Ensemble predictions for week 9 of college football

ncaaf2014Last week, I wrote about how neither Florida State nor Notre Dame would make the college football playoff. One piece of evidence was that Florida State had less than 50% chance to win at Miami by my numbers.

People called bullshit on that one. A commenter said:

If your numbers say Miami over FSU, you ain’t got good rithm

I think he meant algorithm.

It’s easy to understand why people doubt numbers that predict Miami over Florida State. Miami has lost to Louisville, Nebraska, and Georgia Tech. They start a true freshman, Brad Kaaya, at quarterback.

Breaking down Miami into offense and defense

However, The Power Rank looks deeper at college football teams by breaking them down into offense and defense. It starts with a statistic like yards per play, an efficiency metric for football mostly immune from the randomness of turnovers. My algorithm adjusts yards per play for strength of schedule to rank offense and defense.

Before last night, Miami had the 12th and 14th ranked offense and defense respectively. While Kaaya will probably make freshman mistakes, he has weapons like Duke Johnson to break off big plays. The defense is solid against both the pass and rush.

These adjusted yards per play numbers also predict a point spread for any game. Before last night, yards per play predicted a 11.7 point win for Miami at Virginia Tech.

That margin was probably too large. This season, I’ve been working with Mike Craig to aggregate many different predictions for a better ensemble prediction. In addition to yards per play, we use calculations based margin of victory and data from the markets. The public predictions on The Power Rank is just one member of the ensemble.

This ensemble predicted a 1 point win for Miami over Virginia Tech. The line opened favoring Virginia Tech by 3 points before ending with Miami as a 2 point favorite.

Miami thrashed Virginia Tech 30-6 last night. The Hurricanes gained 6.61 yards per play compared with 4.44 for Virginia Tech. Miami led 24-0 at half without the benefit of any Virginia Tech turnovers.

It doesn’t always work out so nicely in the random world of college football. But yards per play has an amazing power to identify undervalued teams at this point in the season.

Let’s at some other ensemble predictions for this weekend.

Mississippi at LSU

Mississippi will beat LSU by 3.3. The teams will score 47.5 points.

Life is never easy in the SEC West. Ole Miss is undefeated with a huge win over Alabama. They attempt to continue their fairy tale season at LSU on Saturday.

Even though LSU has lost twice already, they are still solid on both sides of the ball. The Tigers rank 35th and 32nd on offense and defense respectively in my adjusted yards per play.

Ole Miss has a lights out defense ranked 2nd. On offense, they have struggled to run the ball, so they rely on the arm of quarterback Bo Wallace. Their passing game ranks 14th in adjusted yards per attempt.

Grantland asked me to predict the Heisman winner for their mid-season predictions article. These predictions make me uncomfortable because my analytics do not directly apply.

I picked Wallace to win the Heisman. He has a great defense to help win games, and he has increased his completion percentage 66% this season. (The blond hair probably helps as well.) If my prediction has any chance of coming true, Wallace must have a Heisman moment at LSU to win the game. Perhaps a leap over a defender for a go ahead touchdown?

Ohio State at Penn State

Ohio State will beat Penn State by 5.8. The teams will score 50.4 points.

Matt Hinton of Grantland had a nice article on Ohio State QB J.T. Barrett. The freshman has been impressive recently, and some think he’s better than the injured Braxton Miller. With only six games played for Barrett, that’s absurd.

It’s easy to get excited about a small sample size, especially when two of Ohio State’s opponents were Rutgers and Cincinnati. Remember when Texas A&M freshman Kenny Hill destroyed South Carolina in their first game and jumped to the front of Heisman lists? Hill has stumbled against the elite defenses in the SEC West.

Barrett faces a Penn State defense ranked 16th, a unit which will make it difficult for Barrett to put up the same numbers he did against Cincinnati and Rutgers. Ohio State probably wins this game at Happy Valley, but it won’t be easy.

Oregon at California

Oregon will beat California by 11.8. The teams will score 75.8 points.

Offensive linemen usually don’t get much credit. However, tackle Jake Fisher got props from CBS writer Rob Rang for jump starting Oregon’s offense when he returned from injury against UCLA.

I didn’t see that game, but Oregon’s offensive line looked terrible against Arizona. And I like it when the big guys up front get credit.

Oregon should take care of business against an improved California team.

Michigan at Michigan State

Michigan State will beat Michigan by 14.4. The teams will score 45.8 points.

For Michigan to capitalize on their 15% chance to beat Michigan State, quarterback Devin Gardner must put on the superman cape that he lost after last year’s Notre Dame game.

Sample of other ensemble predictions

In the following, the first number gives the point spread for the home team; a negative number implies a victory for the home team. The second number gives total points scored in the game.

UNLV at Utah State: -20.7, 54.5.

Northern Illinois at Eastern Michigan: 18.1, 62.4.

Texas Tech at TCU: -20.8, 68.3.

Central Michigan at Buffalo: 3.3, 56.7.

Akron at Ball State: 2.6, 50.9.

UCLA at Colorado: 18.9, 65.1.

West Virginia at Oklahoma State: 2.5, 64.5.

Vanderbilt at Missouri: -22.3, 45.0.

Maryland at Wisconsin: -7.9, 55.7.

Louisiana Tech at Southern Miss: 6.8, 52.1.

Temple at UCF: -11.7, 51.2.

UAB at Arkansas: -17.2, 59.1.

Massachusetts at Toledo: -17.6, 68.8.

Texas State at Louisiana Monroe: -2.1, 49.6.

Minnesota at Illinois: 2.1, 56.0.

UTEP at UTSA: -19.1, 52.2.

Texas at Kansas State: -10.4, 44.6.

Kent State at Miami (OH): -9.1, 51.9.

Troy at South Alabama: -16.0, 54.1.

Members of The Power Rank have access to these ensemble predictions and my yards per play calculations. To learn more, sign up for the free email newsletter. Just enter your best email and click “Sign up now.”

World Series win probability for 2014

world_series_winprobAt the end of the regular season, I ranked baseball teams based on expected runs scored and allowed. The formula comes from Dave Smyth’s Base Runs, which on average estimates a team’s season run total within 1%.

The first number after a team gives the differential in expected runs, followed by offense and defense in parentheses. The record comes from baseball’s Pythagorean theorem with an exponent 1.83.

1. Los Angeles Angels, 131.99. (728.60, 596.61). Record: 95-67.
2. Washington, 125.88. (695.92, 570.04). Record: 95-67.
3. Oakland, 117.29. (684.77, 567.49). Record: 94-68.
4. Los Angeles Dodgers, 115.87. (737.74, 621.87). Record: 93-69.
5. Pittsburgh, 91.53. (724.80, 633.26). Record: 90-71.
6. Baltimore, 72.00. (718.53, 646.54). Record: 88-74.
7. Detroit, 59.78. (771.46, 711.68). Record: 86-76.
8. San Francisco, 52.06. (657.84, 605.78). Record: 86-75.
9. Tampa Bay, 35.93. (642.20, 606.27). Record: 85-77.
10. Seattle, 31.37. (606.33, 574.97). Record: 84-78.
11. Toronto, 28.01. (735.04, 707.04). Record: 83-79.
12. Cleveland, 22.12. (686.96, 664.84). Record: 82-79.
13. St. Louis, 16.59. (628.22, 611.63). Record: 82-80.
14. Kansas City, -4.46. (634.16, 638.61). Record: 79-82.
15. Milwaukee, -5.71. (659.09, 664.80). Record: 80-82.
16. New York Yankees, -20.64. (638.51, 659.14). Record: 78-84.
17. Chicago Cubs, -25.19. (634.64, 659.83). Record: 77-84.
18. Atlanta, -28.47. (602.24, 630.71). Record: 77-85.
19. Miami, -29.88. (662.45, 692.33). Record: 77-85.
20. Colorado, -34.93. (784.56, 819.49). Record: 77-84.
21. New York Mets, -46.29. (620.00, 666.29). Record: 75-87.
22. Houston, -49.93. (648.16, 698.10). Record: 75-87.
23. San Diego, -55.98. (539.81, 595.78). Record: 73-89.
24. Cincinnati, -58.94. (582.96, 641.90). Record: 73-89.
25. Chicago White Sox, -69.15. (670.81, 739.96). Record: 73-89.
26. Minnesota, -71.83. (700.10, 771.92). Record: 73-89.
27. Philadelphia, -72.87. (611.56, 684.42). Record: 72-90.
28. Boston, -80.61. (647.36, 727.97). Record: 72-90.
29. Arizona, -111.52. (620.73, 732.25). Record: 68-94.
30. Texas, -150.93. (629.36, 780.30). Record: 65-97.

Kansas City had a negative run differential by expected runs. They should have lost more games than they won during the regular season. Now the Royals are playing in the World Series.

San Francisco had the 4th best run differential among 5 NL playoff teams. They beat that 5th team (St. Louis) to return to the World Series.

Playoff baseball is crazy.

World Series Prediction

My numbers gives San Francisco a 53% chance to beat Kansas City in the World Series. The markets favorite a Kansas City team that has not yet lost a game in the playoffs.

Kansas City has home field because the AL won the All-Star game. If San Francisco had home field, their win probability goes up to 55%. I thought the gap would be bigger, but the 0.3 runs I use for home field advantage has a small effect on series odds.

The series win probabilities start with my MLB team rankings, which take raw run differential and adjust for strength of schedule. Also, I adjust for cluster luck based on the regular season.

In addition, the projections consider starting pitching through xFIP, an ERA type statistics that captures the skill of a pitcher through strike outs, walks and fly ball rate.

Daily predictions for each game appear on the predictions page.

Why neither Florida State nor Notre Dame will make the 2014 college football playoff

The Power Rank's top 10 as of Oct 16th, which doesn't include Florida State or Notre Dame.

The Power Rank’s top 10 as of Oct 16th, which doesn’t include Florida State or Notre Dame.

You expect the winner of the Notre Dame and Florida State to make the first college football playoff. It might be the easiest prediction of the season given their unblemished records and the recent track record of these programs.

Florida State won the national title last season and returns Heisman winning QB Jameis Winston. Only an incredible run of wins by Mississippi State has knocked them from first in the AP poll.

Notre Dame played in the national championship game two years ago. Now Everett Golson, the quarterback that led the Fighting Irish during their championship run, has returned after getting removed from school for year. Notre Dame seems like an elite team again.

The past few season has brought a return to former greatness for both programs. This visual shows a 30 year history of Florida State.

Florida State

The bottom panel shows a rating, or an expected margin of victory against an average team from The Power Rank’s algorithm.

Former Florida State Bobby Bowden took the program to incredible heights before regressing in the new millennium. Current coach Jimbo Fisher had the Seminoles on an upward trajectory before making a huge leap last season.

Notre Dame has also enjoyed some terrific seasons over the last 30 years.

Notre Dame

Former coach Lou Holtz capture the national championship in 1988 and had amazing teams in 1989 and 1993. The program went in decline after his departure in 1996 except for a one year blip in 2005 (Charlie Weis’s first year with Tyrone Willingham’s players). Current coach Brian Kelly had the Fighting Irish on the rise until a slight drop last season.

Despite this return to greatness for these programs, neither Notre Dame nor Florida State will make the college football playoff this season. Let me explain.

Florida State

No, my pessimism towards Florida State isn’t based on the high likelihood Jameis Winston gets kicked off the team. The news cycle has brought us constant stories about his problems over a rape case and whether he took money for signing autographs.

Florida State’s offense has been good this season. The Power Rank takes yards per play and adjusts for strength of schedule through a proprietary algorithm to rank offense and defense. These rankings put the Seminole offense at 6th in the nation. They might even be better, as this calculation includes the Clemson game in which Winston didn’t play.

Florida State’s problems are on defense. In the past two season, this unit has ranked in the top 5 by my adjusted yards per play. However, they have dropped to 35th this season.

There are many possible reasons for this decline. Florida State lost tackle Tim Jernigan and cornerback LaMarcus Joyner to the NFL draft. Coordinator Jeremy Pruitt left for Georgia, who pitched a shutout at Missouri this past weekend.

The numbers single out the pass defense, which ranks 51st by yards per pass attempt adjusted for schedule. Florida State can’t generate a pass rush as they have sacked the quarterback on 4.2% of pass attempts compared to a 6% FBS average.

To confirm these numbers, I watched the first half of Florida State’s game against Syracuse. The Seminoles only pressured the quarterback three times on 18 pass attempts. None of these pressures led to a sack, although one did result in an interception.

With this struggling defense, Florida State becomes vulnerable to a few of their conference foes. (Well, they were also vulnerable 2 seasons ago when they lost at North Carolina State.) My best predictions at The Power Rank come from aggregating a number of predictions based on stats such as margin of victory and yards per play.

These ensemble predictions make Florida State an underdog at Louisville (44% win probability) and Miami (38%). Louisville has the 2nd best defense in the nation, while Miami looks strong on both sides of the ball despite 3 losses already.

Notre Dame

The Everett Golson story has been heart warming. The kid makes a mistake and cheats on a test. Notre Dame finds out and kicks him out of school, causing him to miss the 2013 season. Then Golson comes back and has led the Fighting Irish to a 6-0 record this season.

However, the numbers suggest that Notre Dame’s offense is not any better than last season. With Tommy Rees at quarterback last season, Notre Dame ranked 23rd in yards per pass attempt adjusted for schedule. This season, Notre Dame ranks 22nd.

On both offense and defense, Notre Dame looks the same as last season if not slightly worse. In my numbers, their offense is ranked 35th after ending last season 28th. Their defense is ranked 31st after ending last season 28th. For comparison, one loss Alabama has the 11th and 7th ranked offense and defense respectively this season.

In addition, Notre Dame hasn’t played any quality teams except Stanford. They beat Stanford when converted a 4th and 11 for a go ahead touchdown when Stanford coach David Shaw inexplicably rushed 3 guys (no, this Stanford alum is not bitter. Not at all.). Notre Dame also won a strange 31-0 game against Michigan in which they had fewer total yards than Michigan.

Even if they beat Florida State, Notre Dame will find their remaining schedule difficult. My numbers have them as underdog at USC and Arizona State, two teams that have struggled this season. They also have a slim 51% win probability against Louisville at home.


Notre Dame at Florida State should be a fantastic game. The environment will be electric, the offenses will score points.

The best predictions at The Power Rank come from aggregating a number of different predictions. In collaboration with Mike Craig, we use my numbers as well as data from the betting markets and other trusted rankings. This ensemble method predicts a 9 point win for Florida State, which implies a 25% win probability for Notre Dame.

However, these are not two of the best five teams in the nation. Florida State might sneak back into the top 5 if their defense returns their level of play over the last two seasons. The numbers suggest Notre Dame is a solid top 25 team but nothing more.

Expect neither Florida State nor Notre Dame to end up in college football’s first playoff.

Championship series win probabilities for the 2014 MLB playoffs

boy_hitting_foreheadIt’s an embarrassment, but I’m not going to hide from it. All the favored teams by my numbers (and the markets) lost in the division series.

And it’s not like the favored teams lost in a game 7 coin flip. The favored teams won two games (Dodgers over Cardinals, Nationals over Giants).

As Billy Beane said about his analytics in Moneyball, “my shit doesn’t work in the playoffs.”

Someone tweeted that quote at me before the division series started. I blew it off. If you’re betting on heads, you want the coin to come up heads 52% instead of 50% of the time. Always.

But Billy’s words have new meaning after the division series. It’s not that analytics don’t work in the playoffs. It’s that we should appreciate the randomness of a short series.

Let’s also remember that the team that wins the World Series might play more playoff games than a college football team does all season.

Here are numbers for the championship series.

  • San Francisco has a 52.6% chance to beat St. Louis. No, there is Cardinal Devil Magic in this prediction.
  • Baltimore has a 56.6% chance to beat Kansas City.

These win probabilities start with my MLB team rankings, which take run differential and adjust for strength of schedule. Also, for the first time, I adjust for cluster luck based on the regular season.

In addition, the projections consider starting pitching through xFIP, an ERA type statistics that captures the skill of a pitcher through strike outs, walks and fly ball rate.

Daily predictions for each game appear on the predictions page.

As of noon Eastern on October 10th, the markets give both San Francisco and Baltimore an implied odds of 55.1%.

Ensemble predictions for college football, week 6, 2014

Over the last year, I’ve started aggregating many predictions to a single prediction. Research in diverse areas shows that this ensemble of predictions gives better predictions.

As the college football season continues, I’ve been working on ensemble game predictions for members of The Power Rank. These predictions aggregate not only my calculations but also other trusted sources. Some of these predictions use margin of victory while others use statistics such as yards per play and yards per pass attempt.

The predictions below also included totals (total points scored in the game). These calculations are a collaboration with Mike Craig, my partner in the college football prediction service.

Here are 3 interesting predictions on a terrific slate of games this Saturday.

Stanford at Notre Dame

Stanford will win by 4.1 points. Stanford and Notre Dame will score 43 points.

Stanford has struggled with mistakes in their two biggest games. They couldn’t punch the ball in the endzone against USC, losing by 3 on a 53 yard field goal. Washington returned a Stanford fumble for a touchdown last weekend, even though Stanford survived for the win in Seattle.

These mistakes have little affect yards per play statistics. Hence, Stanford looks better by these numbers. For example, yards per play predicts a 8 point road win over Notre Dame. This visual shows how Stanford matches up with Notre Dame by yards per play adjusted for schedule.

(In the visual, better defenses appear further to the right. This facilitates comparisons, as the unit further to the right is predicted to have an advantage.)

Screen shot 2014-10-03 at 2.35.32 PM

The ensemble likes Stanford to beat Notre Dame. No, there’s no personal bias in these numbers from this Stanford alum.

Nebraska at Michigan State

Michigan State will win by 4.1 points. Nebraska and Michigan State will score 57.5 points.

This Big Ten showdown features Nebraska’s 5th ranked offense against Michigan State’s usually strong defense.

Screen shot 2014-10-03 at 2.31.01 PM

The yards per play numbers in the visual use data from last season since rankings with only data from this season are volatile. For example, Michigan State had the 28th ranked defense yesterday, a low but potentially believable ranking for a defense that was elite last season.

Then Oregon’s offense has a terrible game against Arizona’s defense last night. Since Oregon played Michigan State earlier this season, Michigan State’s defense drops to 57th in rankings that only use this year’s data. With some input from last year’s games, a ranking of 26th seems more reasonable.

Ohio State at Maryland

Ohio State will win by 0.2 points (a 50-50 game). Ohio State and Maryland will score 59 points.

The predictions are all over the map for this game. Maryland looks the equal of Ohio State by yards per play. Both offenses have a slight edge, as shown in this visual.

Screen shot 2014-10-03 at 2.33.38 PM

Yards per play favors Maryland by 3 points. My model gives the home team 3 points, so this prediction says Ohio State and Maryland are equal teams.

However, the markets favor Ohio State by 8.5 points. Some of this advantage probably comes from the history and tradition of Ohio State. However, this spread also considers injuries. Maryland QB C.J. Brown is listed as questionable, and Maryland has suffered a rash of other injuries on both sides of the ball.

Become a member of The Power Rank

Members have access to all the ensemble predictions as well as interactive versions of these match up visuals. To learn more about my methods, sign up for my free email newsletter. Enter your email and click on “Sign up here.”