NFL Picks – Week 11
Total (weeks 1-11) – SU: 111-49-1 (69.25%) ATS: 81-78-2 (50.93%, -4.8 units) O/U: 84-75-2 (52.80%, +1.5 units)
Week 1 – SU: 9-7-0 ATS: 8-8-0 O/U: 13-3-0
Week 2 – SU: 10-6-0 ATS: 10-6-0 O/U: 10-6-0
Week 3 – SU: 12-4-0 ATS: 9-6-1 O/U: 8-8-0
Week 4 – SU: 7-6-0 ATS: 5-7-1 O/U: 5-8-0
Week 5 – SU: 14-2-0 ATS: 6-9-0 O/U: 9-6-0
Week 6 – SU: 11-3-1 ATS: 8-7-0 O/U: 6-9-1
Week 7 – SU: 11-4-0 ATS: 7-8-0 O/U: 8-7-0
Week 8 – SU: 11-3-0 ATS: 8-7-0 O/U: 8-7-0
Week 9 – SU: 9-4-0 ATS: 8-5-0 O/U: 4-8-1
Week 10 – SU: 9-4-0 ATS: 4-9-0 O/U: 6-7-0
Week 11 – SU: 9-5-0 ATS: 8-6-0 O/U: 7-7-0
Buffalo at Miami
Prediction: Dolphins 23-20
Pick: Bills +5.5 (57.09%)
Total: Over 42
Detroit at Arizona
Prediction: Cardinals 22-21 (52.0%)
Pick: Detroit +1.5 (52.26%)
Total: Over 42
Atlanta at Carolina
Prediction: Panthers 24-21 (57.5%)
Pick: Panthers -2.5 (50.46%)
Total: Under 46.5
Minnesota at Chicago
Prediction: Bears 24-20 (63.4%)
Pick: Bears -3.5 (53.71%)
Total: Under 46.5
Houston at Cleveland
Prediction: Texans 22-21 (50.2%)
Pick:Texans +3.5 (60.04%)
Total: Over 41.5
Philadelphia at Green Bay
Prediction: Packers 27-23 (60.9%)
Pick: Eagles +5.5 (54.61%)
Total: Under 55.5
Seattle at Kansas City
Prediction: Seahawks 21-19 (56.1%)
Pick: Seahawks +1.5 (60.31%)
Total: Under 42
Cincinnati at New Orleans
Prediction: Saints 27-23 (63.6%)
Pick: Bengals +7.5 (57.41%)
Total: Under 51
San Francisco at NY Giants
Prediction: 49ers 22-20 (54.3%)
Pick: Giants +4.5 (58.48%)
Total: Under 44.5
Oakland at San Diego
Prediction: Chargers 26-18 (69.6%)
Pick: Raiders +10.5 (59.42%)
Total: Under 45
Denver at St. Louis
Prediction: Broncos 27-21 (65.8%)
Pick: Rams +9.5 (60.75%)
Total: Under 51.5
Tampa Bay at Washington
Prediction: Washington Football Team 24-20 (62.4%)
Pick: Buccaneers +7.5 (58.72%)
Total: Under 45.5
New England at Indianapolis
Prediction: Patriots 27-26 (52.2%)
Pick: Patriots +2.5 (59.26%)
Total: Under 58.5
Pittsburgh at Tennessee
Prediction: Steelers 22-21 (52.2%)
Pick: Titans +5.5 (63.25%)
Total: Under 46.5
College Football Playoff = AP poll + RPI
A few weeks ago the College Football Playoff (CFP) committee published their first rankings. That led to this train of though:
- Wow. Those look a lot like the AP rankings.
- I wonder what rankings the CFP is closest too?
- Where can I get a bunch of different ratings to compere? MASSEY!
So, I downloaded all of the ratings that Massey had collected and compared each of them to the CFP rankings using Kendall’s Tau. And guess what? The AP rankings are the most highly correlated rankings with the CFP with tau=.8478261. So basically, we’re back to the days when the AP picked the national champion? Well not exactly. Now we have the AP AND the RPI. Because guess what’s next most highly correlated? Real Time RPI! At tau=.833333333333, the RPI falls just behind the AP in the most highly correlated rankings. This is rather alarming, because basically anyone who has any idea what they are doing, knows the RPI is terrible. (See this, this, and this.) To highlight this, I’ll take a quote from that second link from Nate Silver in regards to the RPI in terms of selecting NCAA basketball teams:
Over the long run, R.P.I. has predicted the outcome of N.C.A.A. games more poorly than almost any other system.
Basically, the RPI is terrible, everyone who knows what they are doing knows it, but it still gets used because….I have no idea.
Following the RPI, you have the USA Today Coaches poll with a tau of .8115942. So basically the CFP is using some sort of secret combination of two human polls and basically the worst non-human ranking that is available. As opposed to the BCS which was a known combination of two human polls and 6 computer systems that ranged from very good to dreadful. (At least the BCS didn’t involve completely inexplicable people like Condoleezza Rice. Seriously, why is she involved in this? If you really want a woman on the committee aren’t there literally thousands of women more qualified for this than her?)
Also, for the sake of it, I made a decision tree to try to predict the CFB rankings. It looks like the formula for getting into the playoff is be top 15 in coaches poll and then be top 3 in RPI. The three variables are the USA Today coaches poll (X.USA), the AP poll (X.AP), and Real Time RPI (X.RTR). My response was the CFP rank with all teams not ranked set to 26.
All of this of course would infuriate me if the NCAA was earnestly trying to find the 4 best teams in college football to put into a playoff. But I suspect that’s not necessarily their goal. The NCAA is about one thing: MONEY. They can say whatever they want, but they are ruled by money (none of which goes to the actual people who produce the product!). Because of this, this sham process somehow bothers me less. The NCAA is trying to get the 4 “best” teams into the college football playoff. They are just using their definition of best, which is probably much different than that of the average fan. And besides I’ve got bigger problems with the NCAA. (See here, here, here, here, here, and here for starters.)
Cheers.
NCAA Football Rankings – 11/10/2014
Mississippi State stays number 1 virtually everywhere (including my rankings) after remaining undefeated at 9-0. My rankings have Alabama and Ole Miss moving up to numbers 2 and 3 after an OT victory over LSU and a blowout win over Presbyterian, respectively. Auburn moves in the other direction, dropping 2 spots to number 4 after losing to Texas A and M. So, if I were choosing the college football play-off, I’d take 4 teams from the SEC West. Will this happen? No. For starters, a number of these top 4 teams have to play (beat the crap out of each other) in the next few weeks (i.e Miss St-Alabama, the Egg Bowl, Alabama-Auburn). As far as I am concerned, all of these are de facto play-off games.
The rest of my top 10 includes: TCU, Oregon, UCLA, Baylor, Florida St, and Georgia. TCU is up 2 spots with it’s win over Kansas State (who fell 3 spots and out of my top 10). Oregon remains in 6th, with UCLA up 2 spots to 7th. Baylor made the biggest move into the top ten this week jumping 11 spots to number 8 after their destruction of Oklahoma. If Baylor and TCU win out, I see Baylor being ranked higher than TCU as a results of Baylor’s head-to-head win over TCU and Baylor’s remaining game against Kansas State. TCU’s remaining schedule is Kansas, Texas, and Iowa State whose combined records are 10-18. TCU basically can’t improve, whereas Baylor has one big game left to help it jump into the top 4. Florida state continues to slowly climb the rankings moving from 10 to 9. They are hurt by there very weak schedule, and that “big win” over Notre Dame looks a lot less impressive after Notre Dame got blown out by Arizona State. If they go undefeated, their obviously in playoff. No questions asked; No matter how weak their schedule. Rounding out the top 10, I have Georgia after a blowout victory over Kentucky a week after recovering from their dreadful loss to the Florida Gators.
As for teams that dropped out of my top 25 this week: Notre Dame fell from 14 to 27; Michigan State fell from 21 to 32; Louisiana Tech fell from 25 to 34.
An finally, congratulations to New Mexico State for moving out of the basement, swapping places with Georgia State. Go team!
Full Rankings: http://wp.me/PlZJR-Hv
|
NFL Picks – Week 10
Total (weeks 1-10) – SU: 102-44-1 () ATS: 73-72-2 () O/U: 77-68-2 ()
Week 1 – SU: 9-7-0 ATS: 8-8-0 O/U: 13-3-0
Week 2 – SU: 10-6-0 ATS: 10-6-0 O/U: 10-6-0
Week 3 – SU: 12-4-0 ATS: 9-6-1 O/U: 8-8-0
Week 4 – SU: 7-6-0 ATS: 5-7-1 O/U: 5-8-0
Week 5 – SU: 14-2-0 ATS: 6-9-0 O/U: 9-6-0
Week 6 – SU: 11-3-1 ATS: 8-7-0 O/U: 6-9-1
Week 7 – SU: 11-4-0 ATS: 7-8-0 O/U: 8-7-0
Week 8 – SU: 11-3-0 ATS: 8-7-0 O/U: 8-7-0
Week 9 – SU: 9-4-0 ATS: 8-5-0 O/U: 4-8-1
Week 10 – SU: 9-4-0 ATS: 4-9-0 O/U: 6-7-0
Cleveland at Cincinnati
Prediction: Bengals 25-19 (65.8%)
Pick: Browns +6.5 (52.3%)
Total: Under 46
St. Louis at Arizona
Prediction: Cardinals 23-19 (61.6%)
Pick: Rams +7 (58.1%)
Total: Under 43.5
Tennessee at Baltimore
Prediction: Ravens 24-18 (66.3%)
Pick: Titans +10 (61.5%)
Total: Under 44
Kansas City at Buffalo
Prediction: Bills 21-20 (53.3%)
Pick: Bills +2.5 (60.3%)
Total: Under 42
Miami at Detroit
Prediction: Lions 24-20 (60.1%)
Pick: Lions -3 (51.7%)
Total: Under 44
Chicago at Green Bay
Prediction: Packers 26-22 (62.5%)
Pick: Bears +7.5 (58.5%)
Total: Under 53.5
Dallas at Jacksonville
Prediction: Cowboys 24-20 (61.8%)
Pick: Jaguars +7.5 (59.3%)
Total: Under 45
San Francisco at New Orleans
Prediction: Saints 25-22 (56.1%)
Pick: 49ers +5.5 (59.4%)
Total: Under 49.5
Pittsburgh at NY Jets
Prediction: Steelers 21-20 (51.1%)
Pick: Jets +6 (65.6%)
Total: Under 46
Denver at Oakland
Prediction: Broncos 28-20 (70.3%)
Pick: Raiders +12 (62.7%)
Total: Under 50
Carolina at Philadelphia
Prediction: Eagles 25-22 (58.7%)
Pick: Panthers +6.5 (59.7%)
Total: Under 49
NY Giants at Seattle
Prediction: Seahawks 25-17 (69.3%)
Pick: Giants +9.5 (56.9%)
Total: Under 45.5
Atlanta at Tampa Bay
Prediction: Falcons 22-21 (50.7%)
Pick: Buccaneers +3 (57.8%)
Total: Under 46
2014 Senate Election Prediction Results
Nate Silver and Sam Wang are two of the most popular prognosticators of elections in the United States. They are also known to be openly critical of each other’s methods. But now that the elections are over, let’s take a look and see how each of them fared.
Wang suggests using the Briere Score to evaluate predictions, so I’m going to use that as one measure. This measure compares the predicted win probabilities to the actual outcomes and takes the average the squared differences. I also looked at the mean squared error of the predicted margins of victory as another measure of accuracy.
Silver amassed a Briere score of .14656 based on the 12 closest senate elections (Kentucky, Arkansas, Louisiana, Georgia, Colorado, Alaska, Iowa, Kansas, North Carolina, New Hampshire, Minnesota, and Virginia). For those same races, the Silver put up a mean squared error of 38.727 for margins of victory. Wang fared worse by both these measures with a Briere score of .17018 and a mean squared error of 44.981.
I’ve summarized the results in the graphic below. Each square represent the squared error of margin of victory within a state, and the color of the square corresponds to which direction Wang or Silver was wrong in. For instance, Silver predicted a 5 point win for the Republicans in Arkansas, but they actually won by 17.13. So the square is very red because they Republicans exceeded Silver’s expectations. Alternatively, Louisiana was just slightly bluer than both Wang and Silver projected, so the square gets a slight hint of blue.
So I’m declaring Silver the winner of this round in the batter of Silver vs Wang.
However, if you look at all of the other popular senate predictions, Silver was not the winner. According to Sam Wang, that title goes to Drew Linzer of the DailyKos for his Senate predictions based on Briere score and followed by the Washington Post. FiveThirtyEight came next in a group that included the Huiffington Post and Betfair.com with Briere scores of 0.14. (These Briere scores are slightly different than the ones I calculated, as I used a few more states.) This was then followed by the Upshot and the Princeton Election Consortium. So while Silver, the highest profile of all the prognosticators, won the Silver vs Wang battle, he sort of finished in the middle of pack in terms of predictions. At least he can take solace in the fact that he beat his old employer!
Cheers.
We always think we’re right, but we don’t think we’re always right.
Jordan Ellenberg on how many states Nate Silver is going to get wrong, according to Nate Silver. (This refers to the elections of US Senators taking place tomorrow.) For each state Silver gives a probability of winning; we can give a probability that Silver will be wrong which is just his own predicted probability that the underdog wins. The answer is an an expected value of 2.5. Silver has been saying since the 2012 election that he got lucky in calling all fifty states correctly. In some sense it would have been more impressive if he’d missed a couple, which would have shown his predictions were calibrated correctly. (I remember trying to explain this to colleagues at my job at the time, where I’d been for a bit over a month; I think I did so successfully, but it’s a subtle point.)
Silver’s famous 50-for-50 2012 presidential predictions are
View original post 107 more words
4 College Football rankings I have no faith in: AP, USA Today, Playoff Committee and Fremeau
Massey has compiled a list of 114 different college football ratings for comparison. Go take a look at them. Now go to the bottom and look at the conference rankings. Scroll across the page and you’ll see that the SEC is literally the unanimous….wait, what? Only 110 of the 114 rankings have the SEC rated number 1.
This means there are four that rankings don’t have the SEC at the top. Which rankings are these? 3 of them are polls: The AP Poll, the USA Today Poll, and the College Football Playoff Committee. However, you rank college football teams this year, if you don’t have the SEC rated as the best conference, you’re ranking is absolutely meaningless in my book (and I think many people would agree). It’s absolutely insane that the three polls that get this wrong are 1/3 of the old BCS, a coaches poll, and the rankings that will determine the college football playoffs. (At this point, is it reasonable to assume the the NCAA is just messing with college football fans in an elaborate work of performance art?) It doesn’t take a rocket scientist to see that the SEC is the best conference in college football this year. If you’re rankings don’t reflect that, it’s simply hard to have any faith that you know what you are doing.
But wait, it gets worse. Something called FEI has the SEC ranked THIRD (?!?!) best conference in football this year behind the Pac-12 and the ACC. What in the world is the FEI? FEI is apparently and abbreviation for the Fremeau Efficiency Index. Apparently this is one of the two components that Football Outsiders uses in it’s F/+ ratings. Now I swear to you, I’m not actively going out of my way to pick on Football Outsiders for bad statistics, but it just seems to keep happening. (See for instance, former FO writer Bill Barnwell doing a terrible “study” and some of their “analysis” of place kicking.) Now add in that their overall ranking system for college football has the SEC ranked 3rd and it’s even harder to take their slogan seriously: “Innovative Statistics, Intelligent Analysis”.
There is a fine line between genius and insanity and Football Outsiders are either geniuses who see something in the Pac-12 and the ACC that literally no one outside of the media polls sees or they are living in a completely alternate college football reality.
Cheers.
NFL Picks – Week 9
Total (weeks 1-9) – SU: 93-40-1 (69.8%) ATS: 69-63-2 (52.2%, -.2 units) O/U: 71-61-2 (53.7%, +4 units)
Week 1 – SU: 9-7-0 ATS: 8-8-0 O/U: 13-3-0
Week 2 – SU: 10-6-0 ATS: 10-6-0 O/U: 10-6-0
Week 3 – SU: 12-4-0 ATS: 9-6-1 O/U: 8-8-0
Week 4 – SU: 7-6-0 ATS: 5-7-1 O/U: 5-8-0
Week 5 – SU: 14-2-0 ATS: 6-9-0 O/U: 9-6-0
Week 6 – SU: 11-3-1 ATS: 8-7-0 O/U: 6-9-1
Week 7 – SU: 11-4-0 ATS: 7-8-0 O/U: 8-7-0
Week 8 – SU: 11-3-0 ATS: 8-7-0 O/U: 8-7-0
Week 9 – SU: 9-4-0 ATS: 8-5-0 O/U: 4-8-1
New Orleans at Carolina
Prediction: Saints 25-24 (50.3%)
Pick: Panthers +3 (58.2%)
Total: Over 49
Jacksonville at Cincinnati
Prediction: Bengals 26-16 (76.2%)
Pick: Jaguars +11.5 (54.3%)
Total: Under 43.5
Tampa Bay at Cleveland
Prediction: Browns 23-19 (58.9%)
Pick: Buccaneers +7 (60.8%)
Total: Under 44
Arizona at Dallas
Prediction: Cowboys 25-21 (61.2%)
Pick: Cardinals +4.5 (51.5%)
Total: Over 45 PUSH
Philadelphia at Houston
Prediction: Texans 24-22 (55.8%)
Pick: Texans +2.5 (62.7%)
Total: Under 48.5
NY Jets at Kansas City
Prediction: Chiefs 21-17 (62.3%)
Pick: Jets +10 (65.6%)
Total: Under 42
San Diego at Miami
Prediction: Dolphins 23-21 (53.7%)
Pick: Dolphins -1 (50.8%)
Total: Under 44.5
Washington at Minnesota
Prediction: Vikings 22-21 (53.7%)
Pick: Washington Football Team +2.5 (53.5%)
Total: Under 43.5
Denver at New England
Prediction: Patriots 28-27 (53.9%)
Pick: Patriots +3.5 (63.6%)
Total: Under 56
Baltimore at Pittsburgh
Prediction: Steelers 23-21 (55.1%)
Pick: Steelers -1 (52.3%)
Total: Under 48.5
St. Louis at San Francisco
Prediction: 49ers 25-17 (71.6%)
Pick: Rams +10 (55.7%)
Total: Under 44
Oakland at Seattle
Prediction: Seahawks 26-15 (78.2%)
Pick: Raiders +15 (61.5%)
Total: Under 43.5
Indianapolis at NY Giants
Prediction: Giants 24-23 (54.0%)
Pick: Giants +3.5 (63.7%)
Total: Under 52
NCAA Football Rankings Top 25 – 10/28/2014
|
Chicago, Statistics, and Networking: Part 1
I’ve been in Chicago for about 2 months now. And I f****** love it. (We’ll see how I feel in January/February, though). I’ve never lived in a big city (or near a big city, as I actually live outside of the city) before in my life, and I’m constantly amazed by normal city things. Public transportation, coffee houses, walking and biking places, never using my car (I’ve gone to the gas station exactly twice since I have been here). And I love all of it.
On the statistics front, there are so many people doing statistics here it’s hard for me to comprehend. As a way to meet some more of those people, I attended the Chicago Chapter of the ASA statistician of the year award (which has a pretty impressive list of winners) dinner on October 22nd. At dinner, I sat with two of my students from my non-parametric class and met a statistician at Nielsen and another one who worked for Mathematica Policy Research. Needless to say that led to a very interesting dinner conversation.
This years winner was Andrew Gelman, who spoke for maybe 20-30 minutes about his thoughts on statistics.
My two take aways from his talk were:
- Gelman mentioned that he had a conversation with Don Rubin year ago about deciding whether to pursue physics or statistics and Rubin told him that in statistics you get to work on whatever you want. The rest is history and statistics has Gelman and his wide array of work in many different areas.
- I hope I can capture the essence of what he was saying here. Gelman mentioned that there was this sort of mentality of us versus them (statisticians versus scientists), in that statisticians are always trying to train scientists in proper statistics. But Gelman argued that we’ve trained scientists TOO well. To the point where scientists believe that once they show that the p-value is below 0.05 that they have discovered some object truth about the universe. (Statisticians know better, of course.) And that this has led to many borderline crazy studies. Gelman mentioned a few including a study that attempted to demonstrate that single women who were on their menstrual cycles were more like to vote for Obama, but married women on their cycle were more likely to vote for Romney.
After dinner and Gelman’s talk I introduced myself to Gelman and told him that I really enjoyed his blog. I asked him how he was capable of posting something every day and he said that he just writes a bunch of posts and then schedules them for future dates and that he is about 2 months (!) ahead on blog posts. He also mentioned that blogging is what he does when he is avoiding work. Most people waste time checking emails, apparently he blogs. He also mentioned that he saw it as service to the field.
And it was this short conversation with him that has led me to spending the day in a coffee shop writing blog posts when I should be checking my email and revising manuscripts. But I’ll get to that later. I love Chicago.
Cheers.
Stats in the Wild 