Last week, I made a post looking back at last year's football season, specifically at my picks last summer and how those projections played out. What I found surprised me a bit. As it turns out, my probability-based predictions, specifically what I refer to as the "win distribution matrix" for each team were mathematically pretty accurate. Essentially, if I predicted a certain team had only a 10% chance of winning "x" number of games, that wound up being true for about one out of 10 teams. In other words, the math checked out. Furthermore, in my conference-by-conference analysis, I did a fairly good job of predicting what would happen, and when my predictions failed, I could usually trace it back to a single game here or there or a team or two that was either way worse or way better than expected. Finally, I only got 6 of the 12 NY6 participants correct, but again the reasons why were clear. And so, with this all in mind, it is time to once again look forward and to apply my methodology to the upcoming 2018 season.
(Here I will insert a repeated paragraph on my methods from last week's post.) I can (and have in the past) spend several paragraphs explaining my methodology. If you are curious about the details, more information is found here. But, the basic idea is that over the years I have developed power rankings and algorithms that allow me to predict point spreads and victory probabilities for each game of the entire season. I have also built in some historical data that takes into account the uncertainly that comes with the preseason rankings (which, on average, are off by about 20 slots, per team, as I showed in this previous post). With all this information, I can predict the probability of each team winning "x" number of games, and I can simulate the final standings in each conference based on a reasonable number of upsets. While my methods are not perfect (I rely heavily on the idea home underdogs are the biggest source of upsets) I do believe that they provide insight in the coming season, especially as it relates to the impact of schedule. If nothing else, my predictions usually provide something a little different than just saying the top 4 teams in the preseason rankings will make the playoffs.
So, it keeping with last year's format, I will attempt to answer several burning questions about the 2018 season. Let's begin.
Basically, my math suggests that MSU is most likely to win 9 games and has a 32% chance to win at least 10 regular season games, and a 95% chance to at least make a bowl game. If you think a 5% chance to have a losing season is high, I would only caution us all to remember 2016, where in this space I predicted MSU to make the play-offs. On the super-optimistic side, I would add that MSU also has a 2.5% chance to run the table at 12-0, which is 9th highest among Power 5 schools and better odds than Oklahoma and Penn State (as examples). By comparison, #1 ranked Clemson's odds are 6.8% and #2 Alabama's (2) are 6.5%.
As for actual wins and losses, my standings simulation likes MSU's schedule. As we all know, MSU is scheduled to play three other Big Ten East team that are ranked in the Top 10: Ohio State (consensus rank of 3), Penn State (9), and Michigan (10). But, MSU gets two of those three teams at home (OSU and UofM) and only PSU on the road. Therefore, the most likely scenario is for MSU to win the 2 home games and lose the road one. Other than that, the schedule appears very manageable with the next toughest games projecting as the home game with Northwestern (34) and the road game at Nebraska (52), both of which MSU is expected to win. So, if we add that all up, I project MSU to finish the regular season at 11-1, with the sole loss in (un)Happy Valley. But, will that be enough to make it to Indy? Let's turn to that question next.
As expected, the two highest ranked Big Ten schools, Ohio State (3) and Wisconsin (6) have higher expected values for their win totals and most "experts" project them to win their respective divisions. But, the standings simulation paints a much different picture. In the East, it is clear that it is a 4-team race between OSU (3), PSU (9), UofM (10), and MSU (13). When each team's schedule is taken into consideration, the team with the leg up is actually Penn State.
The logic is simple. All four teams are good, and all four would have good odds to beat almost anybody in their home stadiums. MSU and PSU have the advantage of having 2 home games and only 1 road game in that 4-team round robin. This means that the most decisive game in the Big Ten East this year is very possibly MSU's road game to Penn State on October 13th. In this scenario, the winner of that contest would go to Indy. In this universe, Penn State would also go 11-1, with their sole loss coming in Ann Arbor. It won't be easy, however, as the Lions do have tough home games vs. OSU (3), MSU (13), Wisconsin (6), and Iowa (31). I would not discount the road game at Pitt (48) either, but the remainder of their road schedule looks very manageable. If PSU were to trip up in one other conference game, MSU might just find themselves in Indy anyway.
As for the Buckeyes, despite their talent and high ranking, this schedule is a bit problematic for them. As stated above, since both PSU and MSU have an edge on the schedule and since OSU must travel to both places, it creates a bit of a double disadvantage. If OSU were to beat only one of those two teams, they are still likely to lose the tiebreaker with the other team. For example, if OSU were to win at PSU, but lose at MSU (keeping all other games the same) OSU would tie for the division lead, but lose the tie-breaker to MSU. Similarly, if OSU were to win in East Lansing, but lose in Happy Valley (keeping all other games the same) OSU would lose the tie-breaker to PSU. Basically, there is a very good chance that OSU needs to beat both MSU and PSU on the road to make it to Indy. My simulation suggests the Buckeyes will lose both and finish 10-2. But, I do project them to at beat TCU in one of marquee September non-conference match-ups, so at least it is not a total loss.
And then there is Michigan (10). Despite the annual round of irrational exuberance coming out of Ann Arbor, their schedule is, by my analysis, the most difficult in the nation. As the win distribution matrix above shows, their expected value of overall and conference wins is noticeably lower than the other Big Ten contenders. Not only do the Wolverines have to travel to Ohio State and MSU (2 likely losses), they have Wisconsin on the docket as well. And, it is actually a little worse than that, as one of their other road games is at Northwestern (34), who is just good enough to cause problems. Actually, my standings simulation has the Wolverines dropping the game in Evanston to finish Big Ten play at 6-3. Oh, and then there is the opener at Notre Dame (11) which is also projected as an "L". Add it up, and Harbaugh might just be starring at 8-4 once again, and that is assuming the Wolverines beat both Penn State and Wisconsin at home. A little bad luck, and 6-6 is a real possibility. My simulation even predicts a 10.1% chance that the Wolverines are home for the holidays.
In the West, conventional wisdom suggests that Wisconsin (6) will win the Division (and, let's be honest, they probably still will). Their projected total and conference wins are both quite high. But, my standings simulation says, "not so fast my friends." The problem is Wisconsin's schedule, especially on the road. For the East cross overs, the Badgers must travel to both Penn State (9) and Michigan (10). Those games both project as losses. So, in order to feel comfortable still winning the West, Wisconsin essentially needs to run the table in their Division. Unfortunately for them, they also must travel to face three of the top four other teams in the West: Iowa (31), Northwestern (34), and Purdue (55). None of those games are a gimme.
In contrast, the Hawkeye's schedule is MUCH more manageable. While Iowa does have to travel to Penn State (9), their other two East cross-overs are at Indiana (61) and vs. Maryland (68). Furthermore, Iowa draws Wisconsin (6), Northwestern (34), and Nebraska (52) all in Iowa City. Outside of PSU, their second toughest road game is at Purdue (55). Based on all these factors, my standings simulation actual projects Iowa to win the West by 2-games and finish at 7-1 and 11-1 (with a loss to Penn State, and with a non-conference win over Iowa State (37)). The decisive game out West is the 9/22 match-up between Iowa and Wisconsin in Iowa City. If Iowa does win that game, Bucky Badger is in B1G trouble. Even if the Badgers win, they can't afford to drop a game in Evanston (for example) unless they can win in Ann Arbor or Happy Valley. Northwestern is a threat on paper, but I project them to finish 5-4 with losses at Purdue, Iowa, MSU, and vs. Wisconsin, but with an upset win over Notre Dame in Evanston, which would certainly help the conferences image during bowl selection.
In this universe, Penn State (9) would square up against Iowa (31) in Indy, where the Lions would be expected to win this regular season rematch again and run their record to 12-1. Now, if the Badger can survive their schedule and make it to Indy, Wisconsin would be favored in this hypothetical universe.
Bama's schedule this year seems extremely manageable, which explains the full one-game gap in expected conference wins between them and everyone else in the West. Their SEC East cross overs are Missouri (33) and at Tennessee (66) and they face the 3 next-best SEC West teams all at home: Auburn (8), Mississippi State (15), and Texas A&M (22). The Tide do travel to LSU (25) and my simulation actually does predict a loss here, but even so I have Alabama at 7-1 and 11-1 overall. The decisive game out West appears to be on 11/10 when Mississippi State comes to Alabama. An upset here would possibly knock Bama out of the SEC Title game, as the Bulldogs also have a favorable schedule (especially relative to Auburn, who must travel to Alabama, Georgia (4), and Mississippi State). But, let's be honest, even if Bama does lose to the "other MSU", they will likely still make the playoffs and still win the National Title.
In the SEC East, the picture is a bit more murky. Based on the win distribution matrix, Georgia (4) seems to have a similar advantage as Alabama, and it would be natural to assume that the Bulldogs will roll again to the SEC Title game. But, their schedule does appear to contain a few potential landmines. The first, and most important of which is the 9/8 SEC opener at South Carolina (26). The Gamecocks are projected to be just good enough to be dangerous and have just a manageable enough schedule that my simulation actually projects them to beat UGA and win the East with a record of 7-1 (10-2 overall, with a loss at Florida and at Clemson on the books). As for Georgia, even if they can beat the Cocks, they have some tricky contests at Missouri (33), at LSU (25), and vs. Florida (29). My simulation sees them tripping up at South Carolina and at LSU to finish at 10-2 overall. The Gators are potential dark horse as well. If they can beat Georgia, they would gain the tie-breaker upper hand over USC if the outcomes of all the others games are the same.
In this universe, Bama would likely destroy South Carolina in the SEC Title game to finish at 12-1, while 10-2 Georgia would be at home, wondering if they can still sneak into the playoffs.
In conference play, the Tigers do travel to Tallahassee this year to face FSU (17) and my simulation does predict a loss here. But, I also have FSU finishing 5-3, so they do not look like a real threat. The only potential road bump is the 11/3 match-up with Louisville (32) visiting Clemson. The Cardinals have a more favorable schedule, and if they were to upset the Tigers, they would be in the driver seat to win the Atlantic Division. But, as the win distribution matrix shows clearly, Clemson is head-and-shoulders above every other team in the Division, and it would be a shock for them not to make the conference title game. Perhaps more important is the non-conference game in Week 2 where Clemson travels to Texas A&M (22). My simulation favors the Aggies here, which would put Clemson at 10-2 overall. Will that be enough to make the play-offs?
In the Coastal Division, the picture is a little less clear, maybe. Based on the win distribution matrix, it is clear that this division is a two-team race between Miami (12) and Virginia Tech (19). While Miami is projected to be a little bit better, than must travel to Blacksburg this year on 11/17, and Tech is likely to win the head-to-head match-up. But, both teams also play Florida State (17), with Miami getting them at home (likely a win) and Tech drawing them on the road (likely a loss). Both teams then have a couple of sneaky tricky conference road game. Miami travels to BC (38) and GA Tech (42), while VA Tech has to travel to Duke (43) and Pitt (48). On paper, those all should be wins, but the margin of error for both squads is razor thin. My simulation has both teams dropping one of those games (BC and Duke), so in the end it is the head-to-head result that will most likely prevail, sending VA Tech to the conference title game, where Clemson will have there way with them.
As a final note, once again the ACC has a large number (7) of non-conference match-ups with another Power 5 conference, the SEC. This is almost twice the number of inter-sectional match-ups between any other pair of Power 5 conferences. While the SEC has historically dominated this series, in 2018 the ACC looks like they will be likely favored in 4 of the 7 match-ups (South Carolina at Clemson, Florida at Florida State, Kentucky at Louisville, and LSU vs. Miami), while the SEC will only be clearly favored in 2 (Alabama vs. Louisville, and GA Tech at Georgia). The Clemson at Texas A&M game is a bit of a toss up, but Clemson is still likely to at least be favored. The overall result of this "ACC-SEC Challenge" is likely to be at least a footnote in the Playoff / NY6 selection process, and the ACC is poised to have the upper hand.
I am going to go with "blip." After the disastrous 2016 season in which no one was even sure if the Big 12 deserved to be called a Power 5 conference, Oklahoma had a great year last year, making it all the way to the Playoffs. This year, the Sooners (7) start the season in the Top 10, and as the win distribution matrix shows, they have the highest expected value of wins of any Big 12 team again this season. But, trouble may come in as there is no shortage of other dangerous teams such as TCU (18), Texas (20), Oklahoma State (23), and West Virginia (28). Unfortunately for OK, of the teams above, they only draw OK State at home this year, and thus my simulation has them finishing at only 6-3 in conference play. In contrast, both Texas and TCU have slightly lighter schedules, and I project them to tie for the Big 12 Title at 7-2. While Texas hosts TCU this year in the regular season (and thus I project the Longhorns to prevail), the new (and completely unnecessary) Big 12 Championship game would dictate that Texas and TCU would have to play again and in this case, TCU would be slightly favored to win the conference. The bright side for the Sooners is that if they can beat either TCU or Texas, they would slide up to 2nd place and due to the new Championship game format, they would have a shot to still win the Title. So, I am going to designate the both the TCU - Oklahoma game and the Red River Shootout (Texas vs. Oklahoma) as the 2 pivotal games in the Big 12 this year. Honestly, Oklahoma getting at least one of those two "W"s seems quite likely. But, even so, their record would be only 10-2 and seeing that Ohio State has been replaced on the schedule with a very mediocre UCLA (47) squad as their marquee non-conference game, I doubt this would be enough to make the Playoffs.
Yes, I think that there will be. Based on the preseason rankings, Washington (5) is considered the team to beat, and the win distribution matrix above tends to support this. However, there is also a batch of teams right behind them including Stanford (14), USC (16), Oregon (24), Utah (27), and Arizona (35) who all have similar win distribution profiles. Because of the tight grouping of the teams, the schedule will once again likely be a factor. In this case, it is the Oregon Ducks that have the edge in the North, seeing as they face both Washington and Stanford at home this year. Washington does draw Stanford at home, which means the Washington at Oregon game on 10/13 is most likely decisive game in the North. My standings simulation has all three teams finishing at 7-2 in conference play, which give Oregon the spot in the Pac 12 title game based on head-to-head. Unfortunately for Washington, my simulation also projects them to lose to Auburn in Week 1, so even a NY6 Bowl might be out of reach if the Huskies do finish at 9-3. Similarly, Stanford's road trip to South Bend this year is likely also to give them a third "L" as well.
In the South, the competition is a bit weaker, but the analysis is very similar. The three main contenders for the south crown (USC, Utah, and Arizona) all play each other, and in this case Utah is the team with the schedule edge. Utah draws both teams at home, while USC is the team that must play both other teams on the road. Interestingly, Arizona does have an easier cross-over schedule (as they draw Oregon State, Washington State, and Oregon in the North while both Utah and USC draw Stanford on the road). At the end of the day, my simulation has Utah finishing at 7-2, a game ahead of both Zona and the Trojans. But, both 10/13 Arizona at Utah and 10/20 USC at Utah contests would swing control of the conference to the winner, with the Arizona at Utah game being perhaps a bit more significant. If this all comes to pass, Oregon would be favored to beat Utah in the Pac 12 Title game (in a regular season rematch) to finish at 11-2. But, there putrid non-conference schedule (no offense Bowling Green and San Jose State) and losses at Utah (in the regular season) and at Arizona would likely keep them out of the Playoff hunt.
Perhaps. The Irish are on of the more interesting teams this year, as their schedule is an odd combination of manageable and tricky. Notre Dame has on its schedule six games against teams ranked in the top 35, and my projections suggest the spreads in all 6 games will be tighter than one TD. Because of all these potentially tight games, I only have their expected wins at 7.73 and their odds of winning 10 games or better is just below 20%. Similarly, my standings simulation projects losses at VA Tech (19), at USC (16), and a close lose at NW (34) to drop the Doomers to 9-3 overall. All this assumes wins vs. Michigan (10), at Wake Forest (44), vs. Stanford (14). If I stick with my simulation, ND is clearly out. But, realistically, NONE of those six teams really scare me, especially since UofM and Stanford need to come to South Bend. USC is good on paper, but that is often the only place that they are good, and it is hard to believe that the Wildcats will actually upset the Doomers even in their own stadium (which will be filled with ND alums). The VA Tech game will certainly be a challenge, but I could easily see the Irish at 11-1.
In the AAC, both UCF and Memphis look to have fairly smooth sailing to their respective division titles, with UCF expected to lose at Memphis in a cross over, but with each team winning every other conference contest. UCF projects to win the rematch in the AAC Title game and also finish at 12-1. But, UCF in this scenario would score two Power 5 victories as they travel to UNC (58) in Week 3 and host Pitt (48) in Week 5, two games I project them to win. Memphis, incidentally, gets a shot at Missouri (33) on 10/20, but a win is unlikely. If both Boise and UCF win their conference titles and finish 12-1, it is a virtual coin flip as to which team would finish higher in the final playoff poll. Just for the sake of variety, I am going to go with Boise, but I don't feel great about it.
If the Mountain West and AAC both implode somehow, Florida Atlantic could make an interesting case for the NY6. I have them running the table in C-USA, but they actually get a shot at UCF in Week 4, and an upset could leave FAU at 12-1 to end the season, with the sole loss occurring in Week 1 at Oklahoma. The MAC does not look like a threat to the NY6 this year, as I project everyone to have at least 3 losses, with Northern Illinois beating Buffalo in the Title game. But, the Sunbelt could produce a 12-1 team in Arkansas State (71) or Appalachian State (75), depending on the winner of the 10/13 contest at Arkansas State and the new Sun Belt Championship game. But, neither team is at all likely to go undefeated, as they face Penn State and Alabama, respectively.
So, how would this shake out? Two playoff teams would be very clear: Bama and Penn State, and although there would be some debate over the #1 vs. the #2 seed, I find it hard to believe that Bama would not get the 1-seed, even if Penn State's resume, honestly, looks a bit stronger.
As for the 3 and 4-seeds, as much as it makes me sound like a homer, if MSU does finish as one of only 3 teams in the Power 5 with 1 loss, and that 1 loss is to another Playoff team, MSU has to be in, right? For the final spot, Clemson, OSU, and Georgia would all seem to have the strongest cases, but the nod would most likely go to Clemson, based on the fact that they would have a conference title, and the overall strength of their resume. As for the final seeding, I just can't image that the committee would line up Clemson and Alabama in the semi's again with two Big Ten on the other side. So, I am guessing it would look like this:
As for the remaining four NY6 Bowls, the Rose and Sugar Bowl are contractually obligated to certain conferences, so I would project the following match-ups:
I should also note for completeness that the values above differ from the ones in the win distribution tables because above I use the less conservative value that I get before I apply the correction for the odds that the preseason ranking are wrong. If I were to use the more conservative numbers, it just tends to drag down the O/U for the top rated teams and bump up the O/U for the low ranked teams. Those teams basically have no where to go but down or up and so those number tend to regress more to the middle.
Well, that is all for now. Enjoy!
(Here I will insert a repeated paragraph on my methods from last week's post.) I can (and have in the past) spend several paragraphs explaining my methodology. If you are curious about the details, more information is found here. But, the basic idea is that over the years I have developed power rankings and algorithms that allow me to predict point spreads and victory probabilities for each game of the entire season. I have also built in some historical data that takes into account the uncertainly that comes with the preseason rankings (which, on average, are off by about 20 slots, per team, as I showed in this previous post). With all this information, I can predict the probability of each team winning "x" number of games, and I can simulate the final standings in each conference based on a reasonable number of upsets. While my methods are not perfect (I rely heavily on the idea home underdogs are the biggest source of upsets) I do believe that they provide insight in the coming season, especially as it relates to the impact of schedule. If nothing else, my predictions usually provide something a little different than just saying the top 4 teams in the preseason rankings will make the playoffs.
So, it keeping with last year's format, I will attempt to answer several burning questions about the 2018 season. Let's begin.
1) What is the outlook for MSU this year?
Quite positive and perhaps even better than expected. MSU's "consensus ranking" (based on the average preseason ranking from Phil Steele, Athlon, Lindy's, ESPN's FPI, S&P+, and USA Today) is 13. Based on my simulation, MSU's "expected value" of wins is 8.56, with 6.76 conference wins. That might seem a bit low (as this is a bit of a conservative calculation) but it is actually the 8th highest among Power 5 teams and in this case the full win distribution tells a more complete story:Basically, my math suggests that MSU is most likely to win 9 games and has a 32% chance to win at least 10 regular season games, and a 95% chance to at least make a bowl game. If you think a 5% chance to have a losing season is high, I would only caution us all to remember 2016, where in this space I predicted MSU to make the play-offs. On the super-optimistic side, I would add that MSU also has a 2.5% chance to run the table at 12-0, which is 9th highest among Power 5 schools and better odds than Oklahoma and Penn State (as examples). By comparison, #1 ranked Clemson's odds are 6.8% and #2 Alabama's (2) are 6.5%.
As for actual wins and losses, my standings simulation likes MSU's schedule. As we all know, MSU is scheduled to play three other Big Ten East team that are ranked in the Top 10: Ohio State (consensus rank of 3), Penn State (9), and Michigan (10). But, MSU gets two of those three teams at home (OSU and UofM) and only PSU on the road. Therefore, the most likely scenario is for MSU to win the 2 home games and lose the road one. Other than that, the schedule appears very manageable with the next toughest games projecting as the home game with Northwestern (34) and the road game at Nebraska (52), both of which MSU is expected to win. So, if we add that all up, I project MSU to finish the regular season at 11-1, with the sole loss in (un)Happy Valley. But, will that be enough to make it to Indy? Let's turn to that question next.
2) Which two teams from the Big Ten will make it to Indy?
The full win distribution matrix for the Big Ten is shown here:As expected, the two highest ranked Big Ten schools, Ohio State (3) and Wisconsin (6) have higher expected values for their win totals and most "experts" project them to win their respective divisions. But, the standings simulation paints a much different picture. In the East, it is clear that it is a 4-team race between OSU (3), PSU (9), UofM (10), and MSU (13). When each team's schedule is taken into consideration, the team with the leg up is actually Penn State.
The logic is simple. All four teams are good, and all four would have good odds to beat almost anybody in their home stadiums. MSU and PSU have the advantage of having 2 home games and only 1 road game in that 4-team round robin. This means that the most decisive game in the Big Ten East this year is very possibly MSU's road game to Penn State on October 13th. In this scenario, the winner of that contest would go to Indy. In this universe, Penn State would also go 11-1, with their sole loss coming in Ann Arbor. It won't be easy, however, as the Lions do have tough home games vs. OSU (3), MSU (13), Wisconsin (6), and Iowa (31). I would not discount the road game at Pitt (48) either, but the remainder of their road schedule looks very manageable. If PSU were to trip up in one other conference game, MSU might just find themselves in Indy anyway.
As for the Buckeyes, despite their talent and high ranking, this schedule is a bit problematic for them. As stated above, since both PSU and MSU have an edge on the schedule and since OSU must travel to both places, it creates a bit of a double disadvantage. If OSU were to beat only one of those two teams, they are still likely to lose the tiebreaker with the other team. For example, if OSU were to win at PSU, but lose at MSU (keeping all other games the same) OSU would tie for the division lead, but lose the tie-breaker to MSU. Similarly, if OSU were to win in East Lansing, but lose in Happy Valley (keeping all other games the same) OSU would lose the tie-breaker to PSU. Basically, there is a very good chance that OSU needs to beat both MSU and PSU on the road to make it to Indy. My simulation suggests the Buckeyes will lose both and finish 10-2. But, I do project them to at beat TCU in one of marquee September non-conference match-ups, so at least it is not a total loss.
And then there is Michigan (10). Despite the annual round of irrational exuberance coming out of Ann Arbor, their schedule is, by my analysis, the most difficult in the nation. As the win distribution matrix above shows, their expected value of overall and conference wins is noticeably lower than the other Big Ten contenders. Not only do the Wolverines have to travel to Ohio State and MSU (2 likely losses), they have Wisconsin on the docket as well. And, it is actually a little worse than that, as one of their other road games is at Northwestern (34), who is just good enough to cause problems. Actually, my standings simulation has the Wolverines dropping the game in Evanston to finish Big Ten play at 6-3. Oh, and then there is the opener at Notre Dame (11) which is also projected as an "L". Add it up, and Harbaugh might just be starring at 8-4 once again, and that is assuming the Wolverines beat both Penn State and Wisconsin at home. A little bad luck, and 6-6 is a real possibility. My simulation even predicts a 10.1% chance that the Wolverines are home for the holidays.
In the West, conventional wisdom suggests that Wisconsin (6) will win the Division (and, let's be honest, they probably still will). Their projected total and conference wins are both quite high. But, my standings simulation says, "not so fast my friends." The problem is Wisconsin's schedule, especially on the road. For the East cross overs, the Badgers must travel to both Penn State (9) and Michigan (10). Those games both project as losses. So, in order to feel comfortable still winning the West, Wisconsin essentially needs to run the table in their Division. Unfortunately for them, they also must travel to face three of the top four other teams in the West: Iowa (31), Northwestern (34), and Purdue (55). None of those games are a gimme.
In contrast, the Hawkeye's schedule is MUCH more manageable. While Iowa does have to travel to Penn State (9), their other two East cross-overs are at Indiana (61) and vs. Maryland (68). Furthermore, Iowa draws Wisconsin (6), Northwestern (34), and Nebraska (52) all in Iowa City. Outside of PSU, their second toughest road game is at Purdue (55). Based on all these factors, my standings simulation actual projects Iowa to win the West by 2-games and finish at 7-1 and 11-1 (with a loss to Penn State, and with a non-conference win over Iowa State (37)). The decisive game out West is the 9/22 match-up between Iowa and Wisconsin in Iowa City. If Iowa does win that game, Bucky Badger is in B1G trouble. Even if the Badgers win, they can't afford to drop a game in Evanston (for example) unless they can win in Ann Arbor or Happy Valley. Northwestern is a threat on paper, but I project them to finish 5-4 with losses at Purdue, Iowa, MSU, and vs. Wisconsin, but with an upset win over Notre Dame in Evanston, which would certainly help the conferences image during bowl selection.
In this universe, Penn State (9) would square up against Iowa (31) in Indy, where the Lions would be expected to win this regular season rematch again and run their record to 12-1. Now, if the Badger can survive their schedule and make it to Indy, Wisconsin would be favored in this hypothetical universe.
3) Will Bama run roughshod over the SEC?
Probably, yes. At this point, it is probably best to pencil the Crimson Tide into the National Title game pretty much every year, and 2018 is no different. The preseason magazines are virtually evenly split between Alabama (2) and Clemson (1) for the top spot, with the Tigers having a very narrow edge. But, as far as the SEC goes, I don't expect much drama. The win distribution matrix is shown here:Bama's schedule this year seems extremely manageable, which explains the full one-game gap in expected conference wins between them and everyone else in the West. Their SEC East cross overs are Missouri (33) and at Tennessee (66) and they face the 3 next-best SEC West teams all at home: Auburn (8), Mississippi State (15), and Texas A&M (22). The Tide do travel to LSU (25) and my simulation actually does predict a loss here, but even so I have Alabama at 7-1 and 11-1 overall. The decisive game out West appears to be on 11/10 when Mississippi State comes to Alabama. An upset here would possibly knock Bama out of the SEC Title game, as the Bulldogs also have a favorable schedule (especially relative to Auburn, who must travel to Alabama, Georgia (4), and Mississippi State). But, let's be honest, even if Bama does lose to the "other MSU", they will likely still make the playoffs and still win the National Title.
In the SEC East, the picture is a bit more murky. Based on the win distribution matrix, Georgia (4) seems to have a similar advantage as Alabama, and it would be natural to assume that the Bulldogs will roll again to the SEC Title game. But, their schedule does appear to contain a few potential landmines. The first, and most important of which is the 9/8 SEC opener at South Carolina (26). The Gamecocks are projected to be just good enough to be dangerous and have just a manageable enough schedule that my simulation actually projects them to beat UGA and win the East with a record of 7-1 (10-2 overall, with a loss at Florida and at Clemson on the books). As for Georgia, even if they can beat the Cocks, they have some tricky contests at Missouri (33), at LSU (25), and vs. Florida (29). My simulation sees them tripping up at South Carolina and at LSU to finish at 10-2 overall. The Gators are potential dark horse as well. If they can beat Georgia, they would gain the tie-breaker upper hand over USC if the outcomes of all the others games are the same.
In this universe, Bama would likely destroy South Carolina in the SEC Title game to finish at 12-1, while 10-2 Georgia would be at home, wondering if they can still sneak into the playoffs.
4) Clemson is probably pretty good again this year, right?
Yup. As stated above, Clemson (1) gets the very slight edge in my consensus rankings as the preseason #1 team and they are quite likely to repeat as ACC Champs. The ACC win distribution is shown below.In conference play, the Tigers do travel to Tallahassee this year to face FSU (17) and my simulation does predict a loss here. But, I also have FSU finishing 5-3, so they do not look like a real threat. The only potential road bump is the 11/3 match-up with Louisville (32) visiting Clemson. The Cardinals have a more favorable schedule, and if they were to upset the Tigers, they would be in the driver seat to win the Atlantic Division. But, as the win distribution matrix shows clearly, Clemson is head-and-shoulders above every other team in the Division, and it would be a shock for them not to make the conference title game. Perhaps more important is the non-conference game in Week 2 where Clemson travels to Texas A&M (22). My simulation favors the Aggies here, which would put Clemson at 10-2 overall. Will that be enough to make the play-offs?
In the Coastal Division, the picture is a little less clear, maybe. Based on the win distribution matrix, it is clear that this division is a two-team race between Miami (12) and Virginia Tech (19). While Miami is projected to be a little bit better, than must travel to Blacksburg this year on 11/17, and Tech is likely to win the head-to-head match-up. But, both teams also play Florida State (17), with Miami getting them at home (likely a win) and Tech drawing them on the road (likely a loss). Both teams then have a couple of sneaky tricky conference road game. Miami travels to BC (38) and GA Tech (42), while VA Tech has to travel to Duke (43) and Pitt (48). On paper, those all should be wins, but the margin of error for both squads is razor thin. My simulation has both teams dropping one of those games (BC and Duke), so in the end it is the head-to-head result that will most likely prevail, sending VA Tech to the conference title game, where Clemson will have there way with them.
As a final note, once again the ACC has a large number (7) of non-conference match-ups with another Power 5 conference, the SEC. This is almost twice the number of inter-sectional match-ups between any other pair of Power 5 conferences. While the SEC has historically dominated this series, in 2018 the ACC looks like they will be likely favored in 4 of the 7 match-ups (South Carolina at Clemson, Florida at Florida State, Kentucky at Louisville, and LSU vs. Miami), while the SEC will only be clearly favored in 2 (Alabama vs. Louisville, and GA Tech at Georgia). The Clemson at Texas A&M game is a bit of a toss up, but Clemson is still likely to at least be favored. The overall result of this "ACC-SEC Challenge" is likely to be at least a footnote in the Playoff / NY6 selection process, and the ACC is poised to have the upper hand.
5) Was last year just a blip for the Big 12, or are they actually a real conference now?
I am going to go with "blip." After the disastrous 2016 season in which no one was even sure if the Big 12 deserved to be called a Power 5 conference, Oklahoma had a great year last year, making it all the way to the Playoffs. This year, the Sooners (7) start the season in the Top 10, and as the win distribution matrix shows, they have the highest expected value of wins of any Big 12 team again this season. But, trouble may come in as there is no shortage of other dangerous teams such as TCU (18), Texas (20), Oklahoma State (23), and West Virginia (28). Unfortunately for OK, of the teams above, they only draw OK State at home this year, and thus my simulation has them finishing at only 6-3 in conference play. In contrast, both Texas and TCU have slightly lighter schedules, and I project them to tie for the Big 12 Title at 7-2. While Texas hosts TCU this year in the regular season (and thus I project the Longhorns to prevail), the new (and completely unnecessary) Big 12 Championship game would dictate that Texas and TCU would have to play again and in this case, TCU would be slightly favored to win the conference. The bright side for the Sooners is that if they can beat either TCU or Texas, they would slide up to 2nd place and due to the new Championship game format, they would have a shot to still win the Title. So, I am going to designate the both the TCU - Oklahoma game and the Red River Shootout (Texas vs. Oklahoma) as the 2 pivotal games in the Big 12 this year. Honestly, Oklahoma getting at least one of those two "W"s seems quite likely. But, even so, their record would be only 10-2 and seeing that Ohio State has been replaced on the schedule with a very mediocre UCLA (47) squad as their marquee non-conference game, I doubt this would be enough to make the Playoffs.
6) Will there be any surprises in the Pac-12 this year?
Yes, I think that there will be. Based on the preseason rankings, Washington (5) is considered the team to beat, and the win distribution matrix above tends to support this. However, there is also a batch of teams right behind them including Stanford (14), USC (16), Oregon (24), Utah (27), and Arizona (35) who all have similar win distribution profiles. Because of the tight grouping of the teams, the schedule will once again likely be a factor. In this case, it is the Oregon Ducks that have the edge in the North, seeing as they face both Washington and Stanford at home this year. Washington does draw Stanford at home, which means the Washington at Oregon game on 10/13 is most likely decisive game in the North. My standings simulation has all three teams finishing at 7-2 in conference play, which give Oregon the spot in the Pac 12 title game based on head-to-head. Unfortunately for Washington, my simulation also projects them to lose to Auburn in Week 1, so even a NY6 Bowl might be out of reach if the Huskies do finish at 9-3. Similarly, Stanford's road trip to South Bend this year is likely also to give them a third "L" as well.
In the South, the competition is a bit weaker, but the analysis is very similar. The three main contenders for the south crown (USC, Utah, and Arizona) all play each other, and in this case Utah is the team with the schedule edge. Utah draws both teams at home, while USC is the team that must play both other teams on the road. Interestingly, Arizona does have an easier cross-over schedule (as they draw Oregon State, Washington State, and Oregon in the North while both Utah and USC draw Stanford on the road). At the end of the day, my simulation has Utah finishing at 7-2, a game ahead of both Zona and the Trojans. But, both 10/13 Arizona at Utah and 10/20 USC at Utah contests would swing control of the conference to the winner, with the Arizona at Utah game being perhaps a bit more significant. If this all comes to pass, Oregon would be favored to beat Utah in the Pac 12 Title game (in a regular season rematch) to finish at 11-2. But, there putrid non-conference schedule (no offense Bowling Green and San Jose State) and losses at Utah (in the regular season) and at Arizona would likely keep them out of the Playoff hunt.
7) Will Notre Dame be a factor in the Playoff hunt this year?
Perhaps. The Irish are on of the more interesting teams this year, as their schedule is an odd combination of manageable and tricky. Notre Dame has on its schedule six games against teams ranked in the top 35, and my projections suggest the spreads in all 6 games will be tighter than one TD. Because of all these potentially tight games, I only have their expected wins at 7.73 and their odds of winning 10 games or better is just below 20%. Similarly, my standings simulation projects losses at VA Tech (19), at USC (16), and a close lose at NW (34) to drop the Doomers to 9-3 overall. All this assumes wins vs. Michigan (10), at Wake Forest (44), vs. Stanford (14). If I stick with my simulation, ND is clearly out. But, realistically, NONE of those six teams really scare me, especially since UofM and Stanford need to come to South Bend. USC is good on paper, but that is often the only place that they are good, and it is hard to believe that the Wildcats will actually upset the Doomers even in their own stadium (which will be filled with ND alums). The VA Tech game will certainly be a challenge, but I could easily see the Irish at 11-1.
8) Which Group of Five team will get the NY6 Invite?
Most of the preseason magazines say Boise State, but there are a few other contenders. Based on my consensus rankings, the teams ranked in the Top 60 include Boise (21), UCF (30), Florida Atlantic (36), Memphis (41), San Diego State (46), Fresno State (49), Houston (57), Marshall (59), and USF (60). Not only is Boise State the highest ranked, they also have a very favorable schedule, as they draw both San Diego State and cross-over Fresno State at home. Outside of their Week 3 game at Oklahoma State (23), which is likely their only loss, their next toughest road game is at Wyoming (80). I project Boise to run the table in the Mountain West and to beat Fresno in the Title game (again) to finish at 12-1.In the AAC, both UCF and Memphis look to have fairly smooth sailing to their respective division titles, with UCF expected to lose at Memphis in a cross over, but with each team winning every other conference contest. UCF projects to win the rematch in the AAC Title game and also finish at 12-1. But, UCF in this scenario would score two Power 5 victories as they travel to UNC (58) in Week 3 and host Pitt (48) in Week 5, two games I project them to win. Memphis, incidentally, gets a shot at Missouri (33) on 10/20, but a win is unlikely. If both Boise and UCF win their conference titles and finish 12-1, it is a virtual coin flip as to which team would finish higher in the final playoff poll. Just for the sake of variety, I am going to go with Boise, but I don't feel great about it.
If the Mountain West and AAC both implode somehow, Florida Atlantic could make an interesting case for the NY6. I have them running the table in C-USA, but they actually get a shot at UCF in Week 4, and an upset could leave FAU at 12-1 to end the season, with the sole loss occurring in Week 1 at Oklahoma. The MAC does not look like a threat to the NY6 this year, as I project everyone to have at least 3 losses, with Northern Illinois beating Buffalo in the Title game. But, the Sunbelt could produce a 12-1 team in Arkansas State (71) or Appalachian State (75), depending on the winner of the 10/13 contest at Arkansas State and the new Sun Belt Championship game. But, neither team is at all likely to go undefeated, as they face Penn State and Alabama, respectively.
9) After all this analysis, who makes the Playoffs and the NY6?
If we assume that everything listed above comes to pass (which it won't) what would the final playoffs standings look like? Interestingly, no team is projected to finish the season undefeated, but there are several teams with only one loss:- Alabama: 12-1, SEC Champs, loss @ LSU (25), wins over Auburn (8), Miss St (15), A&M (22)
- Penn State: 12-1, Big Ten Champs, loss @ Michigan (10), wins over OSU (3), Wisconsin (6), MSU (13)
- Michigan State: 11-1, loss @ Penn State (9), wins over OSU (3), Michigan (10), NW (34)
- Boise State: 12-1, MWest Champs, loss @ OK State (23), wins over SDSU (46) and Fresno (49) 2x
- UCF: 12-1, AAC Champs, loss @ Memphis (41), wins over FAU (36), Memphis (41), Pitt (48), and UNC (58)
- Arkansas State: 12-1, Sunbelt Champs, loss @ Alabama (2), wins over App State (75) 2x
- Clemson: 11-2, ACC Champs, losses @ FSU (17) and A&M (22), wins over VA Tech (19), South Carolina (26), and Louisville (32)
- Ohio State: 10-2, losses @ PSU (9) and @ MSU (13), wins over Michigan (10) and TCU (18)
- Georgia: 10-2, losses @ LSU (25) and South Carolina (26), wins over Florida (29) and Missouri (33)
- Oregon: 11-2, Pac-12 Champs, losses @ Utah (27) and @ Arizona (35), wins over Washington (5) and Stanford (14)
- Miami: 10-2, losses at VA Tech (19), and Boston College (38), but wins over FSU (17), and LSU (25)
- Iowa: 11-2, Division Champs, both loses to PSU (9), wins over Wisconsin (6) and NW (34)
- TCU, 10-3, Big 12 Champs, losses to OSU (3), Texas (20), and WVU (28), wins over Oklahoma (7), Texas (20), and OK State (23)
So, how would this shake out? Two playoff teams would be very clear: Bama and Penn State, and although there would be some debate over the #1 vs. the #2 seed, I find it hard to believe that Bama would not get the 1-seed, even if Penn State's resume, honestly, looks a bit stronger.
As for the 3 and 4-seeds, as much as it makes me sound like a homer, if MSU does finish as one of only 3 teams in the Power 5 with 1 loss, and that 1 loss is to another Playoff team, MSU has to be in, right? For the final spot, Clemson, OSU, and Georgia would all seem to have the strongest cases, but the nod would most likely go to Clemson, based on the fact that they would have a conference title, and the overall strength of their resume. As for the final seeding, I just can't image that the committee would line up Clemson and Alabama in the semi's again with two Big Ten on the other side. So, I am guessing it would look like this:
- Cotton Bowl: #1 Alabama vs. #4 MSU
- Orange Bowl: #2 Penn State vs, #3 Clemson
As for the remaining four NY6 Bowls, the Rose and Sugar Bowl are contractually obligated to certain conferences, so I would project the following match-ups:
- Rose Bowl: Ohio State vs. Oregon (where OSU is favored)
- Sugar Bowl: Georgia vs. TCU (where UGA is favored)
- Washington: 9-3 with loses vs. Auburn (8), at Oregon (24), and at Utah (27), but wins over Stanford (14) and UCLA (47)
- Oklahoma: 9-3 with losses at TCU (18), Texas (20), and WVU (28), but wins over OK State (23) and Iowa St (37)
- Auburn: 9-3 with losses at Alabama (2), Georgia (4), and Miss State (15), but wins over Washington (5), Texas A&M (22), and LSU (25)
- Notre Dame: 9-3 with loses at USC (16), VA Tech (19), and NW (34), but wins over Michigan (10), Stanford (14), and Florida State (19)
- Mississippi State: 9-3 with loses at Alabama (2), LSU (25), and Kansas State (40), but wins over Auburn (8), Texas A&M (22), and Florida (29)
This one is a really tough call, but I think Auburn's projected non-conference win over Washington and the fact that 2 of their 3 losses are to other NY6 teams would give them the edge. So, the final two match-ups would be:
- Peach Bowl: Iowa vs. Miami (where Miami is favored)
- Fiesta Bowl: Auburn vs. Boise State (where Auburn is favored, but likely loses anyway)
10) What are the key games to watch out for, based on this analysis?
I like the following match-ups:- 09/01: Washington vs. Auburn (the winner is much more likely to grab a NY6 bid)
- 09/08: Georgia at South Carolina (the winner is likely to win the SEC East)
- 09/15: Boise State at Oklahoma State (If Boise scores the upset, they could run the table and if they do, will the committee put them in the playoffs?)
- 09/22: Wisconsin at Iowa (the winner has the inside track to Indy, which could have a huge impact on the entire Playoff and NY6 picture)
- 10/06: Oklahoma vs. Texas (If Oklahoma wins here, I think they will eventually win the Big 12)
- 10/13: MSU at Penn State (the winner has a great shot to win the Big Ten East)
- 10/13: Washington at Oregon (the winner has the inside track in the Pac 12 North)
- 10/13: Arizona at Utah (the winner is likely to win the Pac 12 South)
- 11/03: Louisville at Clemson (The Tigers can likely lock up the ACC Atlantic Division)
- 11/03: Notre Dame at Northwestern (this contest could mean the difference between the NY6 bid for the Domers)
- 11/10: Mississippi State at Alabama (the winner has the inside track to the SEC West)
- 11/17: Miami at VA Tech (Winner take all contest for the ACC Coastal Division)
12) Can we use any of this information in Vegas?
I have thought about this over the years and never came up with any foolproof method. But, the closest thing I have found are over-under numbers. Just for fun, here is a comparison of the over-under values that I found, relative to my expected value of wins for each team. As you can see, my values correlate very well to the O/U, but there are a handful of teams that deviate by more than a full game. This year, I set the cut-off as 1.2 games over or under. So, in other words, my spread says to take the over on UCF, Memphis, South Carolina, and Northwestern, but take the under on Georgia Southern, Syracuse, Bowling Green, SMU, Minnesota, and Nevada. Use at your own risk!
I should also note for completeness that the values above differ from the ones in the win distribution tables because above I use the less conservative value that I get before I apply the correction for the odds that the preseason ranking are wrong. If I were to use the more conservative numbers, it just tends to drag down the O/U for the top rated teams and bump up the O/U for the low ranked teams. Those teams basically have no where to go but down or up and so those number tend to regress more to the middle.
Well, that is all for now. Enjoy!
Comments
Post a Comment