Home » BYU Football Articles, DSB Analysts' Articles, Featured, Headline

BYU’s Experience a Predictor of a Big Season?

28 August 2012 Louis E. Deaux 19 Comments

BYU fields a highly experienced team in 2012.

Louis E. Deaux is a contributor to Deep Shades of Blue. Deaux is a member of the Football Writers Association and has covered college athletics for decades, including writing for the Football News for many years, covering the WAC and MWC. —

Statistics and mathematical patterns have always intrigued me.

As a professional that works with numbers and geometry in multiple dimensions, my mind tends to wrap itself around problems that require the methodology of analysis most often associated with changing forces.

Calculus is of course the branch of mathematics that deals with changes in forces and the outcome of forces as conditions change. For this reason, I am always intrigued by the way in which different mathematicians approach the otherwise difficult to quantify value of an entity — any entity.

College football is a wonderful example of fluid entities that change in time for a variety of measurable reasons: coaching changes, talent levels, the strength that comes to a program from simply winning, and the negative impact of constantly losing. Football Teams are an assemblage of minds and bodies for the purpose of achieving victory. However every team is different. Every team changes in time.

In my roughly three decades of watching, analyzing and evaluating football; after the many interviews with coaches, players, and other writers or analysts, I’ve come to one conclusion. The MAXIMUM level of performance any team is likely to achieve can usually be predicted and is based on the cumulative value of all related EXPERIENCE factors.

That performance quotient or predictor is quantified as a projected WON/LOSS record. In other words, the value and weight associated with the EXPERIENCE LEVEL of all components that comprise a team will ultimately predict it’s season outcome.

Of course, the use of the variables that comprise the algorithm predictor for “experience” are not going to be an individually accurate for every single team over the course of the year. Some teams will suffer inconsequential losses of personnel while others will suffer devastating losses of key players, or team members. The death of a key player’s family member may not be predictable, but a negative can affect a team and change anticipated high outcomes. Those predictors are difficult to quantify because they are very random.

Suffice to say, some teams ranked very high, may be over-ranked based on their overall experience rating going into a season. Some teams may be ranked way too low based on the same factors.

The purpose of this exercise was to look at the experience factors under the weight of each added together and then used to devise logical EXPERIENCE VARIABLES from which to construct a rating system:

2012 NCAA Team Experience Ratings  — see the results (click on the Experience Ratings again on the next page when it opens)

V1. Returning Starters divided by 22 = a quotient – The higher this number, the stronger a starting lineup a team has going into a season.

V2. Returning Letterman = a solid variable. ALL teams usually lose a few players from this total between spring and fall for various reasons. Inasmuch as that is universally accepted to be commonly the same for all teams, the loss is ignored in the evaluation process. Keep in mind, the loss of 3 players during that time frame for a school that has 50 returning lettermen is far less impactful than it would be to a team that only returns 35 lettermen.

V3. Returning Quarterback – If yes, the Returning QB is worth a factor of +10 in the algorithm. Without a starting QB, teams are much more likely to struggle. Rarely does a new starting quarterback come into a system and have an immediate positive impact. The learning curve is steep at the major college level and so the new quarterback will result in a negative impact = -10 points.

V4. Winning is important. Teams that experience the thrill of competing and winning get a solid mental mindset for winning in the future. Winning in the recent past can provide a confidence level that is translated into play on the field. A winning history can become a powerful force for predicting how a team views its aggregated capability.

The win/loss aspect of the predictor directly relates to the collective psyche of an organization. You win most often because you expect to win. The positive value of a winning tradition in the prior year is worth twice its face value as a positive number. A ten win team is therefore worth 20 experience points to the plus side.

V5. Losing is predictive and more devastating to team psyche. When teams perpetually lose, they fail to connect their effort with progress that can eventually lead to winning. Negative thought is significantly more powerful to persuade failure, than positive mental attitude is to lifting a team to wins.

Losses can result from lots of losses, not just on the field, but the losses that also result from individual failures, injuries, and arrests, anything that can instill in an organization a sense of failure. A negative loss number is worth three times its face value. So the 10 win season that lost three regular season games loses 9 basis points. The sum of V4 + V5 for a 10-3 Team is equal to +20-9 = 11 positive points. That is a solid predictor that a team will do well simply because it’s prior year was successful.

V6. Talent matters. – Upper echelon BCS and BCS like programs (Note Dame, BYU, Boise State) have talent, some more, some less, but these teams do rise to the top. Talent can be defined as the developed potential of the recruiting classes. It’s futile to use Scout, Rivals or other so called scouting services because they do not really evaluate all incoming FBS level talent.

We won’t get into that in this article, but suffice to say, the teams with great talent do usually succeed. But not always! UCLA , Texas and Notre Dame have for example done far less with great talent in the last decade than say TCU, Boise State and Georgia Tech with purportedly less talent.

A Talent quotient is important, but should not be over-valued. A strong experienced team with less purported talent can lay the wood to the team that is supposedly a lot more talented (BSU demolishing Georgia and Virginia Tech) in recent seasons is a perfect example, or Utah over Alabama in the Sugar Bowl a few years ago.

V7. Coaching and Staff experience is huge – The more experienced a staff is, the more it can relate to the experienced players. New staffs are even occasionally resented and so it takes time for a staff to mesh with a group of players. Staffs change for many reasons. The point is, the value of staff experience is more important than talent.

V8. Prior Year Schedule – The stronger a schedule, the more valuable the success experience. This factor should not however be over-valued. But there is no question that LSU gains a great deal of strength because it returns a lot of players that experienced success against an SEC schedule as well as an out-of-Ccnference (OOC) schedule that included Oregon and West Virginia. When a team has success against really good competition, it strengthens the overall experience value for returning starters and back-ups who have lettered.

V9. Regardless of all other factors, teams must play a new schedule. Playing at home a lot means more to experienced teams. Couple experience with home games and you get an environment that is guaranteed to promote success. One more home game can often offset the loss of one key player. Therefore, it is important to factor in the current schedule to the extent a team gets to appear before the home crowd. A six game home schedule is neutral because the team must also travel six times. A seven game home schedule is worth two extra starters. An eight game home schedule is worth 4 extra starters. Home games really are that valuable.

I did not factor in the value of where teams travel. It should be noted that teams which do NOT travel very far for their few home games are often over-valued or given too much credit for schedule difficulty. While I did not take those factors into account, it is meaningful psychologically and physically on members of the traveling team.

Alabama will only play four road games and none are more than a long bus trip or hour plane flight away. Like Alabama; LSU will never leave the south, never play at altitude, it will never experience driving snow or the bitter cold. We think of LSU and Alabama as power teams, but in some ways, they and other SEC teams almost never prove themselves the way Michigan, Nebraska, Ohio State, BYU, UTAH or Wisconsin do in late November.

They will not be tested playing at nearly a mile above sea level where the oxygen content of the air is 14% lower than at sea level. They experience humid heat, but do not know the difficulty of desert; an oven-like dehydration and 110 degree heat. We think of SEC teams as being formidable, but they really play most games in friendly venues or travel very little, and rarely experience some of the more difficult natural elements of weather and geography that other teams see year in and year out.

With that, I constructed the algorithm to produce a new ranking based on the revised W/L record predictor. The formula weighs all factors appropriately as follows:

[(V1/22 x V2) + V3 + (V4 x 2) + (V5 x 3) + (V6 x 3) + (V7 x 4) + V8 + v9] = ΣtER [Total Experience Rating} = ΣtER /12 (games) = WP (Win Predictor) Note: teams with a 13th game at Hawaii were not adjusted in.

This number is rounded to the nearest whole integer because you cannot really win a fraction of a game. The number of predicted Regular Season Losses is the difference between 12 – WP. Teams with the same W/L record are ranked on the basis of Score such that the team nearest the beginning of that WIN LEVEL is most likely to win that many games, and the one at the bottom is least likely in the group.

All things being equal, the one closest to the top of a WIN LEVEL is also more likely to over-achieve and win more than predicted, and the one at the bottom is most likely to win less. Notice that the Predictor Record W/L percentage is not exactly based on the whole integer .pct. This value is based on the projected non-rounded record so that each team can be sorted based on which is by record more likely to hit that projected value and which is less likely.

Also, it is important to realize that a predictor is based on the best case result for any given team. Again, any significant personnel losses can change the outcome. Inasmuch as those are difficult to predict from one team to another, the reality cannot be predicted. Utah returns a starting QB in Jordan Wynn.

But if he’s lost due to an injury, the highly regarded UTES that rank 10th in the “revised expectation” would drive a team from +5 to -5 in the QB column. That would alter UTAH’s score from 117.648 and a 10-2 regular season projected record to 102.648 and a 9-3 record.

Some Surprising Results

With that, we notice some immediate trends. Some teams that are projected to finish very high by pollsters do not project as strong based on the overall experience factor.

USC is loaded with talent and has a plethora of returning starters, and a Heisman Candidate at Quarterback, but is thin. Years of probation have left the Trojans with fewer experienced athletes behind the front line starters and that will haunt the Trojans as they get into the meat of their very difficult schedule.

They are ranked #1 in the Polls, but the experience predictor says they will do no better than 9-3 in the regular season. A 9-3 record is still an excellent season but Stanford, Oregon and Utah are all significantly deeper and have better depth and even more highly valued experienced coaching staffs AT THIS TIME.

In the National Picture, Florida State is by experience and talent, right in the middle of the mix. Teams that did not fare as well, which surprised me, included Alabama and LSU. Both fell to 8th and 17th respectively. Alabama returns a starter at QB and LSU does not.

BYU’s Experience Moves the Cougars Up

Oklahoma comes on strong in the Big-12 as does Georgia and Arkansas in the SEC. One team that is way undervalued besides Utah is their arch rival Brigham Young. The Cougars ended the spring with a whopping 65 lettermen and 14 solid returning starters. Not factored into this is the consideration of BYU’s average age for athletes that is nearly one year senior to most other teams because of LDS Missions.

Missions are problematic in that athletes often lose their physical and competitive edge while away from the game. But BYU gains in exchange considerable mental maturing in many cases. It’s probably a wash but the mental side of the game is a big plus.

BYU is simply mature, experienced and undervalued based on that. Yes, it has one of the more difficult road schedules in the country, but the team experience and talent is probably greater than at any time in its storied history. Look for BYU to have a possible BCS busting year.

Other teams that should have big years include Nebraska, Texas and Michigan. Surprise teams like BYU include Tennessee, Texas Tech and Georgia Tech. Greatly overvalued teams besides USC include West Virginia, Michigan State, and South Carolina.

Again, as with any algorithm, it cannot predict the unpredictable. Furthermore these are best record scenarios. It is easier for teams to under perform than over perform. In the end, what matters is what transpires on the field. What this measurement does is to value teams realistically based on their overall Experience factor and presents a snapshot before the season begins that mirrors what SHOULD BE the W/L record of teams by the end of the regular season.

Check out additional articles on Deep Shades of Blue.

Welcome to the Deep Shades of Blue Community!


  • Trey said:

    Fascinating! Thanks for the insight. I like your algorithm. I will be paying more attention to the way things finish this season and compare the results to your rankings based on the algorithm and pollsters preseason rankings.

    I personally think that there should be more value added to the previous year than most pollsters give credit for. Your algorithm accounts for that.

    I would prefer rankings to never be adjusted year to year. Or in other words, maintain the final rankings of the previous season to begin teh current season. That way a team that earned a final ranking only falls out when it fails to perform. I believe It would balance itself out by the end of the year, but no team would be given undue credit with a previous crappy season.

  • Casey said:

    That is an interesting idea from Trey. It seems a little crazy but I think that would be a much better pre-season ranking than the current one which is mostly based on reputation.

    I wonder how well this algorithm has worked on predicting previous seasons, any examples from last year or other years?

  • Jeffrey said:

    Very impressive, and my knee jerk reaction is this would be a great way to rank teams in the preseason. Polls in general really should not be valid until the 4th or 5th game of the season to really position teams properly. It will be interesting to see what the new format will bring. Will it be revamped at all or will the cronies maintain their grip on the voting system that is so biased?

  • Zen said:

    Your algorithm is interesting, but beyond the talking points, I don’t understand the basis. It appears that you have done some sort of multi-linear regression analysis to develop your variable selection and weightings, but then again, it sometimes looks like the weights are arbitrary. Given your stated math background, I expect there must be some rigor behind your numbers.

    Have you run this algorithm against past seasons to determine its predictive strength? If so, what is the R value?

  • Rob H. said:

    I’ve been thinking that this year will be special. Too many people were psychologically damaged from the Utah game to believe in their team big time for this year.

  • Walt Hanssen said:


    I like your article and mathematical matrix. How long have you been doing this and what is your record versus the actual finish of the teams you predicted. Finally, are you able to list your top 25? Maybe if you start tracking your predictions you will be able to get ESPN, USA Today, Fox Sports, etc. to start carrying your Top 25 or power ranking. Thanks

  • Kcatch9 said:

    I, like Zen, would like to see some predictive strength.

    Assumingly, some work has been done to see how predictive this algorithm is (because any good mathematician/statistician would always test his theory on historical data). Thus, I find it interesting that it has been omitted completely from this article.

    Until I see some comparative analysis done on historical data, this algorithm means as much to me as the dart board ranking methods that we currently have.

  • Louis E. Deaux said:

    This is the first year I’ve used this method for looking at logical rankings. I” attempt to answer the many questions.

    1) It was the high ranking for USC, West Virginia, Boise and Michigan State that originally made me want to look at overall experience in multiple dimensions, starting with how many good bodies a team could put on the field. Finding out that Utah, Florida State, Michigan, BYU and Uah were major climbers or at least verified Top 25 teams was actually a surprise.

    2) I had difficulty deciding how to weagh each factor in developing the formulas. In the end, I relied on my observational experience, gut instinct and sure knowledge that “as important as talent may be, coaching competence is clearly more important. I look at Coach Leach at Washington State and and fairly certain that if he had been able to recruit the defenders Mack Brown had gotten over in Austin during his salad days at Texas Tech, his Red Raiders would have been in a national Title game at some point. The man can practically take take playground girls and they’ll score points, he’s that good. STOPPING OPPONENTS WAS ALWAYS tECH’S ISSUE. Ultimately, my formula is a best guess, a theory. In future seasons we’ll be able to see if the theory plays out.

    3) In the past I have done extensive conference analysis, evaluating some of the myths of power and trappings of “traditional” though. At the end of the article, one notices that the SEC gets an enormous amount of media credit and deservedly so in some respects. But the SEC teams mostly play only 4 road games, do not travel, play in weather, at altitude, etc. It is difficult to know how good Alabama is if they only play in the comfort of their geography. How do we know Alabama would beat Michigan in Ann Arbor in early December or late November? We don’t! We only know Ohio State has done that. Conditions matter and the SEC gets a constant media pass on conditioning play outside the south. Mighty Texas struggled for an entire half against Wyoming (elevation 7,200 ft). More important is the fact that Texas was willing to play at laramie in the first place.

    4) The records predicted are best case scenarios. Some teams will perform better, some worse than what is predicted.

    5) In the end I cannot emphasise enough that I was not surprised that Florida State came in #1. I was totally surprised that BYU came in #2 and Utah #4. Of note, many teams on those teams schedule are also predicted to perform better than polls would suggest. Something has to give. Utah and BYU face each other in Salt Lake…edge to Utah. But a BYU win there and that experience of winning, the mental confidence that comes from a victory can be converted by a good coaching staff into follow-up victories…even in Boise the following Thursday Night.

    6) Of corrective note, the loss of a quarterback like Jordan Winn should have read a change = +/-10 points, not 5.

    7) Enjoy the season and don’t worrty too much about any of what we all predict. Football is about the event. have fun and love the event win/lose. Your egos are NOT based on whether your team wins or loses…just who pays for lunch the next day!

  • Seabass said:

    Ehhhh, I was told there would be no math on this site :)

    Interesteing stuff nonetheless. Dabbling in this type of thing is like being a meteorologist, there are so many variable factors when you’re wrong it isn’t a big deal but if your right you’re instantly Nostradamus…

  • BlueHusky said:

    I like the concept but the methodology, as suggested before, it lacks empirical validation. It would be easy to set up a database containing those factors, dating back, say 20 years, and then use a regression-type approach to “train” the model. Validation on a hold-out random sample would be good, then of course, apply the model year by year to predict the results, and feed back into the model. This way the model would “learn”.

    If anybody would like to pull the data together, I have the machine learning tools to test the idea. Reply to this and let me know if you want work with me on this. Perhaps the author?


  • Martin said:

    LOL at all these comments. And Bronco says BYU fans are not intelligent. ;-)

  • FL Cosmo said:

    Loved the post. Very interesting. Couple thoughts, though.
    First, one team’s underperformance directly results in another’s overachievement, so we can’t really say from a holistic view that one is easier than the other. In otherwords, underperformance by teams above us in the polls translates into our overachievement if we win games. Second, it doesn’t make a ton of sense to say that an extra home game has a similar effect on predictability as having two additional returning starters. Extra home games are usually a gimmie and should be counted as a solid win with no predictability impact on the rest of the season in my opinion.

  • FL Cosmo said:

    Check that, for one entity in a vacuum it is definitely easier to underachieve than overachieve. I should specify that I meant in the context of predictability, especially relating to polled rankings, they are the same. At the end of the day, BYU’s got to win four very tough road games, and if they can do it, I think the rest will fall into place and BYU will likely have a solid case for much more than a BCS berth.

  • Jared (the original) said:

    I definitely want to see a reverse or regression analysis.

    For the BB world, Pomeroy has developed the number one predictor of wins of any computer model. This sounds like it has potential for football. But there may be other factors that need to be considered. A regression analysis could go a long way in verifying this.

    Anybody taking you up on the analysis BlueHusky?

    And are we really supposed to be #2 in the nation. Based on this that is amazing. But UTAH at 4.

    I guess it depends on who has the more durable QB.

    If as he says we are that highly prepared for a strong season, there is no one on our schedule including Utah that should beat us. I’m hoping!

  • SoCal Cougar said:

    Historical testing would be great. I think some variables are being weighted incorrectly. For instance the ranking of USC. Yes, they lack depth because of the scholarship restrictions, but their front line talent is tremendous. I don’t think the best case scenario is 3 losses. I would not be at all surprised at an undeafted season for them.

  • Louis Deaux said:

    I agree w/Florida Cosmo. Although UTAH residents tend to poo poo BYU’s reputation, nationally BYU has major 4 decade kwan. BYU is a winner and in spite of the three years of bad football earlier in this century, BYU has been a solid performer, finishing in the top 25 in 18 of 36 prior seasons. The rest of the nation respects BYU. Therefore, a 12-0 season and BYU will be in the middle of talk about a National Championship in addition to a BCS birth. BYU’s schedule is tough and not dissimilar from other major conference schedules. There are six major programs on tap, three from the PAC-12, and projected championship contenders for the MWC and ACC plus Notre Dame. That is every bit as good as any ACC, Big-TEN or PAC-12 schedule.

  • Louis Deaux said:

    SoCal Cougar…I get your love of the Trojans, and clearly my love of the Briuins is not affecting my judgment (LOL), but I did in fact weigh USC’s talent as significantly better than most teams. The problem for USC, Penn State, West Virginia and several other major players in college football, running on the field with just 30-40 letterman means you are fielding an extremely thin squad. You are not even 2 letters deep at every position and while you might have a lot of talent, you need cohesion in a program. Freshman just don’t create cohesion at the get go. I almost always agree with you…but not this time.

  • Casey Adams said:

    What do your numbers say about offensive coordinators with experience? If there is a limiting factor at BYU this year it is the lack of experience of Brandon Doman.

  • top jimmy said:

    I belive Byu will have a good season but unfortinetly BYU would have be be perfect to be as a mormon would call it celestial or BCS worthy and unfortinetly I would Bet my eternal Salvation and Im not Mormon that they will loose at least one game and I am a die hard BYU football fan.

Leave your response!

Add your comment below, or trackback from your own site. You can also subscribe to these comments via RSS.

Be nice. Keep it clean. Stay on topic. No spam.

You can use these tags:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

This is a Gravatar-enabled weblog. To get your own globally-recognized-avatar, please register at Gravatar.

1 × = two