TCU has played much better in the second half of games this year. What other schools have seen dramatic shifts of play after halftime?
04 Apr 2007
Guest Column by Mike Horn
Now that free agency has drifted into the second phase (also known as "Daniel Snyder has run out of cap room"), the draft is the next big event on the NFL calendar. I needed a football fix not generated by a Peter King column. I'm not a draftaholic, but I started wondering what positions get drafted most frequently by which teams. Regular readers of Football Outsiders know about the Patriots and tight ends, but what other patterns exist?
To look for those patterns, I compiled five years of data from the extraordinarily useful drafthistory.com. (Why five years? So we don't have to adjust for the non-existent Houston Texans.) What follows is more reference material than analysis, although there are some comments. I don't think this data can necessarily help you predict this draft -- and you probably know more about your favorite team than I do. But when your local blowhard says (or writes in his blog), "Team X always takes a defensive back in the draft", you have some charts to refer rather than digging around at drafthistory.com for half an hour on the company's time. And maybe you can point out to your new best friend that Team X has drafted fewer defensive backs than any team except the Cardinals over the last five years.
Here's the raw data of teams and their draft picks by position for 2002-2006:
|Draft Picks From 2002-2006 By Team and Position|
Again, the draft information is from drafthistory.com, so safeties are not separated for cornerbacks and fullbacks are not separated from halfbacks. Positions reflect what the players were drafted as, not what they play in the NFL. Wins are regular season only, so please no complaints that your franchise has been disrespected.
The cells highlighted in green reflect a maximum value for that category; cells in yellow reflect a minimum. At the end of the article, I've included tables breaking the draft into its first and second day.
Here's the same information as in the previous table, but displayed in a different way (plus a little Day 1/Day 2 data thrown in):
|Summary by Position of Draft Picks 2002-2006|
|POS||Total||MAX/TM||MIN/TM||Day 1||Day 2||D2/D1|
The color-coding reflects positions drafted at roughly the same frequency, for example, the numbers of WR and LB drafted have more in common with each other than with the other positions.
The MAX/TM columns capture the maximum number drafted at each position by a specific team (or teams) in this five-year period. The MIN/TM is the same information for the minimums. "Various" means a lot of teams drafted the maximum or minimum number.
The Day 1 column shows the number taken at each position in Rounds 1-3 of the draft; Day 2 is for Rounds 4-7. D2/D1 shows the ratio of players at each position drafted on the second day vs. the first. Since there were more second day picks (800) than first day (488), the overall ratio of 1.64:1 is reflected in the Grand Total row.
I've highlighted the "LB" D2/D1 number in dark blue font because it is the smallest number by a considerable margin. NFL teams would seem to want to draft linebackers on Day 1. On the other hand, I've shown the "G" D2/D1 number in red -- guards seems to be something the NFL goes after on Day 2 of the draft. Kickers and Punters are also Day 2 positions (must... not... make... Mike Nugent... joke...).
Not surprisingly, there are more defensive backs drafted than any other position: defenses almost always have more DBs on the field at a time than players at any other position, and often play five or more. Plus DBs have value on return and coverage teams. What did interest me was that the Colts had drafted 14 DBs, easily the most players at one position by any one team over the last five years. The second most common choice was DBs by the Titans (12) who had the most total picks (50) in five years, so they could spread them around more. The Colts used a third of their picks on DBs -- no other team used over a quarter. Now the Colts have been known to use a DB as a linebacker, but that's still a lot of DBs (and they drafted an above average number of LBs too).
At the other end of the DB spectrum are the Cardinals. They only drafted two DBs in five years: both on Day 1 in 2005 (Antrel Rolle and Eric Green). No other team drafted fewer than five DBs. There doesn't seem to be much correlation between drafting DBs and wins (0.23, but only 0.03 if I take the Colts out and -0.17 if both the Cardinals and Colts are removed from the sample). I wonder what the Arizona front office thinks it knows that the rest of the league doesn't.
Wide receivers are the second most-drafted position. I would have thought it would be linebackers, as most teams have three or four LBs in their base defense compared to two or maybe three WRs in their most-used offensive alignment. But except for the odd quarterback or two converted to WR, most NFL receivers seem to come into the league at that position, while some LBs are drafted as DBs or DEs. So perhaps that explains it.
In any case, the 49ers have been the team drafting the most WRs. They've also tied for the top spot in drafting QBs and third in drafting TEs. My guess is that the constant changes in offensive systems have caused them to have to repeatedly draft players with different skills at these positions. In the final year under Mariucci and first year under Erickson, San Francisco was running the West Coast Offense. In Year 2 of the Erickson Era, quite a big deal was made of abandoning the WCO that had been in place since the Walsh Dynasty. Then the Nolan Restoration brought back the WCO in 2005 with OC Mike McCarthy, but only briefly because Norv Turner came in with his offense in 2006 (the last two 49er OCs got head coaching jobs -- I guess the NFL really does reward excellence). Perhaps all this churning of systems led to draft pick after pick being spent on players suited to the offense du jour who became superfluous in the next OC's scheme. Or maybe the 49ers just found WRs, QBs, and TEs to consistently be the best players available. Or they could just have drafted badly...
Then there's the Colts, who haven't drafted a wideout since Reggie Wayne in 2001. And have kept their offensive system in place since before Jim Mora Sr. was wondering if his team would make the playoffs. Interestingly, the Colts have been emulating the old 49ers, who went 20-some years with one offensive system. Or maybe the Colts were so happy with their two main wideouts, Manning-favorite Brandon Stokely, and TE/WR Dallas Clark that they didn't need any more receivers.
I have a sneaking suspicion that the Browns are to linebackers what the 49ers are to wide receivers. I'm not as familiar with the Brown defensive coordinator travails as I am with the 49ers' OCs. I do note that the Romeo Crennel/Phil Savage Browns have drafted four LBs in two years. Meanwhile, the Patriots have drafted one "LB" in five years (a fifth rounder at that). New England did take a median number of DEs (four) in this period, two of whom are now listed at linebacker (Tully Banta-Cain and Jeremy Mincey -- Mincey never played with the Pats and was picked up by the Jaguars). In any case, the New England front office has tried to man its 3-4 with linebackers accumulated less through the draft than the Browns even though their head coaches presumably run comparable defensive systems (the Patriots drafting three LB-types in five years vs. the Browns taking four LBs in the last two years alone).
I also noticed that two teams that I think of as having strong linebacking corps, Pittsburgh and Denver, each only drafted two LBs. When I looked at the Steeler depth chart, I saw that only Larry Foote (Round 4, 2002) among its 2006 starters entered the league in the time frame I'm examining. I'm sure a lot of you knew this, but I hadn't realized the degree to which the Steelers LB corps was aging and not being replenished through the draft. The Broncos had used a #1 on D.J. Williams in 2004, and need fewer LBs in their 4-3 alignment than the Steelers' 3-4 requires, so their low number of LBs drafted is a little less remarkable.
Meanwhile, the Broncos spent six picks accumulating running backs to plug into their vaunted star-RB assembly line, and that's without drafting any RBs 2006. Of course, last year they signed undrafted rookie free agent Mike Bell instead of using a pick on him. Neither Denver nor Tennessee, the other team drafting six RBs from 2002-2006, drafted one in the first round. The teams only taking one back (Giants, 49ers, and Packers) took their choices in round 3 or later as well.
The Chargers have consistently taken an offensive tackle in the late rounds (missing only 2003 but picking two in 2004) -- even last year when they found wunderkind Marcus McNeill 50 players into the draft they also took Jerome Clary in the sixth round. Meanwhile the Patriots haven't spent a pick on an OT since 2001 (Matt Light and Kenyatta Jones). Just to illustrate how drafting patterns can change, though, New England took five OTs from 1999-2001.
Defensive tackle, which falls right in the middle of all the positions in popularity as a draft pick, seems to inspire strong opinions. Eight different teams tied for the maximum number of DTs chosen (five) while four other teams only drafted one each. Aside from kickers and punters, only centers had that many teams at the extremes. Six teams took the maximum number of centers (three) and six drafted the minimum (zero).
Probably no one who has bothered to read this far is surprised that the Patriots are a leading consumer of tight ends. But the Rams? For the last five years, the Rams have thrown the ball to their tight ends less than any other team in the league (226 pass targets, just ahead of the Steelers' 230 and of course Pittsburgh threw the ball a lot less overall than the Rams). Even last year with Mike Martz gone and two rookie tight ends taken on the first day of the draft, the Rams seldom threw to their TEs.
The Falcons took Alge Crumpler in 2001 and haven't drafted a TE since. The Packers last took a TE in 2000 (Bubba Franks) not counting WR-convert David Martin, picked in 2001. That ties the Ravens (guard) for the longest current streak of not drafting at a position.
The Ravens do like to draft QBs, four in six years, although they mixed in three sixth or seventh rounders with Kyle Boller before turning to FA to fill the position with Captain Checkdown. And the Texans, despite taking David Carr at #1 in their very first draft, didn't hesitate to take three more QBs in 2003-2004. Meanwhile the Cowboys and Dolphins haven't drafted a QB since 2001 (Quincy Carter and Josh Heupel) although they've certainly picked up quarterbacks through other means, including Drew Henson, who was originally a sixth round pick in 2003 - by the Texans.
Centers were the least drafted non-special teams position. But when you consider that there are half as many centers (and quarterbacks) starting in college as guards, proportionately a larger number of those centers are getting drafted by the NFL. Guards are the least likely position to get drafted relative to their supply (and demand). And, as already mentioned, guards have been viewed as "second day" players in the draft. It will be interesting to see if the increased value placed on guards as measured by their highly-publicized large FA contracts translates into a different draft demand -- or if those contracts were just a small sample aberration.
Say what you want about the Jets, at least they only drafted one kicker. The Jags, Chargers, and Colts have taken two each -- accounting for half the kickers chosen in the last five years. And the Colts don't even have their choices (admittedly a sixth and a seventh) on their roster anymore.
That brings me to Seattle and its punters. There's a tale to tell. But it's already been told (see Pro Football Prospectus 2006).
The next chart rolls up the positions into some larger groupings:
|Draft Picks From 2002-2006 By Team and Position Grouping|
I've used bright green and bright yellow as I did in the first table. I'll explain the paler colors below. A key to the letter codes:
D = Defense; total defensive players drafted.
O = Offense.
KP = Kitchen Police (just checking, actually Kickers and Punters).
D/O = Percentage of defensive players chosen; the chart is sorted on this column.
DB = Still defensive backs.
B7 = Back Seven, DB + LB in the first table.
F7 = Front Seven, LB + DT + DE.
DL = Defensive Linemen, DT + DE.
OL = Offensive Linemen, C + G + T.
OO = Other Offensive Players, commonly mis-termed "skill players" as if any big ol' fat guy can play the line.
TR = Trench Ratio, the number of lineman (OL + DL) chosen divided by the total number of offensive and defensive players drafted. Not actually a ratio, now that I think about it.
The Colts have a distinct profile on this chart: they have drafted the most defensive (27) and fewest offensive (13) players, both in raw numbers and in the percentage of defensive players taken (0.68). Those defenders have been heavily weighted toward their back seven, and have minimized the number of non-linemen on offense (only six OO). The result has been the most regular season wins over the last five years. Obviously, having and retaining key offensive stars has been part of that and cap management has been linked to their draft strategy.
I was under the impression that defenders were more often drafted than offensive players -- maybe because I'm a Chiefs fan and they have been one of the heaviest drafters of defenders (percentage-wise). In fact, the league slightly favors drafting for offense (650 to 615).
The Bucs have been the team drafting the most on the offensive side of the ball, both in total numbers (28) and weighting (D/O = 0.36). I think that equates to "aging defense". The fact that Tampa has also drafted the fewest front seven defenders, tied with the draft pick-deprived Redskins (DPDR), in the last five years is also notable.
In the D/O column, I've highlighted in pale green and yellow the teams with the second highest and lowest percentages of defensive players drafted. You can see that the gaps between the Colts (0.06) and the Bucs (0.03) and the next closest team on either extreme are pretty large; only the spread from the Vikings to the Chiefs (0.05) is that big anywhere else. Those two franchises have really been outliers in how they've drafted.
In the various position groupings, the Cardinals are notable because of their focus on their front seven. They tied for the lead in most DL drafted, and that, combined with an above average number of LBs selected, also gave them the lead in the front seven category.
Several teams tied with the Cardinals in drafting nine DL, but the Bears stand out because they didn't place similar emphasis on the OL, where they and the Falcons only drafted four players. There is always a lot of talk about whether the Falcons' woes in the passing game are due to the quarterback or the receivers. But at some level you have to wonder if the lack of talent drafted along the line has also been a factor. I know the zone-blocking scheme and line guru Alex Gibbs were supposed to compensate, but drafting no first day lineman since 2000 might not be the way to go.
I already mentioned the Chargers' having drafted the most offensive tackles. They also totaled double-digit OL overall, the only team besides the Eagles to spend that many picks on the O-line. Because of that, San Diego had the highest percentage of players drafted for the "trenches" (0.46) by a couple of percentage points. The Eagles and Dolphins (who both took the most guards) tied for second in this category (0.44), also a few percentages points above the next team. The Eagles, though, drafted 19 lineman to the Dolphins' 14, as Philadelphia had many more picks than Miami.
Then there's the Lions. Apparently former LB Matt Millen, among his other quirks, doesn't have much use for drafting linemen. His team chose the lowest percentage of players in the trenches from 2002-2006 -- and only the DPDR drafted fewer lineman total. None of the Lions' OL picks were on Day 1.
Noted with interest: the Ravens drafted the most OO players of any team, over three times as many as the Colts. No wonder that playoff game was the Ravens offense vs. the Colts defense.
Thanks to drafting three punters, the Seahawks dominate the KP category. Over half of all franchises have not drafted a KP in the last five years. The correlation between KPs drafted and wins was higher (0.40) than for any other category in the table. Not a real strong correlation and I have no explanation for it other than perhaps good teams are more willing to draft specialists than bad ones, but there it is. Several of the correlations were close to zero and I've already discussed the DB-wins correlation. Two positive correlations with wins might have some meaning, OL (0.21) and TR (0.25), and obviously they are not independent of each other. The largest negative correlation (-0.15) was for the front seven. The more statistically-educated readers can speak more authoritatively than I can on this, so I'm not going to comment further.
The next chart really just summarizes a lot of the information in the first table: the total number of draft picks each team had and the positions on which they focused their drafting (or ignored) relative to the rest of the league:
|Summary of Maximum and Minimum Draft Picks
By Position From 2002-2006
|TEN||50||2||RB, DT||2||K, P||40|
|SF||47||3||WR, DT, QB||1||RB||30|
|BAL||45||2||QB, C||2||G, K||45|
|TB||44||0||--||3||DT, K, P||39|
|GB||44||2||DT, C||3||RB, TE, K||44|
|HOU||44||3||DT, QB, C||2||K, P||24|
|CHI||42||1||DT||3||C, K, P||40|
|SD||42||3||T, C, K||0||--||47|
|IND||42||2||DB, K||2||WR, P||60|
|CAR||41||0||--||3||DT, K, P||44|
|NE||41||2||DT, TE||3||LB, T, P||59|
|OAK||41||0||--||3||DT, K, P||26|
|DAL||40||0||--||4||DT, QB, K, P||39|
|DET||39||0||--||4||DT, C, K, P||22|
|ARI||36||1||DT||3||DB, K, P||25|
|NYG||36||1||DE||3||RB, K, P||39|
|ATL||34||0||--||4||TE, C, K, P||40.5|
|MIA||32||0||--||4||DE, QB, K, P||38|
|WAS||29||0||--||4||DE, C, K, P||33|
The MAX/POS section shows at how many positions the team took more players than any other team (including ties) and lists what those positions were. MIN/POS is the same thing for the fewest picks made at the various positions.
The shaded area includes the teams within +/- 2 total draft picks of the mean (40.3). This includes half the league, with the first quartile above the shaded area and the fourth quartile below. The first quartile averaged 39.5 wins per team. The middle two quartiles averaged 41.6, and the bottom quartile averaged 37.5 wins. It would generally appear that it's better to have more picks than fewer, but that gathering too many picks is not a great strategy in the short run. The overall correlation between the number of picks and wins is only 0.14 so perhaps I don't know what I'm talking about.
Only two teams were more than two standard deviations away from the mean number of draft choices; the Titans on the high side and the DPDR on the low end.
Three teams had extremes (Max or Min) at five positions. That one of them was the Patriots argues for the value of staking out a definite drafting philosophy and sticking to it. That the Packers (slightly above average in wins from 2002-2006) and the Texans (one of the worst franchises) were the others weakens this argument.
Three other teams had only one extreme position, and that was either kicker or punter. The Seahawks, Jets, and Vikings have been the teams least likely to focus on a particular position when drafting (well, except for the Seahawks and punters). The Saints and the Steelers both only had extremes (minimums) at kicker and punter, so they too didn't have a noticeably strong preference or aversion for the other positions. Every other team recorded a maximum or minimum for some position besides the kicking specialists.
Lastly, here's the Day 1 -- Day 2 splits:
|DAY 1 Draft Picks From 2002-2006 By Team and Position||2-Day
The Day 1-2 tables use the same formatting as the table for all draft picks except the minimums are only selectively highlighted as there were so many zeros.
The samples are pretty small to analyze meaningfully. Some brief observations of where particular teams compared to the league or to themselves on (and between) Day 1 and Day 2 choices:
As I said at the beginning of the article, you can't necessarily predict this year's draft with all this information. You can see tendencies, but a team that always drafts LBs that just signed three in free agency is probably going to go away from that tendency. Teams that haven't taken a player at a position in a while may now have a need there. And franchises that have taken a lot of linemen may now be full up. You need a lot more detailed knowledge to make draft predictions. This may be one small piece.
Mike Horn writes a stats-based column for FantasyGuru.com. In real life he is a retired Army officer working as a military defense analyst in the Washington, DC area. If you are interested in writing a guest column for Football Outsiders, something with a unique take on the NFL, please e-mail info-at-footballoutsiders.com. Readers should know that, with less football news to run on FO, May and June are great months for submitting a guest column.
44 comments, Last at 23 Apr 2007, 10:47pm by Mark