season for BYU, I’m sorry to be the one to tell you this, but assuming my individual game probabilities turn out to be anywhere close to accurate, then you’re probably going to be disappointed. My model gives BYU about a 1-in-1,000 chance of putting together an undefeated regular season. Many pundits surmise that with the strength of BYU’s 2016 schedule, even a 1-loss team may have a shot at the College Football Playoff, provided there aren’t multiple undefeated teams left at that point. But, even the possibility of a 1-loss regular season would only be about 1-in-100 based on my probabilities.

The question is: what should we expect for BYU in 2016? In an interview on BYU Sports Nation on Monday, Trevor Matich set the bar fairly low by saying that “7 wins would be a phenomenal season against this schedule. I think 8 wins would be beyond phenomenal. And 9 wins would be a dream.” Here are the words I would use to describe each regular season win total for the 2016 BYU football season:

  • 5 wins or fewer – Inadequate: Based on my model, there’s a 88% chance that BYU will finish the regular season with enough wins to qualify for a bowl game. I think it’s fair to say that it would be a disappointing season if the Cougars are forced to stay home in December.
  • 6 wins – Acceptable, but somewhat disappointing: If BYU ends the season with 6 wins (88% likely), I’ll feel that this team’s potential was not fully realized, but I think 6 wins and a trip to the Poinsettia Bowl would be acceptable given all of the scheme/coaching changes and the difficulty of this year’s schedule. That level of acceptability would only increase if we find that BYU had been hit by injuries to key players or position groups.
  • 7 wins – Expected; Satisfying: I think we will all be satisfied if BYU ends the year with 7 wins. This outcome has a likelihood of 69%, which means that in 2/3 of the season simulations I ran, BYU finished with 7 wins or more. And while I think we’ll all be happy with that outcome, I wouldn’t go as far as Matich in proclaiming it “phenomenal.”
  • 8 wins – Nice; Exceeded expectations: BYU has about a 43% chance of winning 8 or more regular season games in 2016. Based on the simulations, an “average” season (half of the results are better and half worse) would yield about 7.7 wins. As such, I would be excited to finish with 8 or more wins. And I think it would build a lot of momentum towards Year 2 of the Sitake era in 2017.
  • 9 wins – Great: I would be ecstatic if BYU manages to win 9 or more games this season. According to my model, there’s only about a 1-in-5 chance that we’ll see a win total that high. A 9-win regular season would likely still mean a trip to the Poinsettia Bowl, but would certainly represent an improvement over 2015’s 9-win campaign with the increase in schedule strength. It would mean that BYU is on the rise moving forward.
  • 10 wins – Among the best BYU teams ever: There’s only a 6% chance that the 2016 Cougars will put everything together well enough to win at least 10 games against a daunting schedule. It would mean only two losses against a slate of teams including Michigan State, Boise State, UCLA, Utah, Arizona, West Virginia, Mississippi State, and Cincinnati. Those are 6 quality wins and you have to figure that BYU would be ranked in the Top 25 and at least in consideration for a NY6 bowl game. And if that were the case, you’d have to rank a 10-win season among the top 5 or 10 greatest seasons in BYU history.
  • 11 wins – Dream; Possibly the best All-time BYU team: As mentioned earlier, if BYU were to run through this gauntlet of a schedule and emerge with only one loss, you’d have to figure the Cougars would be a shoo-in for a New Year’s Six bowl and potentially in the CFP conversation. And if that were to be the case, the 2016 team would be in the conversation to be the best team in the history of BYU football. A 1-in-100 season would certainly be the dream scenario for 2016.
  • 12 wins – Impossible; The best All-time at BYU: If BYU were to pull off the unthinkable and finish the season with 12 wins, a trip to the college football playoff would be the likely outcome, barring an absurd shun by the Selection Committee. And just having the opportunity to win a National Championship on the field would rank the 2016 team as the best ever to play at BYU.

So, what’s my prediction for 2016? I see an 8-4 season. I think BYU will go 4-4 in its first four games, winning: at home against Toledo, one of the neutral games, and two of Utah/Mississippi St./Boise St./UCLA. But I see the Cougars finishing strong with four November wins heading into the Poinsettia Bowl (a likely win against a Mountain West opponent) and into the 2017 season. What are your thoughts? Join the discussion by adding your predictions in the comment section below or on Cougarboard.

3 Comments

  1. David Strong

    September 2, 2016 at 11:15 am

    Brandon,

    Tell me about he inputs you used in your model to determine winning percentage projections. I have my own algorithm that I have used to create a program that simulates BYU football games, and while the total number of wins might be similar, the ranking of most difficult games is somewhat different.

    While our models for projecting the outcome of the games are doubtless different, I would assume that given the same inputs, they would yield similar out puts, so what inputs do you feed into your model?

  2. Brandon Jones

    September 2, 2016 at 11:51 am

    I don’t use the model to predict the winning percentages. Those are the inputs based on my own feel. My program only gives the likelihood of each win total based on the given percentages.

    For example, if you played two games and had a 25% chance of winning the first and a 75% chance of winning the second; the odds would be: 18.75% chance of winning both (25% x 75%), 18.75% chance of losing both (75% x 25%), and 62.5% chance of winning one of the two (25% x 25% + 75% x 75%). If you extrapolate this same method out for 12 games, that’s what my model does. It gives the exact chance of winning X number of games based on provided game-by-game odds.

    Obviously a lot of factors play into the game-by-game percentages, but I don’t have anything mathematical, just a gut feeling based on my own weighting of dozens of factors.

    I’m curious though, what are your game-by-game percentages? I could even put them into my model and compare the final results with my own.

    • David Strong

      September 2, 2016 at 8:58 pm

      What my model does is simulate the most likely outcomes of BYU football games based on prior statistics with a lot of modifiers and calculations to make their stats mean the same thing. For example, a QB from the MAC with identical stats to one from the SEC would have markedly different values attached to them.

      The values for the different players and position groups are ranges instead single values. The more uncertain a player or position groups likely performance the wider the range. each simulated game pulls a random number from within that range and uses that to represent that player or position group in that particular phase of the game.

      I run the simulation 1000 times. From those results I can gather a projection of how the simulation thinks the games is likely to look like. I submit the Write up to Torn By Sports. For the Arizona projection go to:

      http://www.tornbysports.com/simulations-show-byu-beating-arizona/

      This simulation does not yield a “percent chance of victory”. It can tell me what percentage of the 1000 simulations BYU wins, but because the simulations simulate the most likely outcome, the larger the projected margin of victor, the more over exaggerated the % chance of victory becomes. For SUU, USU, and UMass, that percentage is 100%.