I didn't get a chance last year to do my N.F.L. rankings, but here they are far this year. I made a slight tweak and used a logistic function to rate each win. A small point differential resulted in a half-win to each time. A large point differential resulted in one team getting much closer to a win (nearly 1) and the other getting a loss (closer to 0). However, with a logistic function, the differential does not result in a linear change. Basically, I set it up so that anything beyond a two-possession game -- for example, one team wins by 16 points -- is above a 0.90 for the winning team. Running up the score further pushes the win closer to a 1 in an ever slower fashion.
The next step was to use these adjusted results to do a strength-of-schedule adjustment.
Here is my logistic function:
And here are my rankings:
The link to download it. No real surprises, except that the Ravens are ranked in my method ahead of the Bengals, Colts, and Steelers.
My wild-card round predictions: Ravens over Steelers (the only "surprise"), Colts over Bengals, Cowboys over Lions, Cardinals over Panthers.
---
Update, 1/5/2015: Got the Ravens, Colts, and Cowboys right. Got the Panthers wrong (although this should have been the one I was the most confident about). Score: 3/4.
Next predictions for the divisional playoffs, using the unrevised list above: Patriots over Ravens, Broncos over Colts, Packers over Cowboys, Seahawks over Panthers.
Update, 1/13/2015: Got the Patriots, Packers, and Seahawks right. Got the Colts wrong. Score: 3/4.
Next predictions: Seahawks over Packers, Patriots over Colts.
Update: 1/21/2015: Got both right. Score: 2/2.
Next predictions: Seahawks over Patriots in the Superbowl.
Update: Wrong Superbowl prediction. 0/1.
Total score: 8/11, or 73%.
Showing posts with label NFL. Show all posts
Showing posts with label NFL. Show all posts
Monday, December 29, 2014
Thursday, January 3, 2013
Literature base
I'm going to enter into cantankerous old man territory by making this post: the National Forensic League is not doing a great job with Lincoln-Douglas topics, but it is doing a decent job with Public Forum topics. My problem is the scope of the topic literature bases. It is unreasonable (and counter-productive to education) to give students a big topic to research and little time to do it. Students learn best when they can thoroughly explore a topic and are able to prepare on most of the key arguments. Of course, because debate is a competitive activity, some opponents will always search out unusual, squirrelly arguments, but my point is that being caught off-guard should be a rare experience for debaters. If the topic is too broad to prepare and debaters are regularly caught off-guard, then the value of research and preparation is undermined.
Given that topics are used for one month in PF, two months in LD, and ten months in Policy debate, I think the literature bases should follow a 1:2:10 proportion for these various formats. In other words, the length of time a topic is used should be roughly commensurate with how big the literature base is. This only makes sense: LD debaters might prepare for four to six topics during the year, so it only makes sense that they would invest about one-fifth the research and preparation time into each one as policy debater do; PF debaters might prepare for nine to eleven topics, so one-tenth the work seems fair. How would the N.F.L. do against this benchmark?
I started with the last five years' policy topics. The N.F.L. does NOT write the policy debate topics; those are written by the National Federation of High School Associations. Generally, the policy topics are just about the right breadth to spend a whole year researching. This result is probably no accident, because their process requires the topic framers to consider the literature base qualitatively, and to a lesser extent, quantitatively. To make a very rough gauge of the size of the literature bases, I typed in key terms and terms of art into ProQuest. ProQuest is a general use research database, containing both news articles and some academic journal articles, which is commonly available to high school students. I chose search terms that seemed appropriate to find the core articles of each topic. I make no apologies for the very provisional method; don't infer too much from this. I think this gives a sense of scale for comparison, but not much else.
Policy topics:
2012-2013 topic: "The United States federal government should substantially increase its transportation infrastructure investment in the United States."
Search terms: "transportation infrastructure" and ("United States" or "U.S.")
Results: 16,069
2011-2012 topic: "The United States federal government should substantially increase its exploration and/or development of space beyond the Earth’s mesosphere."
Search terms: ("space exploration" or "development of space") and ("United States" or "U.S.")
Results: 25,657
2010-2011 topic: "The United States federal government should substantially reduce its military and/or police presence in one or more of the following: South Korea, Japan, Afghanistan, Kuwait, Iraq, Turkey."
Search terms: ("military presence" or "police presence") and ("United States" or "U.S.") and ("South Korea" or Japan or Afghanistan or Kuwait or Iraq or Turkey)
Results: 17,954
2009-2010 topic: "The United States federal government should substantially increase social services for persons living in poverty in the United States."
Search terms: "social services" and poverty and ("United States" or "U.S.")
Results: 22,133
2008-2009 topic: "The United States federal government should substantially increase alternative energy incentives in the United States."
Search terms: "alternative energy" and ("United States" or "U.S.")
Results: 36,728
There is a wide variation from topic to topic. But the mean of 24,000 articles seems about right and a reasonable place to start from.
How do PF topics compare? They are in the right ballpark.
2012-2013 PF topics:
Sept. topic: "Congress should renew the Federal Assault Weapons Ban."
Search terms: "assault weapons ban"
Results: 3,367
Oct. topic: "Developed countries have a moral obligation to mitigate the effects of climate change."
Search terms: "developed countries" and ("mitigate" or "mitigation") and "climate change"
Results: 2,546
Nov. topic: "Current U.S. foreign policy in the Middle East undermines our national security."
Search terms: ("U.S. foreign policy" or "United States foreign policy") and "Middle East" and "national security"
Results: 3,105
Dec. topic: "The United States should prioritize tax increases over spending cuts."
Search terms: ("tax increase" or "spending cut") and "fiscal cliff" and ("United States" or "U.S.")
Results: 1,200 (66,279 without "fiscal cliff" term)
Jan. topic: "On balance, the Supreme Court decision in Citizens United v. Federal Election Commission harms the election process."
Search terms: "Citizens United" and election
Results: 3,885
Feb. topic: "On balance, the rise of China is beneficial to the interests of the United States."
Search terms: China and ("United States interests" or "U.S. interests")
Results: 5,756
Search terms: "rise of China" and ("United States" or "U.S.")
Results: 3,350
The China topic seems to be dangerously large, but the remaining seven topics have an average of 2,800, just about one-tenth of a policy topic. How do LD topics compare?
2012-2013 LD topics:
Sept./Oct. topic: "The United States ought to extend to non-citizens accused of terrorism the same constitutional due process protections it grants to citizens."
Search terms: terrorist and "due process" and ("United States" or "U.S.")
Results: 5,505
Nov./Dec. topic: "United States ought to guarantee universal health care for its citizens."
Search terms: "universal health care" and ("United States" or "U.S.")
Results: 10,696
Jan./Feb. topic: "Rehabilitation ought to be valued above retribution in the United States criminal justice system."
Search terms: (rehabilitation or retribution) and "criminal justice" and ("United States" or "U.S.")
Results: 10,546
With the exception of the first topic, they are all quite large topics. The health care topic is essentially the 1993-1994 policy debate topic! The criminal justice topic is similar to the Jan./Feb. 2011 LD topic, except that one was limited by its focus on juveniles.
What specific recommendations would I make to the N.F.L.?
First, when writing topics, please consider the size and quality of the literature base. Perhaps develop some standard statistics to gauge the size of the literature base. If the potential topic generates too many hits, narrow the topic in some way; if the potential topic generates too few hits, broaden it.
Second, please report out the statistics to us members whenever we vote between different wordings of the same topic. It would be helpful to know which version is the more narrowly worded.
Given that topics are used for one month in PF, two months in LD, and ten months in Policy debate, I think the literature bases should follow a 1:2:10 proportion for these various formats. In other words, the length of time a topic is used should be roughly commensurate with how big the literature base is. This only makes sense: LD debaters might prepare for four to six topics during the year, so it only makes sense that they would invest about one-fifth the research and preparation time into each one as policy debater do; PF debaters might prepare for nine to eleven topics, so one-tenth the work seems fair. How would the N.F.L. do against this benchmark?
I started with the last five years' policy topics. The N.F.L. does NOT write the policy debate topics; those are written by the National Federation of High School Associations. Generally, the policy topics are just about the right breadth to spend a whole year researching. This result is probably no accident, because their process requires the topic framers to consider the literature base qualitatively, and to a lesser extent, quantitatively. To make a very rough gauge of the size of the literature bases, I typed in key terms and terms of art into ProQuest. ProQuest is a general use research database, containing both news articles and some academic journal articles, which is commonly available to high school students. I chose search terms that seemed appropriate to find the core articles of each topic. I make no apologies for the very provisional method; don't infer too much from this. I think this gives a sense of scale for comparison, but not much else.
Policy topics:
2012-2013 topic: "The United States federal government should substantially increase its transportation infrastructure investment in the United States."
Search terms: "transportation infrastructure" and ("United States" or "U.S.")
Results: 16,069
2011-2012 topic: "The United States federal government should substantially increase its exploration and/or development of space beyond the Earth’s mesosphere."
Search terms: ("space exploration" or "development of space") and ("United States" or "U.S.")
Results: 25,657
2010-2011 topic: "The United States federal government should substantially reduce its military and/or police presence in one or more of the following: South Korea, Japan, Afghanistan, Kuwait, Iraq, Turkey."
Search terms: ("military presence" or "police presence") and ("United States" or "U.S.") and ("South Korea" or Japan or Afghanistan or Kuwait or Iraq or Turkey)
Results: 17,954
2009-2010 topic: "The United States federal government should substantially increase social services for persons living in poverty in the United States."
Search terms: "social services" and poverty and ("United States" or "U.S.")
Results: 22,133
2008-2009 topic: "The United States federal government should substantially increase alternative energy incentives in the United States."
Search terms: "alternative energy" and ("United States" or "U.S.")
Results: 36,728
There is a wide variation from topic to topic. But the mean of 24,000 articles seems about right and a reasonable place to start from.
How do PF topics compare? They are in the right ballpark.
2012-2013 PF topics:
Sept. topic: "Congress should renew the Federal Assault Weapons Ban."
Search terms: "assault weapons ban"
Results: 3,367
Oct. topic: "Developed countries have a moral obligation to mitigate the effects of climate change."
Search terms: "developed countries" and ("mitigate" or "mitigation") and "climate change"
Results: 2,546
Nov. topic: "Current U.S. foreign policy in the Middle East undermines our national security."
Search terms: ("U.S. foreign policy" or "United States foreign policy") and "Middle East" and "national security"
Results: 3,105
Dec. topic: "The United States should prioritize tax increases over spending cuts."
Search terms: ("tax increase" or "spending cut") and "fiscal cliff" and ("United States" or "U.S.")
Results: 1,200 (66,279 without "fiscal cliff" term)
Jan. topic: "On balance, the Supreme Court decision in Citizens United v. Federal Election Commission harms the election process."
Search terms: "Citizens United" and election
Results: 3,885
Feb. topic: "On balance, the rise of China is beneficial to the interests of the United States."
Search terms: China and ("United States interests" or "U.S. interests")
Results: 5,756
Search terms: "rise of China" and ("United States" or "U.S.")
Results: 3,350
The China topic seems to be dangerously large, but the remaining seven topics have an average of 2,800, just about one-tenth of a policy topic. How do LD topics compare?
2012-2013 LD topics:
Sept./Oct. topic: "The United States ought to extend to non-citizens accused of terrorism the same constitutional due process protections it grants to citizens."
Search terms: terrorist and "due process" and ("United States" or "U.S.")
Results: 5,505
Nov./Dec. topic: "United States ought to guarantee universal health care for its citizens."
Search terms: "universal health care" and ("United States" or "U.S.")
Results: 10,696
Jan./Feb. topic: "Rehabilitation ought to be valued above retribution in the United States criminal justice system."
Search terms: (rehabilitation or retribution) and "criminal justice" and ("United States" or "U.S.")
Results: 10,546
With the exception of the first topic, they are all quite large topics. The health care topic is essentially the 1993-1994 policy debate topic! The criminal justice topic is similar to the Jan./Feb. 2011 LD topic, except that one was limited by its focus on juveniles.
What specific recommendations would I make to the N.F.L.?
First, when writing topics, please consider the size and quality of the literature base. Perhaps develop some standard statistics to gauge the size of the literature base. If the potential topic generates too many hits, narrow the topic in some way; if the potential topic generates too few hits, broaden it.
Second, please report out the statistics to us members whenever we vote between different wordings of the same topic. It would be helpful to know which version is the more narrowly worded.
Tuesday, August 9, 2011
A modest proposal to ensure geographic mixing at Nationals
I wrote an article (visible below), published in the inaugural edition of the National Journal of Speech and Debate, about using geographic and strength criteria to mix teams in preliminary rounds at N.F.L. Nationals. In other words, I advocate the use of a system that ensures that each team will debate a broad cross-section of different opponents, from different parts of the country and at different skill/experience levels, as measured by N.F.L. debate points. I won't repeat the arguments here about why I think this is a worthwhile goal, except for this one thought: Geographic mixing makes it likely that the less experienced teams -- who probably have not travelled far afield -- will debate opponents at Nationals they have never seen before. Even with geographic mixing, there is still a chance that national circuit teams might face an opponent in preliminary rounds at Nationals they have debated many times during the invitational season. That is why there is also a need for skill/experience level mixing. Both are necessary to make it likely every team will see "new" opponents at Nationals.
I will focus on two technical concerns about my proposal in this blog post.
Concern 1: Can two criteria really be maximized at the same time?
Yes and no. In a strict sense, no: only one variable can truly be maximized at a time. That is to say, you can have a round where the average geographic distance between opponents is maximized, or you can have a round where the average difference of skill/experience between opponents is maximized, but you can not have both at the same time. However, in a looser, more practical sense, the answer is yes: you can have a round where opponents are well-mixed geographically (even though not maximally mixed) AND well-mixed skill/experience-wise (even though not maximally mixed). Let me show you with some sample data.
Here are 26 fictitious teams, spread throughout the country in geographic clusters, and at different skill/experience levels (normally distributed from 0 to 1 in my sample data). I imagined that the N.F.L. points could be scaled so the weakest team to qualify was given a rating of 0 and the most experienced a 1, but they do not need to be scaled at all for this method to work. The median distance between every possible pairing in the whole set is 2138 miles. The median difference in experience is 0.28 units.
If one tries to maximize geographic spread, the round 1 pairings that are selected would look like this:
The median distance between two opponents in each pairing is 3333 miles, and the shortest distance is 2452 miles. In other words, every match chosen is above the average of 2138 miles. This is maximized; it is a Pareto optimal solution, meaning that any change to improve a pairing by swapping opponents would have to make another pairing worse. The net result can not be improved. In the first round, teams in the middle of the country would debate coastal teams. As the tournament proceeds, each team would get opponents from every geographic region of the country.
If one tries to maximize the differences of skill/experience levels, the round 1 pairing would look like this:
The median difference between opponents is 0.45 units, and the least difference is 0.42 units. Every match chosen is above the average of 0.28 units. Again, this is a Pareto optimal solution. In the first round, mid-level teams would debate either inexperienced or highly experienced teams. No inexperienced team would be matched against a highly experienced team -- this would force two mid-level teams to debate. However, in further rounds, each team would get opponents at every different level.
What happens if you try to maximize both? The resulting pairings would look like this:
The median distance is 3092 miles, and the shortest distance is 2003 miles (with only 15% of matches below the average of 2138 miles). The median difference is 0.45 units, and the least difference is 0.23 units (with 23% of matches below the average of 0.28 units). Although these pairings do not maximally mix for geography, they do pretty well. And likewise for difference in skill/experience level. This represents a lower boundary of how well this method could work. If we use a larger data set, such as the 200+ teams at Nationals, then it becomes easier to find pairings that maximize both criteria.
Concern 2: Would this method create the same pairing year after year?
It seems like it might: an optimal solution for one year seems like it might be the same, or very similar, the next year. No one wants to see the same opponent in preliminary rounds two (or more) years in a row at Nationals.
However, this kind of optimization is chaotic, meaning it is extremely sensitive to small changes.
I will focus on two technical concerns about my proposal in this blog post.
Concern 1: Can two criteria really be maximized at the same time?
Yes and no. In a strict sense, no: only one variable can truly be maximized at a time. That is to say, you can have a round where the average geographic distance between opponents is maximized, or you can have a round where the average difference of skill/experience between opponents is maximized, but you can not have both at the same time. However, in a looser, more practical sense, the answer is yes: you can have a round where opponents are well-mixed geographically (even though not maximally mixed) AND well-mixed skill/experience-wise (even though not maximally mixed). Let me show you with some sample data.
Here are 26 fictitious teams, spread throughout the country in geographic clusters, and at different skill/experience levels (normally distributed from 0 to 1 in my sample data). I imagined that the N.F.L. points could be scaled so the weakest team to qualify was given a rating of 0 and the most experienced a 1, but they do not need to be scaled at all for this method to work. The median distance between every possible pairing in the whole set is 2138 miles. The median difference in experience is 0.28 units.
If one tries to maximize geographic spread, the round 1 pairings that are selected would look like this:
The median distance between two opponents in each pairing is 3333 miles, and the shortest distance is 2452 miles. In other words, every match chosen is above the average of 2138 miles. This is maximized; it is a Pareto optimal solution, meaning that any change to improve a pairing by swapping opponents would have to make another pairing worse. The net result can not be improved. In the first round, teams in the middle of the country would debate coastal teams. As the tournament proceeds, each team would get opponents from every geographic region of the country.
If one tries to maximize the differences of skill/experience levels, the round 1 pairing would look like this:
The median difference between opponents is 0.45 units, and the least difference is 0.42 units. Every match chosen is above the average of 0.28 units. Again, this is a Pareto optimal solution. In the first round, mid-level teams would debate either inexperienced or highly experienced teams. No inexperienced team would be matched against a highly experienced team -- this would force two mid-level teams to debate. However, in further rounds, each team would get opponents at every different level.
What happens if you try to maximize both? The resulting pairings would look like this:
The median distance is 3092 miles, and the shortest distance is 2003 miles (with only 15% of matches below the average of 2138 miles). The median difference is 0.45 units, and the least difference is 0.23 units (with 23% of matches below the average of 0.28 units). Although these pairings do not maximally mix for geography, they do pretty well. And likewise for difference in skill/experience level. This represents a lower boundary of how well this method could work. If we use a larger data set, such as the 200+ teams at Nationals, then it becomes easier to find pairings that maximize both criteria.
Concern 2: Would this method create the same pairing year after year?
It seems like it might: an optimal solution for one year seems like it might be the same, or very similar, the next year. No one wants to see the same opponent in preliminary rounds two (or more) years in a row at Nationals.
However, this kind of optimization is chaotic, meaning it is extremely sensitive to small changes.
Subscribe to:
Posts (Atom)