| |
Saturday, December 18, 2004
How the Power Ratings Are Derived
I analyze the score of every game and adjust the actual points scored to reflect the quality of the opponent, and the location of the game (home or away). For example, consider New England's first game, a 27-24 victory over Indianapolis. Indianapolis' defense gives up almost exactly the average number of points that other NFL teams have, 0.994 times as many to be precise. So we can say that scoring 27 points against them is pretty much like scoring 27.2 points against an average team using the formula (27/.994). However, because the game was played at New England, the spreadsheet is not as impressed with New England's performance as it might otherwise be. Home Field Advantage in the NFL amounts to about three points. We allocate half of that to offense and half to defense, so the net adjustment to the offense's score is 1.5 points. Thus, New England's scoring 27 points at home against Indianapolis is considered the equivalent of them scoring 25.7 points against an average team at a neutral site.
For the Patriot's defense, the spreadsheet looks at Indianapolis's scoring per game as compared to the league average. The Colts have scored 1.629 times as many points as an average NFL team, so we could say giving up 24 points to them is like giving up 14.7 points to an average NFL team (24/1.629=14.7). However, the game was at New England, so I add 1.5 points to the opposition's score. Thus, giving up 24 points to the Colts at home is considered the equivalent of giving up 16.2 points to an average team on a neutral field.
After every game has been scored like this, New England comes up with an adjusted total of 374 points scored and 208 points allowed, or 166 points above average for the season. Since there have been 13 games played as of last weekend, the Patriots are about 12.8 points per game better than an average team. To avoid having to deal with negative numbers for bad teams, the Power Rating is added to 100; New England's power rating is thus 112.8.
A team's adjusted points scored and adjusted points allowed are generally fairly close to their actual points scored and points allowed this late in the season. But for teams facing a weak or strong schedule the numbers can vary substantially. Looking at Cleveland, for example, they have given up 98 more points than they've allowed. But they've faced a very difficult schedule, with two games each against Pittsburgh and Baltimore and with Phillly, New England, Buffalo and the Jets among their non-divisional opponents. Their opponents' average Power Rating has been 103.8. So the spreadsheet sees their numbers as a little better than they appear, for a net negative 48 points instead of negative 98 points. Seattle, which has faced the easiest schedule in the league, is adjusted downwards from its positive 16 points to negative 26.
Friday, December 17, 2004
Meet Rafael Peralta, American Hero
Ollie North has the details.
Not only can Rafael's family be proud of him, but his fellow Marines are alive because of him. As Peralta lay near death on the floor of a Fallujah terrorist hideout, he spotted the yellow grenade that had rolled next to his near-lifeless body. Once detonated, it would take out the rest of Peralta's squad. To save his fellow Marines, Peralta reached out, grabbed the grenade and tucked it under his abdomen, where it exploded.
Reggie Rivers: Merry Christmas is Divisive
The former Denver Bronco and current columnist weighs in on the Merry Christmas sign with a novel take.
If the removal of "Merry Christmas" can provoke this type of reaction, then no one can reasonably argue that it's a benign phrase that shouldn't offend anyone.
Get it? It's a classic Catch-22; if you object to the removal of the sign it shows that the sign is divisive and should be removed. And if you don't object to the removal of the sign, it's not divisive, but we're removing it because you didn't object.
Let me pose an alternative scenario. Suppose the City of Denver, in a moment of civic pride, put up a sign that said "Go Broncos!" Some local sports fans object that they are fans of the Raiders. The city decides to take down the sign, causing a brouhaha. Would this prove that the sign should come down, because after all, nobody'd raise a fuss if it weren't so divisive?
Thursday, December 16, 2004
Better NFL Ratings
I've been a little tough on the guys over at Football Outsiders for some questionable comments they've made. But it's time to acknowledge that they are doing some very good work over there.
A big part of the Football Outsiders website are the DVOA ratings derived by Aaron from an analysis of every single play of every single game. When I looked at them last year, the ratings did not seem very good. For example, when I looked at how he'd rated the teams as of Week 6 and how those teams had done since, I found that his top 16 teams combined for a record of 84-97 after that week. By contrast, my top 16 teams as of that week combined for a record of 102-82 after Week 6. It seemed to me that if you were going to claim you had a good rating system, it should at least be able to separate good teams from bad teams. (Correction: All the records shown are after Week 5 of last year's NFL season, not Week 6).
Aaron and I debated this point on his website last year at some length. He stated that the purpose of his ratings was not to predict the future but to analyze the past, but of course that is begging the question: Have you really analyzed the past well when your ratings don't have any relevance to the future?
I emphasized at the time that I thought Aaron's ratings would eventually be better than mine. All I do is enter the scores of the games into a spreadsheet, then translate those scores into an assumed score against an average opponent on a neutral field. My thought was that there might be something in the scores that was not always reflected in the Won/Lost records, and indeed, I have found that is true. So it seemed logical to me that Aaron's analysis of play-by-play data would find things that were not always reflected in the score of the game.
Well, I am pleased to say that either Aaron has made some improvements to his DVOA ratings, or he has gotten very lucky indeed, because his DVOA ratings this year are better than my Power Ratings, and by a significant enough margin that I can recommend them as the best way I've seen to rate the teams.
For starters, let's establish a baseline. The simplest method of rating the teams is by won/lost record. Teams are of course rated this way every day in the standings published in the newspaper. So if you wanted to rate teams, the simplest thing would be to say that New England, Pittsburgh and Philadelphia are the best teams in the league, and that San Francisco and Miami are the worst. And you wouldn't be far wrong, obviously.
So it seems to me that any method of rating the teams should at least do as well as simply using the won/lost record. The way of checking how good the rating is involves a statistical method known as correlation. Correlation is just a way of checking how closely two sets of numbers fit each other. For example, if you had one set of the numbers 1,2,3,4, and another set of the numbers 2,3,4,5, it should be obvious that there is a high degree correlation between the two sets. In fact, the correlation would be 100%, because the second set can be produced by adding 1 to each of the numbers in the first set.
What I did was look at the correlation between the won/lost percentage of each team as of Week 3 and their Won/Lost percentage in the games since then. For example, Philadelphia had one of the best Won/Lost Percentages (1.000) in the league at 3-0 after the third week, and they have gone 9-1 (.900) since then, for one of the best Won/Lost Percentages after Week 3. Obviously there is a great deal of correlation in looking just at Philly. But Seattle was also 3-0 (1.000 and they have gone 4-6 (.400 since) so the correlation isn't always high. Looking at all the records as of Week 3, the correlation between a team's record then and their record since is 33%. I then did the same calculation for the next four weeks as well:
Week 3 : 33%
Week 4 : 33%
Week 5 : 32%
Week 6 : 27%
Week 7 : 32%
Pretty consistent there. So that's the benchmark. How does my system do?
Week 3 : 40%
Week 4 : 41%
Week 5 : 33%
Week 6 : 26%
Week 7 : 33%
Better, except for that Week 6 blip. Must have been some funky games. Now, there are two opposing tendencies in the ratings. Ratings in later weeks should be more accurate than earlier ratings because they are based on more games. However, there have also been fewer games since then, so there is more room for random results to throw off the correlation.
But when I checked Aaron's DVOA ratings, the correlations were higher:
Week 3 : 47%
Week 4 : 46%
Week 5 : 43%
Week 6 : 45%
Week 7 : 47%
That's not to say my Power Ratings are useless. Indeed, because they are denominated in points, they have more immediate utility to people looking to place a bet on a game or fill in the office pool. But I gave Aaron a hard time about the correlation between his ratings and future won/lost percentages a year ago; it seems only fair to note now that he has improved his record markedly.
Wouldn't Bother Me in the Slightest
Denver Post columnist Cindy Rodriguez comes up with what she no doubt thinks is a devastating argument against "Merry Christmas" signs on government property:
Some Christians don't think twice about complaining because they are the majority and feel entitled. But imagine if Jews were the majority and the rest of us had to pass a "Happy Hanukkah" sign on the way into the City and County Building.
Horrors! I don't think I could handle it!
Seriously, this strikes me as similar to John Kerry's comment about Mary Cheney being a lesbian. The similarity is that both Kerry and Rodriguez seem to think that the Christian right is made up of a bunch of bigots, so why not use their bigotry against them? Kerry calculated that the Christian right hated homosexuals, so his mention of Mary Cheney's sexuality might cost the Bush/Cheney ticket some votes. Rodriguez calculates here that the Christian right is made up of religious bigots, so why not use that religious bigotry to good purpose (as she sees it) by confronting them with the notion of governmental Happy Hanukkah signs.
What's next, arguments against Social Security privatization because it will prevent the government from taking money from black people (who have shorter lifespans than average) and giving it to white people?
A Recommendation the Democrats Won't Take
Peggy Noonan suggests that the Democrats start taking on the anti-religious zealots.
It is this: Stop the war on religious expression in America. Have Terry McAuliffe come forward and announce that the Democratic Party knows that a small group of radicals continue to try to "scrub" such holidays as Christmas from the public square. They do this while citing the Constitution, but the Constitution does not say it is wrong or impolite to say "Merry Christmas" or illegal to have a crèche in the public square. The Constitution says we have freedom of religion, not from religion. Have Terry McAuliffe announce that from here on in the Democratic Party is on the side of those who want religion in the public square, and the Ten Commandments on the courthouse wall for that matter. Then he should put up a big sign that says "Merry Christmas" on the sidewalk in front of the Democratic National Committee Headquarters on South Capitol Street. The Democratic Party should put itself on the side of Christmas, and Hanukkah, and the fact of transcendent faith.
The good news is that there is not a chance that the "reality-based community" will accept this advice.
Read the whole article; I especially loved the bit about manipulating symbols. The Democrats are constantly searching for the secret code words that will get them elected despite being anti-relious and anti-American.
Update: Jim Geraghty has the same take over at the Kerry Spot, but he phrases it better:
The Democratic party doesn’t have a reputation as the side hostile to religious faith because its messages have been misinterpreted, or because it has clumsy public relations, or because it can‘t keep up with the GOP “message machine.” It has that reputation because a large number of its members are hostile to religious faith.
Tuesday, December 14, 2004
Fourth And One
Gregg Easterbrook made a comment in his Tuesday Morning Quarterback column last week in favor of teams going for it on fourth down and a yard when inside enemy territory. I took a quick look at it, and it certainly appears that he's right. I took a look at the 32 times teams had gone for the first down in that situation between 10/31/04 and 12/5/04. I ignored cases where there was a penalty that negated the play or where the team went for it while lined up to kick either a punt or a field goal, so that we are concentrating only on the times where it was apparent that the team was going for it.
The results? Teams were successful in getting the first down (or touchdown)) on 23 of 32 attempts, or about 72% of the time. There was little apparent difference between running and passing, with runs succeeding 16 of 22 times (73%) and passes getting the job done 7 of ten times (70%). Just for the heck of it, I took a look at this weekend's games and the pattern was the same; there were four attempts (all runs) and three of them were successful.
Chomsky's Bagels
Benjamin Kerstein's article in Front Page Mag points to a Noam Chomsky recap of the election that reveals just how out of it the cunning linguist is.
The outcome was a disappointment, but there have been disappointments before. Take 1984, when essentially the same gang of thugs—a little less tilted to the extreme reactionary statist side—won by a 2-1 margin, with about the same percentage of the electoral vote as today. And they were engaged in horrendous atrocities abroad and very harsh and destructive programs for most of the population at home. The world didn’t come to an end. In fact, activism proved quite effective.
Hmmm, first of all, Reagan did not win by a 2-1 margin; he won by about 1.45-1. And Reagan's electoral vote win was far larger. Ronaldus Maximus got almost 98% of the electoral vote as compared to President Bush's 53%. I assume that the distinguished professor meant to say about the same turnout percentage as today, which would be reasonably true (56.1% of the voting age population turned out this year, as compared to 53.1% in 1984).
Later, Chomsky claims:
The progressive left is very substantial in scale, and could be far larger, including the large majority of the population, judging by highly credible public opinion studies that the press scarcely mentions, presumably because they understand that it is much too dangerous to allow people to understand that they are not alone in their views.
This is, of course, the "lost tribe of leftists" theory that is so popular in the "reality-based community". In this theory, there exists a huge, untapped base of support for far left positions that fails to show up at the polls every four years because no presidential candidate, even the Democratic nominee, is really speaking to them. If only the Democrats would nominate a true leftist candidate rather than these semi-Republicans like Al Gore and John Kerry, they would sweep to an easy victory. Chomsky makes this clear:
I don’t think that the Kerry campaign even tried to include the opinions of most of the population, including those who voted for Kerry. People will vote their class interests when they see some credible political force that might represent those interests. That’s not Kerry or the DLC.
Of course, Kerry is not affiliated with the DLC (Democratic Leadership Council). Indeed, one suspects that if the DLC had a candidate in the 2004 race, it was the much-maligned (among Democrats) Joe Lieberman.
NFL Power Ratings after Week 14
Comments: New England retains the top spot for the sixth consecutive week. San Francisco put its top draft slot in jeopardy with their win at Arizona, but the spreadsheet still sees them comfortably worse than their competitor for that draft pick. Miami is seven points better than the 49ers (which happens to be reflected in the 24-17 win they notched at the 'Stick)..
The ratings reflect the astounding domination of the AFC over the NFC, with 11 of the top twelve teams coming from the former conference.
NE 112.6
Ind 112.1
Phi 111.1
Pit 107.9
SD 107.0
Bal 106.4
NYJ 105.7
KC 104.3
Buf 103.3
Den 103.1
Jax 101.4
Cin 101.2
GB 101.2
TB 100.8
Atl 100.6
Car 99.1
Min 99.0
Cle 98.0
Was 97.9
Sea 97.8
Ten 96.8
Hou 96.7
Det 96.5
NYG 95.1
Chi 95.0
StL 95.0
Oak 94.9
Mia 94.9
Dal 93.3
AZ 92.1
NO 92.0
SF 87.2
Monday, December 13, 2004
Connecticut To Execute First Killer Since 1960?
According to this story.
Ross, 45, admits killing eight women in Connecticut and New York in the early 1980s, and raping most of his victims. He has been in prison for 20 years — 17 on Connecticut's death row — for four of those murders and faces lethal injection.
Remind Me Never To Do That Again
Shoot off my mouth without checking the facts, that is. I took a quick look at the combined records of all AFC teams every year since 1970 (the year the AFL and NFL began playing regular season games together), and found that the AFC was 16 games over .500 in 1999, and an incredible 20 games over .500 in 1979. Since the merger, all told, the AFC is 60 games over .500 against the NFC in regular season games (and 4 games--all Super Bowls of course--under .500 in postseason games). I'd guess you could win a bar bet or two with that knowledge, since most people assume the AFC has been the weaker conference. Even during the AFC's long dry spell from 1984-1996, when they lost 13 consecutive Super Bowls, they were only 15 games under .500 in games with the NFC.
Dominating the other conference in the regular season does seem to have some predictive value to the Super Bowl. There have been 34 seasons since the merger; in six of those years (1973, 1983, 1984, 1990, 2000 and 2001) the AFC and NFC were exactly at .500, so there are 28 remaining seasons/Super Bowls to check. The team from the conference with the winning record went 17-11 in those games. However, you do not have to search hard for contrary evidence; the 2002 Tampa Bay Bucs, the 1999 St. Louis Rams and the 1996 Green Bay Packers all won the big game despite coming from the weaker conference in the regular season.
How Bad Is the NFC?
Right now the second wild card would go to a team (Carolina) that started their season 1-7. There were four NFC/AFC matchups yesterday; the NFC went 1-3, to run their record for the season to 19-35. I'd have to check, but I'd suspect that's the most games under .500 that any conference has been since the early days of the NFL/AFL merger.
Sunday, December 12, 2004
NFL Notes
In honor of this week's marquee matchup, we present the years in which the 49ers and the Cardinals have both had winning records: 1998, 1984, 1983, 1976, 1970, 1968 and 1960. It is safe to say that they will not be adding another season to that total of seven in 54 years.
The real marquee matchup is the game between the NY Jets and the Pittsburgh Steelers. A simple way to rate the games is to multiply the number of wins the respective teams have. The Jets have won 9 games and the Steelers 11, so the game rates as a 99, easily the best of any contest today, with only the Patriots/Bengals rating within 50 points of that. The spreadsheet sees the Steelers winning by about 18-14, although of course it is unlikely to hit that score exactly, since very few teams actually score 18 points (only two teams have done so this season). The spreadsheet says the game fairly represents the expected matchup, with the assumed home field advantage for the Steelers at 3.0 points (the league average so far this season), although both the Steelers (4.6) and the Jets (3.9) have larger HFAs, which could indicate that Pittsburgh should be favored by a little more in my spreadsheet. The line looks to be just about on the money around 5, whice means I wouldn't touch this game.
I took a quick look at how many times X number of points had been scored this season.
0 2
1 0
2 0
3 10
4 0
5 0
6 10
7 11
8 1
9 5
10 26
11 0
12 5
13 20
14 19
15 4
16 13
17 39
18 2
19 8
20 21
21 17
22 6
23 11
24 22
25 3
26 5
27 23
28 15
29 2
30 7
31 19
32 2
33 1
34 17
35 5
36 0
37 2
38 5
39 1
40 2
41 5
42 3
43 2
44 0
45 3
46 0
47 1
48 2
49 3
56 1
58 1
Looks like I missed the score of one ballgame, as that adds up to 382, but there have been 384 team scores so far this season. As you can see, 17 is the most common score, with about 10% of all teams ending with that point total. Aside from that there are obvious clusters at 6 & 7 and then every 7 points above those--13 & 14, 20 & 21, 27 & 28. Using the scores shown above, we can estimate potential winning percentages for every score as follows:
0 0%
1 1%
2 1%
3 1%
4 3%
5 3%
6 3%
7 6%
8 9%
9 9%
10 10%
11 17%
12 17%
13 18%
14 24%
15 29%
16 30%
17 33%
18 43%
19 44%
20 46%
21 51%
22 56%
23 57%
24 60%
25 66%
26 67%
27 68%
28 74%
29 78%
30 79%
31 80%
32 85%
33 86%
34 86%
35 91%
36 92%
37 92%
38 92%
39 94%
40 94%
41 95%
42 96%
43 97%
44 97%
45 97%
46 98%
47 98%
48 98%
49 99%
50 99%
51 99%
52 99%
53 99%
54 99%
55 99%
56 99%
57 100%
58 100%
As you can see, the Cleveland Browns were unfortunate indeed to lose that game to Cincinnati 58-48. Or rather I should say, their offense was unfortunate, they had done enough to win 98% of all ballgames. Unfortunately their defense couldn't close the deal.
That gave me an idea for a quick rating system. Suppose we take the actual scores of a team's games and look at how the offense did as compared to the defense. Let's start with the Indianapolis Colts. In week one, they lost to the New England Patriots 24-27. Looking at the chart, we see that when scoring 24 points, teams should win about 60% of their games, and teams giving up 27 points should win about 32% of their games (to determine winning percentage for teams giving up X points, just take 100% minus the number shown for that score). Doing this for every game on the Colts schedule gives us an average winning percentage for the Colt's offense of 84%, as compared to their defense's winning percentage of 52%.
I'll take a longer look at this later.
|
|
