Thursday, July 16, 2020

First Half Point Differential Part IV: Using 1HPD to Handicap Conference Title Games and the College Football Playoff

Is the title to this post too long? Yes, but I couldn't think of a better way to title it other than spell it out explicitly. For a primer on 1HPD, check the other posts in this series. It ain't that complicated.

I have 1HPD conference data going back six seasons to the beginning of the College Football Playoff era, so we might as well see if it provides any value in picking the winner (both straight up and Against the Spread) in certain postseason games. I didn't look at all bowl games since there can be varying levels of motivation in such contests. By contrast, conference title games and the College Football Playoff are practically guaranteed to have motivated participants.

Let's start by looking at the Group of Five. If you knew nothing else about the teams participating in conference title games and merely used 1HPD unadjusted for strength of schedule, you would have done pretty well in picking the outright winners and decent when going against the spread.
The MAC would have been responsible for nearly all of your losses at the window, with the lesser team winning half the time outright and covering five of six times. Otherwise, 1HPD is 16-2 straight up and 12-5-1 ATS in picking winners. Note there have been two Sun Belt Championship Games, but in the second game, both Appalachian State and Louisiana-Lafayette entered with an identical 1HPD, so it is thrown out as ties have not been a part of college football for a quarter century.

When we move to the Power Five, 1HPD does an even better job of picking winners against the spread.
In the Power Five, there are no MAC outliers dragging down the numbers.

Overall, in 51 conference title games 1HPD has correctly predicted the outright winner 39 times (76.5%) and ATS winner 31 times (62%) with one push. Not too shabby. But what of the College Football Playoff? Alas, the predictive accuracy is not as strong.
1HPD has been as reliable as a coin flip both straight up and ATS in College Football Playoff games. Note one game is not included as I use 1HPD in conference play and Notre Dame does not belong to a conference (yet). One potential reason for 1HPD's underwhelming performance in the College Football Playoff versus conference title games is the fact it is not adjusted for opponent. The College Football Playoff features teams from different conferences (usually) instead of teams from the same conference. Schedule strength between two teams from different conferences has the potential to be much greater than schedule strength between two teams from the same conference (duh). That being said, the way 1HPD has performed in handicapping conference title games gives me confidence it is a solid measure of team strength.

We'll be taking a little break here at Statistically Speaking, but we'll return in three weeks (the first Thursday in August). Instead of any previews or gambling nuggets related to the 2020 season, I am operating on the assumption the 2020 season will not be played. Coronavirus numbers are simply too high to justify endangering an unpaid labor force and the testing required to put on a season (even an abbreviated one) would sap testing resources from communities that need them. Let me be clear, I am not a journalist and do not have sources other than what any person with a television set, internet connection, or podcast subscription has. With the 2020 season unlikely to be played, I will focus my efforts on Yards per Play and Adjusted Pythagorean Record recaps of seasons gone by. I have been making YPP and APR posts for the past five seasons, but I have ten years worth of data that has not been given any type of write up (2005-2014). My first YPP flashback will be on a conference season near and dear to my heart, the 2006 ACC. My alma mater ended up winning their first conference title in over three decades and likely the only title in my lifetime. I have some other conferences and years lined up, but I wanted to get an idea of what my readers want. I know there are at least a few folks who find their way here on a regular basis, so let me know in the comments what conference and year combination you would like to see recapped in either YPP or APR. I feel like it is going to be a long time before we can enjoy a beautiful fall Saturday, so in the meantime, let's relive some from the recent past.

Thursday, July 09, 2020

First Half Point Differential Part III: The Power Five in 2019

Last week we looked at 1HPD in the Group of Five. This week we examine 1HPD in the power conferences.

Beginning alphabetically as we always do, here are the ACC standings from 2019.
And now the IHPD with conference rank in parentheses.
Its a surprise to no one that Clemson was head and shoulders above the rest of the ACC. The Tigers 1HPD of +182 is more than the combined 1HPD of the other seven teams that finished with a positive differential (+144). However, the Tigers did not lead at halftime of every conference game. They were tied with North Carolina in a game they eventually won by a single point. Speaking of the Tar Heels, while they enter 2020 as the favorites in the Coastal Division, they were actually in fifth in the division in 1HPD and Duke was not very far behind them. Chaos always seems to reign in that half of the ACC, so don't assume Mack has a cakewalk to the ACC Championship Game.

Moving on to the Big 10 standings.
And the Big 10 1HPD.
Ohio State was the only Power Five team to lead at the half in all their conference games (and they played nine lest we forget). The Buckeyes also managed to lead by double-digits at the half in each of their Big 10 games! Wisconsin put up a fight in the conference title game, but that is still quite a feat. On the other end of the spectrum, in a shocking development, Maryland actually finished with the worst 1HPD! While the Terrapins did beat Rutgers (and led by twenty at the half), they trailed by double-digits six times in conference play (Rutgers trailed by double-digits seven times).

Next up is the Big 12.
And the Big 12 1HPD.
Despite Oklahoma and Baylor finishing with two conference losses between them, the Sooners and Bears were hardly dominant. The Big 12 was the lone Power Five conference to not have a team finish in triple digits in 1HPD. If the 2020 season is played, we should have a helluva race in flyover country. Just 38 points separated the third place team (Oklahoma State) from the eighth place team (Texas Tech) in 1HPD. Add to that the fact the ninth place team (West Virginia) is likely to improve in Neal Brown's second season and the perennial punching bag at least showed flashes of competency last season (and is coached by a national champion). I wouldn't go so far as to predict Oklahoma failing to qualify for the Big 12 Championship Game for the first time since its reinstatement, but a title game pitting Iowa State and TCU wouldn't be the most shocking development.

Here are the Pac-12 standings.
And the Pac-12 1HPD.
While the Big 12 had no dominant teams, the Pac-12 had no true dregs. Arizona brought up the rear, but the Wildcats 1HPD of about negative eight points per game was the best of any last place finisher in the Power Five (NC State was second at negative twelve). Oregon had the lowest 1HPD of any Power Five champion. In fact, the Ducks did not lead at the half in four of their nine conference games. By comparison, the other four Power Five champs did not hold a lead five times (out of 34 total conference games).

Finally, here are the SEC standings.
And the SEC 1HPD.
I know its fun to hate on Alabama, but don't go throwing dirt on the Tide just yet. Alabama was nipping at the heels of LSU in 1HPD and if we look at the other seven conference games both teams played, Alabama actually had a better differential (+137 to +113). Of course, that classic in Tuscaloosa does count and lets also give LSU credit for never trailing at the half in any conference game (they were tied against Florida and Auburn). In the East, Georgia finished with a healthy 1HPD margin over Florida and Kentucky and despite the relatively disappointing season, the Bulldogs led by double-digits six times in SEC play (tied with Alabama and LSU for the most double-digit leads).

In the YPP and APR offseason recaps, I sort the teams in each conference by how much they over or under-performed relative to their expected record. Since all the Power Five conferences are grouped together here, I am only going to list those that significantly over or under-performed (a difference of at least .200). We'll start with the overachievers.
Louisville beat one FBS opponent in 2018, but the Cardinals improved to seven regular season wins in their maiden voyage under Scott Satterfield. However, there are a few flashing warning signs if you expect continued improvement in 2020. Despite an 8-5 overall record, the Cardinals were actually outscored in 2019 and in ACC play, they trailed at the half five times. Despite a 5-3 conference record, their 1HPD ranked fifth in the Atlantic Division. NC State and Syracuse probably won't be as bad as they were last season and Florida State is likely to rebound as well. The Cardinals also retain two challenging non-conference games in 2020, as they travel to Notre Dame and host their in-state rival Kentucky. A return to bowl eligibility, even it it means a regression to 6-6 should still be seen as a success as the Cardinals try to move on from Bobby Petrino. A championship repeat by Oregon in the Pac-12 would be quite an accomplishment considering the Ducks lost their starting quarterback in the NFL draft and did not post numbers typically befitting an 8-1 conference record. Tennessee has been one of the hardest teams for me to get a read on this offseason. Were the Volunteers good in 2019? They closed the year on a six-game winning streak, beating three bowl teams in the process (Indiana, Kentucky, and UAB). Their YPP numbers were solid (fifth overall and third in the East). On the other hand, their APR numbers were much less glowing (ninth overall and fourth in the East). They also lost to Georgia State and were pounded by the three best teams on their schedule (Alabama, Florida, and Georgia beat them by a combined 82 points). And since when are Tennessee fans delighted by victories against Indiana, Kentucky, and UAB? Their 1HPD also shows them to be pretty weak. Despite winning five of their eight league games, they were outscored in the first half in conference play and actually trailed at the half five times, including four times by double-digits. The Vols do have the benefit of drawing Arkansas in their rotating cross-division game in 2020, but Alabama, Florida, and Georgia are still on the schedule as well as a non-conference trip to Oklahoma. 8-4 seems like the ceiling for this team with a real possibility of 6-6 or worse.

And now the underachievers.
Earlier I mentioned the Big 12 could have one of the more interesting conference races in the Power Five. Part of the reason for that is because teams like the Red Raiders and Horned Frogs are likely in store for some positive regression in 2020. Texas Tech and TCU combined to finish 5-13 in conference play last season despite outscoring their league foes by seven points in the first half. Those aren't great margins, but are more indicative of a .500 level conference finish. So in effect we have two middling Big 12 teams that were led by a first year head coach and a true freshman quarterback respectively. Both those conditions are often marked by initial inconsistency and then marked improvement as time goes on. Keep an eye on both in the Big 12 race. The only other Power Five team to significantly under-perform relative to their 1HPD was Washington State. In Mike Leach's final season on the Palouse, the Cougars lost twice as many conference games as they won despite outscoring their nine Pac-12 opponents in the first half. The Cougars fielded a very bad defense in 2019 that kept them from closing games. That defense has almost no choice but to improve in 2020. The turnover at the top is always a concern, especially at a locale that does not recruit well, but Washington State can expect to extend their bowl streak to six years in 2020.

Thanks for reading. Check back next Thursday when we see how well 1HPD has done in handicapping conference title games and the College Football Playoff.

Thursday, July 02, 2020

First Half Point Differential Part II: The Group of Five in 2019

Last week I introduced the nuanced and complex statistic of First Half Point Differential (1HPD) with a promise to examine the 2019 college football season through that lens. I keep my promises roughly half the time, so enjoy this review of Group of Five conference play from 2019.

American Athletic Conference
Like we always do, we'll start with a look back at the 2019 AAC standings.
And now we'll look at how the conference stacked up in terms if 1HPD in league play (conference rank in parentheses).
The first thing that jumps out is the two best teams by 1HPD did not win their respective divisions. UCF outscored their conference opponents by nearly 100 more points in the first half than the eventual East division winner (Cincinnati), but lost a tight game to the Bearcats and also dropped another close game to Tulsa. The Knights led both games at halftime by the way (by six and eleven points respectively) before frittering away the leads in the second half. However, despite losing a conference game for the first time in nearly three years, the Knights were still a dominant force in the AAC. In the West division, Memphis and Navy finished tied atop the division, but the Tigers won the head to head matchup (despite trailing at the half) and thus the tiebreaker. Elsewhere in the conference, the departing Huskies finished dead last in 1HPD, but they did lead at the half against Temple in the season finale. That marked the first time they led at the half against an AAC foe since 2017. That also happened to be their last conference win. Of course, they went on to lose to the Owls by 32 points.

Conference USA
The CUSA standings.
And the 1HPD.
Lane Kiffin's parting gift for FAU was a second conference title in three seasons, but the Owls were not nearly as dominant in 1HPD as they were two years ago. In 2017, they outscored their eight league opponents by nearly twice as many points (+121) in the first half of games. Out West, UAB won their second consecutive division title, but the Blazers actually finished just behind Southern Miss in 1HPD. The Eagles had the division in their talons after beating UAB by the odd score of 37-2, but dropped their final two conference games to Western Kentucky and FAU (a pair of East division foes) to give the division to UAB.

The MAC standings.
Did the wrong team win? Despite a 1HPD that was nearly double that of the second best team in the conference, Buffalo did not even win the East division. The Bulls opened conference play with back-to-back losses to Miami and Ohio, but were dominant down the stretch, save for a wild comeback by Kent State. The Bulls actually led at halftime in all eight of their conference games, something no other Group of Five team can boast.  At the other end of the spectrum, Akron did not win a game (in conference play or otherwise) and their 1HPD was indicative of that performance.

Mountain West
Here are the Mountain West standings.
And the Mountain West 1HPD.
Boise won the conference and had the best 1HPD. The interesting thing about the Mountain West is that every team in the Mountain Division, with the notable exception of New Mexico, finished with a positive 1HPD. In the West Division, Hawaii was the only team to finish with a significantly positive 1HPD (Fresno State was slightly above water at +6). With such a scoring imbalance, it probably won't surprise you that the five teams from the Mountain Division not named New Mexico went 13-2 against the West, with both losses coming to San Diego State.

Sun Belt
Finally, the Sun Belt standings.
And the Sun Belt 1HPD.
The two best teams in the conference won their respective divisions and met in the league title game. I was a little surprised at Appalachian State's 1HPD. I figured they would have been more dominant considering they are arguably the best Sun Belt team of all time and the first to finish the season ranked.

In the YPP and APR offseason recaps, I sort the teams in each conference by how much they over or under-performed relative to their expected record. Since all the Group of Five conferences are grouped together here, I am only going to list those that significantly over or under-performed (a difference of at least .200). We'll start with the overachievers.
Cincinnati had a great shot at winning the AAC, holding a lead against Memphis in the AAC Championship Game with under five minutes to play, until the Tigers prevailed. The Bearcats have won 22 games over the past two seasons, but can probably expect some regression in 2020. In Conference USA, a pair of first year coaches exceeded expectations. Will Healy guided the 49ers to the first bowl game in school history and Tyson Helton helped the Hilltoppers get back to the postseason. However, I wouldn't fancy either of these teams as contenders in the East this season. The 49ers were outscored in the first half of their conference games and Healy has a history of regressing after breakthrough seasons. He followed up an eight-win campaign at Austin Peay with a losing record. Meanwhile, despite Helton's background as an offensive coach, the Hilltoppers were led by their defense in 2019. A losing season by either or both in 2020 would not be that shocking. Air Force rebounded from back-to-back losing seasons and featured perhaps the best team of Troy Calhoun's tenure. However, despite cresting, the Falcons were unable to wrest the division away from Boise State and I wouldn't expect them to be able to in 2020 either. Finally, elsewhere in the Mountain West, Nevada managed to qualify for a second consecutive bowl game despite very bad first half numbers. The Wolfpack trailed by double-digits in half their conference games. Despite upheaval within the division (four of the six teams are breaking in new head coaches), I expect Nevada to finish with a losing record in 2020.

And now the underachievers.
Houston underachieved in the AAC, spoiling Dana Holgorsen's debut. It should be noted they still put up solid first half numbers despite the absence of quarterback D'Eriq King for seven of their eight conference games. King won't be back in 2020, but his backup earned valuable experience. Despite winning double-digit games for the third consecutive season, UCF actually underachieved in 2019. The Knights are the favorite heading into 2020, not just in the AAC, but also in terms of finishing as the highest ranked Group of Five team. In Conference USA, Rice actually outscored their opponents in the first half despite posting a losing league record. A tough non-conference schedule with games against Army, Houston, and LSU could keep the Owls out of the postseason, but if you are looking for a sleeper to win a wide open Conference USA, the Owls are a fine pick. Their in-state brethren have won just two games over the past three seasons, but I expect the Miners to dig deep and get to at least three wins in 2020. As I mentioned previously, Buffalo was arguably the best team in the MAC last season which is something considering the talent they lost to the NFL and the transfer portal. If the 2020 season is played, the Bulls will be the best MAC team since Western Michigan in 2016. Colorado State and Fresno State both finished with losing conference records despite outscoring their opponents in the first half. And both now have new head coaches stepping into good situations. Finally, South Alabama has a chance to make a third year leap under Steve Campbell after posting decent numbers hardly befitting a 1-7 outfit.

As always, thanks for reading. We'll be back next Thursday with a look at 1HPD in the Power Five. Enjoy the 4th and wear a damn mask.