Three Super Bowls ago, I wrote this post over at Sabernomics. In it, I looked at your probability of winning a squares pool with any given square. For example, I found that in a one-unit-per-square pool, either of the '0/7' squares would have an expected value of about 3.8 units. Compare that with, say, a '5/6' square, which has an expected value of 0.22, or the lowly `2/2' square and its expected value of .04. Because it was all the data I had at the time, I only considered the last digits of the final scores of games, but someone correctly pointed out in the comments that most pools also give prizes for (the last digits of) the cumulative scores at the end of each quarter.

Well, now I have score-by-quarter data for the entirety of the NFL's 2-point-conversion era (1994--present), so it's time for an update.

(Continued)

Several new quarterbacks started for the first time in 2007. Derek Anderson, Kellen Clemens, Brodie Croyle, Tarvaris Jackson, Sage Rosenfels and Matt Schaub, along with rookie Trent Edwards, all threw 150 passes for the first time in their careers. Jay Cutler and Vince Young got experience their rookie seasons, and were the full-time starters in year two; Jason Campbell and David Garrard had some previous playing experience but also became full-time starters in 2007.

I am going to examine the new 2007 quarterbacks to see what history says about their futures. Right before the 2007 season, I wrote "In Search of the Next Brady or Bulger", which took a look at quarterbacks who were drafted outside the top 50 selections in the draft. One of the interesting things for me was that all of the late round successes showed signs of being successful right away--which may be contrary to some conventional wisdom about quarterback development. All were close to average (if not better) in their first real opportunity to start, and most were above average by their second season. But those were only the later round quarterbacks. Here, we will take a look at all the quarterbacks who have played since 1978, to find the most comparable seasons for each of these new quarterbacks.

(Continued)

This one cuts across all kinds of dimensions of trivial football knowledge:

What player/kicker pair has combined for the most career TD/PATs?

ANSWER UPDATE: Brown/Groza almost certainly is the correct answer, but I should have specified that I only have the data from 1960--2007. In that time period, it's a tie between Rice/Cofer and Riggins/Moseley

Answers from the questions in this post:

1. What player has caught touchdown passes from the most different players?

As was guessed by Ben, the answer is Irving Fryar. Here is the list of players that have thrown him a TD:

(Continued)

Here is some weekend fun with the Touchdown Project that I told you about earlier in the week.

I'll start with a few trivia questions (NOTE: all questions cover the time period 1970--2007):

1. What player has caught touchdown passes from the most different players? HINT: 2nd, 3rd, and 4th places are Ricky Proehl, my main man Joey Galloway, and Jerry Rice, respectively.

(Continued)

When a great pass rushing team -- like the 2007 Giants (New York led the NFL in sacks) -- plays a team with a great tight end -- like the 2007 Cowboys (Jason Witten led all TEs in fantasy points), does something unexpected happen? That is, assuming the teams are of equal strength, does one team win more than 50% of the time?

Here's the theory: when you face a great pass rushing team, you generally need to keep your tight end in to block more often than usual. When the Cowboys face the Giants and Michael Strahan and Osi Umenyiora (23 combined sacks this season), Jason Witten will be called on to help his tackles pass protect more often than when Dallas plays Buffalo (starting DEs had 9 sacks combined; the team ranked 30th with just 26 sacks on the season). So, perhaps when these teams play, the great pass rushing teams win a disproportionate number of times (relative to the differences in team strength of the two teams) because the team with a great tight end is forced to use one of its best weapons in a suboptimal way. This is just a theory, of course -- but it is testable.

How so? It's complicated, so let me break it down into steps.

(Continued)

The approximate value experiment (read about it here and here) is on hold for now because I'm busy with, well, lots of things. One of them will eventually be of interest to readers of this blog and users of this site.

Since the sports-reference merger, one grand idea of Sean Forman's has been to try to catalog every scoring play in NFL history and create from them a database that users (i.e. you) can search through. So I've been doing a little bit of work on that from time to time.

There's still much to be done, but I'm now to the point where I can spout some interesting facts. For example, 49% of Richard Todd's TD passes either tied the game, broke a tie, or gave his team the lead, while only 32% of Ken Anderson's did the same. Let's call such a touchdown a "crucial TD." Anderson and Todd might have played on very different kinds of teams, so it's unclear what that means, but check this out:

                 crucial TDs    total TDs      pct
===================================================
Mark Duper           36            59          61%
Mark Clayton         34            85          40%

This data will also allow me to run studies like: when a team overcomes a 14-or-more-point lead to tie the game, do they go on to win it more than 50% of the time? And, just in time, it will allow me to re-run this post with proper data.

Let's take a look at Marion Barber's historically comparable players at age 24. Here is the method I used:

1. Start with 1000 points;
2. Subtract 1 point for every difference of 1 rushing attempt (Barber had 204 in the regular season);
3. Subtract 20 points for every difference of 0.1 in yards per rushing attempt (Barber averaged 4.78 per attempt);
4. Subtract 2 points for every difference in receptions (Barber had 44)
5. Subtract 20 points for every difference of 1 touchdown (Barber had 12).

I then looked at all backs who had 150 or more rushing attempts at age 24, since 1970. Thirty-four different backs have a similarity score of 800 or better. I am actually going to provide the similar players broken down into three lists.

(Continued)

If you haven't done so yet, you'd better read Approximate Value I. (I know you're thinking, "yeah, yeah, whatever." I do the same thing. But I mean it. Go read the old post.)

I left off last time with this bunch of questions:

• What metric do I use to determine offensive points at the team level?
• What fraction of points should go to the line?
• What is the pass/run split?
• On the passing side, what is the throw/catch split?
• We need to figure a way to give some of those offensive line points to fullbacks and tight ends, many of whose jobs include a lot of blocking.

Hopefully, I'll be able to answer all these today, and run through an example or two. Before I do, I'd like to make a few comments about the method and about my style of doing these sorts of things:

(Continued)

Baseball analysis pioneer Bill James had a tool called the Value Approximation Method. In the 1982 Abstract, he introduced it thusly:

The value approximation method is a tool that is used to make judgements not about individual seasons, but about groups of seasons. The key word is approximation, as this is the one tool in our assortment which makes no attempt to measure anything precisely. The purpose of the value approximation method is to render things large and obvious in a mathemtatical statement, and thus capapble of being put to use so as to reach other conclusions.

[The emphasis was in the original.]

James then goes on to describe the method a bit. He used basic stats like batting average, RBI, stolen bases, pitching wins and losses, strikeouts, and so forth, to assign an integer to each player season. A typical MVP would be around 16 or 17, and all-star around 13, an average starter about 10, and so on. He continues:

(Continued)

The similarity score method, described here, sees a couple of uninteresting games:

                   Rec L6  Marg  H Bye  SIM   Result
====================================================
gnb                 3   1   122  1  0
====================================================
stl 2001  c  phi    3   1    95  1  0   973  W 29-24
phi 2004  c  atl    2   1   123  1  0   899  W 27-10
hou 1991  w  nyj    3   0   114  1  0   892  W 17-10
pit 2001  d  bal    3   1   102  1  1   880  W 27-10
den 2005  d  nwe    3   1    96  1  1   874  W 27-13
phi 2001  w  tam    2   1    91  1  0   869  W 31- 9
gnb 2002  w  atl    3   1   -18  1  0   860  L  7-27
chi 2006  c  nor    3   0    81  1  0   859  W 39-14
min 1992  w  was    2   1    80  1  0   858  L  7-24
dal 1994  d  gnb    3   1    71  1  1   849  W 35- 9
buf 1991  d  kan    3   1    70  1  1   848  W 37-14
pit 1995  c  ind    2   1    65  1  0   843  W 20-16
ind 2006  w  kan    3   0    51  1  0   829  W 23- 8
sea 2005  c  car    2   1    49  1  0   827  W 34-14
stl 1999  d  min    3   1   220  1  1   802  W 49-37
====================================================
WEIGHTED AVERAGE:  86.7 pct chance of victory
PROJECTED SCORE:  27.1-15.9

                   Rec L6  Marg  H Bye  SIM   Result
====================================================
nwe                 5   0   187  1  0
====================================================
hou 1991  w  nyj    3   0   114  1  0   727  W 17-10
kan 1995  d  ind    4   0   102  1  1   715  L  7-10
chi 2006  c  nor    3   0    81  1  0   694  W 39-14
chi 2006  d  sea    4   1   178  1  1   691  W 27-24
ind 2006  w  kan    3   0    51  1  0   664  W 23- 8
was 1991  d  atl    4  -1   238  1  1   649  W 24- 7
den 1998  d  mia    4   1   136  1  1   649  W 38- 3
sfo 1990  d  was    4   0    34  1  1   647  W 28-10
nwe 2006  w  nyj    2   0   127  1  0   640  W 37-16
den 1996  d  jax    4  -1   126  1  1   639  L 27-30
stl 1999  c  tam    2   0   249  1  0   638  W 11- 6
phi 2004  d  min    5   2   116  1  1   629  W 27-14
sea 2005  d  was    3   0   115  1  1   628  W 20-10
nyg 1993  w  min    2   0    96  1  0   609  W 17-10
stl 2001  c  phi    3   1    95  1  0   608  W 29-24
====================================================
WEIGHTED AVERAGE:  86.2 pct chance of victory
PROJECTED SCORE:  24.6-13.0

Over at the footballguys message board, it was noted that 12 of the Packers' 22 starters this season were drafted in the fifth round or later, or not at all. As I pointed out in the same thread, this is a very high number, but not quite a record. The 1984 Steelers, who went 9-7 but won their division and a playoff game to make it to the conference championship game, had 14 of 22 their starters being 5th-or-later or undrafted. A handful of other teams since 1980 had 13.

This got me thinking about the more general question of how teams are built. So I looked at all 22 starters on every team from 1980 to 2006 and sorted them into categories according to what round they were drafted in. Why 1980? Because that's roughly the point where the guys who were drafted by both the NFL and the AFL had disappeared from the league. If a guy was drafted in the 2nd round in the AFL draft and the 3rd round of the same year's NFL draft, how should we classify him? What about a guy who was drafted in the first round of the NFL draft and the 10th round of the AFL draft? Rather than wrestle with that sort of question, I decided to declare that the data from 1980 forward would give us enough to chew on, at least for now.

OK, so here is the the summary data:

(Continued)

Thanks to a reader named Vishal for this email:

Just a general question for perhaps a blog topic. I noticed that not one player over wildcard weekend had more than 70 yards rushing. A few of the teams went over 100 total, but was it anomolous to not see one player post a higher total than 70, especially when you consider than 7 of the top 12 running teams in terms of rush yards per game were in action? Was wondering if we've seen any playoff weekends like this.

Answer:

(Continued)

The footballguys message board has a pretty interesting Parcells vs. Gibbs thread going on right now.

We've all heard that Gibbs is the only coach to win Super Bowls with three different quarterbacks. A poster named dgreen takes that a step further to notes that Gibbs has made the playoffs with six different quarterbacks: Joe Theismann, Jay Schroeder, Doug Williams, Mark Rypien, Mark Brunell, and Todd Collins (or Jason Campbell, depending on how you look at it).

Per my typical MO, I'll not spend too much time debating the importance of this as a measurement of coaching success (while at the same time inviting you to debate it in the comments if you like). Instead, I'll focus on the more trivial task of figuring out exactly how this compares to other coaches.

(Continued)

Last year I whipped up a fun method for predicting playoff games. The idea is this. Let's say you're trying to figure the Chargers/Colts game this weekend. One way to go would be to try to find previous playoff games where the two teams had characteristics that look similar to those of the Chargers and Colts, then see how those games turned out. I don't have much to add, so I'll just quote last year's post, but with the Colts' and Chargers' data filled in.

(Continued)

Browns' rookie Joe Thomas was named to the Pro Bowl today. He'll be replacing the injured Jason Peters of the Bills. Thomas is only the fourth offensive lineman since the merger to make the pro bowl as a rookie. The other three were Marcus McNeill (2006), Richmond Webb (1990), and Chris Hinton (1983).

An email from a good buddy of mine (and occasional commenter here):

i am so sick and tired of hearing "its so hard to beat a team three times in a season". idiots. yes, going 3-0 vs a playoff team is hard. BUT NOT IF YOU SPOT TEAM A TWO GAMES YOU MORONS.

He's right, of course. This is essentially the same reason why black isn't necessarily a good bet on a roulette wheel that's come up red on the last ten spins. Yes, it's incredibly unlikely for a wheel to land on red on eleven straight spins, but given that it's already landed on red for the last ten spins, it's not at all unlikely for it to land on red for the eleventh. Likewise, before the season, the Cowboys beating the Giants three times would have been something of a longshot. But now that they've already done two-thirds of the work, it's not.

Anyone making the hard-to-beat-a-team-three-times claim probably is suffering from a failure to understand conditional probability, but that doesn't necessarily make them wrong. Football games are not roulette wheels. In particular, they're not independent. Maybe teams learn more from losses than from wins, or something like that, and it really is hard to beat a team three times, even if you've already beat them twice. That's an empirical question. Let's check it out.

(Continued)

There seems to be a generally held belief that teams with byes in the first round of the playoffs have an increased advantage in the semifinals. For example, since 1990, the home teams in the semifinals have won 77.9% of games, compared to 69.1% in the wildcard round, and 58.8% in the championship games. Not only has the winning percentage been higher, there have been more noticeable blowouts in the semifinals than any other round.

The problem with these numbers is that they do not control for matchup. To (try to) answer the question of whether the bye week increases home field advantage, I will use the regular season SRS ratings going back to the AFL-NFL merger, and look at the average expected results, based on regular season ratings, and average actual results in the playoffs for the home team. (I did exclude the two strike seasons of 1982 and 1987, but included every other year from 1970-2006).

(Continued)

This post is a fuller description of a study I did for ESPN Magazine. They asked me to look at the NFL success of the winners of the various postseason college football awards, such as the Heisman, the Outland, the Lombardi, and so on. Do Heisman Trophy winners, in general, have more or less NFL success than winners of the Doak Walker Award, or the Jim Thorpe, or the Biletnikoff? Which college award has traditionally produced the best pros? Let's find out.

(Continued)