Warning: Cannot modify header information - headers already sent by (output started at /home/content/85/8979285/html/wp-includes/post-thumbnail-template.php:1) in /home/content/85/8979285/html/wp-content/plugins/wp-super-cache/wp-cache-phase2.php on line 62
Main Menu | The DC Times - Part 2

The DC Times

A New Way to Look at the Cowboys, NFL, and Fantasy Football


The Best of The DC Times

Jonathan Bales

Throughout the years, I have uncovered a wealth of unique information on the Dallas Cowboys.  A lot of my analysis has been harsh, particularly back in the early years, but I wanted to address the fact that the ‘Boys are improving in a vast array of categories which might not be perceptible to the average fan (or any loser who doesn’t track every play in Excel like myself).

To exemplify this, I have listed some of my most riveting work since the start of 2010.  The majority of it is the stat-heavy analysis (often focused on play-calling), and by tracking the progression of the articles, you will see Jason Garrett & Co. are really improving in a number of aspects.  Superior draw play usage, second down play-calling, weak side runs, formation-specific play-calling, and so on have helped the Cowboys’ offense.

Does the team still have room for improvement?  You bet.  In my view, the ‘Boys need to work (a lot) on playaction usage, counters, the deep passing game, predictability and more.  The point, though, is that we are headed in the right direction.

*My personal favorites are in bold.*


“Scouting the Draft” series at New York Times

Assessing Draft Strategy

Interesting 2011 Stats


New York Times: Cowboys Should Have Passed

Advanced NFL Stats: On Cowboys’ Predictability

New York Times: Keys to Defeating Dallas

Dallas Observer: Cowboys v Giants

Ultimate Playaction Pass Guide

Why Cowboys should throw deep more

To run or not to run?

Weak Side Runs and Game Theory

Felix Jones: Turf vs Grass

Garrett Tipping Plays Via Formation

Motion Statistics

Team Position Rankings

Can Ware break sack record?

Jason Witten’s Red Zone Performance

Chris Gronkowski vs Deon Anderson

2011 Draft Guide

Run/Pass Dichotomy

Pass Rates By Personnel

Romo vs Blitz/Perceived Blitz

Attempting More Two-Point Conversions


Ultimate Playaction Pass Guide

Second Down Play-Calling

Draw Plays

Double Tight Strong

Could Emmitt be college football’s all-time leading rusher?

Counter Usage

Will anyone break Emmitt’s rushing record?

Play-Calling By Personnel

Rushing/Passing Efficiency By Down

Tight/Spread Runs

Is Kitna more accurate than Romo?

Motion Stats

Weak Side Runs

Romo’s Audibles

Should Witten block less?

Deon Anderson’s Importance

Third Down Running

Throwing From Double-Tight

Best Running/Passing Formations

Gun Trips

Player Rankings

Wacky Stats

Fourth Down Play-Calling

Fourth Down: Momentum

How costly are false starts?

2nd and 1 Play-Calling

Importance of Kicker

Third Down Play-Calling

2010 Draft Guide

Initial Drive Stats


Reassessing Best Player Available Draft Strategy: Why Teams Should Often Bypass BPA To Maximize Overall Value

Jonathan Bales

“Reassessing Best Player Available Draft Strategy: Why Teams Should Often Bypass BPA To Maximize Overall Value”

Sounds more like the title of a thesis than a blog post.  Nonetheless, I wanted to again delve into what is one of my favorite football topics: draft strategy.  Last year, I published an article called “Why Selecting Best Player Available in NFL Draft a Myth.”

The post highlighted a few of my unconventional (and wildly unpopular) thoughts on draft and game theory, the most intriguing of which is that selecting the best player available, even at a position of need, is often a mistake.  I will recap that article with a few quotes, but I suggest clicking the link above and rereading it if you are bored at work, or entertained but simply don’t have a lot of work to do, or even if you have a ton of work, or if it is nighttime and you are off of work, or if your name is Betty.  Just read it.  Here we go. . .

Like many (or even most) of the long-held NFL “truisms,” the concept of selecting the BPA (best player available) is mistaken.  As is the case with punting on 4th and 1 or always kicking extra points, selecting the BPA will actually lead to sub-par results.

GMs who say they always take the BPA are simply lying.

The key to this strategy is a concept I’ve discussed in a few of my fantasy football articles–VORP (value over replacement player).  In a nutshell, VORP means selecting not the player with the most projected points, but the player with the largest disparity of projected points compared to the next player at the same position who you could secure in a later round.  If you read between the lines, you can see game theory is really the backbone of this strategy.  To effectively maximize value, it is critical to understand perceived worth.

Game theory is all about understanding opponents’ beliefs and using them to your advantage.  If you had perfect knowledge of other teams’ draft boards and knew the top player on your board was rated three rounds lower on everyone else’s, you would wait to select that player.  BPA, even at a position of need, promotes the dismissal of potentially useful information.

Although no team has perfect knowledge of a player’s perceived value, the notion that perceptions can and should alter draft theory remains unchanged.  BPA as a draft strategy is too shortsighted and could force premature selections, ultimately decreasing overall value.

On a real world example from 2011:

Let’s assume the ‘Boys will select a defensive end and an offensive tackle in the first two rounds, but they’re unsure of the order.  Now, let’s provide a numerical value to the possible targets.  As a guide, we will use the NFL draft trade value chart and my own 2011 NFL Draft Big Board to assign these values.

As I’ve already proposed, let’s assume Dareus (ranked No. 2 on my Big Board) is available for the Cowboys.  At that ranking, he’s worth a whopping 2,600 points.  As I’ve argued in the past, however, I think there is a major problem with selecting a defensive end in the first round.  By the time the Cowboys’ 40th selection rolls around, there is zero chance that a top-tier offensive tackle will be left on the board.  My top five tackles–Tyron Smith, Ben Ijalana, Anthony Castonzo, Gabe Carimi, and Derek Sherrod–will almost certainly be gone by the second-round.

Thus, the top offensive tackle that is left to pair with Dareus, according to my personal Big Board, is Alabama’s James Carpenter. . .all the way down at No. 71 overall.  According to the value chart, that selection is worth 235 points, bringing the Dareus/Carpenter duo to 2,835 combined points.  Certainly our VORP has been compromised, as Carpenter is terrible value in the second-round.  But is Dareus’ BPA status enough to compensate?

To determine this, let’s project the Cowboys’ possible selections if they take an offensive tackle in the first-round.  At No. 9, the ‘Boys may very well have their pick of the litter, and according to my board, Tyron Smith (No. 8 overall) is that guy.  The eighth overall selection is worth only 1,400 points–a far cry from the 2,600 that we assigned to Dareus.

We can already see the Dareus/Carpenter duo is going to win out.  Even if the Cowboys somehow land Cal’s Cameron Jordan in the second-round (which is clearly a pipe dream), his 14th overall ranking–worth 1,100 points–would still bring the Smith/Jordan duo to only 2,500 overall points–335 behind Dareus and Carpenter.

Although the Cowboys selected an inside linebacker in the second round, they still landed Smith in the first.  My hunch is that he was their BPA, but he should have been selected even if they had a prospect rated higher than him.

It’s worth noting that, although the optimal tandem turned out to be that which was comprised of the BPA, the process by which we discovered that was still VORP.  Thus, teams will often arrive at the right selection, but implement the wrong method of getting there.  Selecting the correct player helps you now, but selecting the correct player by utilizing the proper draft strategy will help you in the future.

Mathematics often leads to counterintuitive results, but the teams which disregard their “gut” and utilize the numbers on draft day are generally the most successful.  Remember, the “gut feelings” are already implemented into a team’s rankings (whether they admit to it or not).  Draft day is not the time to follow hunches.

What I mean here is that film study, interview results, and other non-measurables are already reflected on your board.  Ironically, if you disregard your board on draft day (in relation to VORP, not BPA), you will actually be forgoing those gut feelings which were already implemented into your rankings.

VORP is an all-encompassing draft strategy that leads to greater ultimate value than BPA–a more short-sighted draft philosophy which disregards the future in favor of optimal value right now.  Would you rather have $100 today (BPA) or $500 tomorrow (VORP)?


After I posted that article, I found more support in my critique of BPA draft strategy from Code and Football in their article on why drafting the BPA is simply a way to optimize buyer’s remorse:

Consider this scenario: you have three players in the middle rounds you are considering, whose “true career value” is about the same. We’ll assume drafting is an efficient market for now, so our estimation of the value of these picks is a normally distributed estimate whose mean is based off their true career value. Which one of these men do we draft? We draft the player whose value we have overestimated the most. Consequently, we draft the player most likely to underachieve our expectations.

Since in most drafts there are very few times a true BPA falls into the lap of teams (i.e. players where one is wildly superior to all other candidates), it would seem that BPA is a way of optimizing how heartbroken a team will be over the draft choices it actually picks. Though this approach would appear to gather the best athletes, in a draft with a large error, and multiple situations where you’re picking from nearly equivalent athletes, perhaps all BPA will get you is maximally suffering from buyer’s remorse.

This idea has differences from my own, but it points out the fact that the value of taking the BPA is often minimized because of a team’s overestimation of a prospect’s ability.  The argument that BPA can help a team secure the most elite prospects, then, seems less compelling.  With 32 teams all acquiring nearly the same information, the chance that a single team will obtain a player whose value is so much greater than the other options that it overrides the value of VORP is slim at best.


In the two main articles I posted on draft strategy (here and here), we had over 50 really insightful comments.  Many of these raised interesting critiques of VORP, game theory, and other draft strategies, and I wanted to address a few of them now. . .

“VORP doesn’t address the ‘real world’ value of specific positions.”

It does.  Much as the measure of “intangible” things like heart and determination are actually reflected in advanced football statistics, the value of a specific position over another (quarterback to linebacker, for example) is reflected in a team’s draft board.

“VORP as a long-term strategy will lead to less overal talent than BPA.”

VORP leads to the greatest overall value because it has a far greater focus than BPA, or BPA at a position of need (BPAAAPON, if you will).  VORP is in the business of temporarily bypassing short-term value to secure greater value in the future.

The most valid critique of the draft theory is that it requires too much knowledge, i.e. you can never have enough knowledge of other teams’ intentions for the draft theory to work in practice.  It is a pipe dream, some might say.

While this might be possible, I think that argument would be a better one when applying VORP to the later rounds of a draft.  For the most part, we all know which players will go in or around the first couple of rounds, so predicting the abundance of talent at a specific position is made easier early in the draft.  That task only becomes very difficult after the first few rounds when boards do not match up as comparably.

In effect, VORP becomes a less valuable draft strategy as the draft rolls along and opponents’ beliefs become less predictable.  In the later rounds, when the goal is to maximize upside (as opposed to the goal of minimizing downside early in the draft), selecting the BPA has more merit.

“The draft is a crap shoot, so VORP is no more valuable than any other strategy.  Just select the best players.”

The draft is a crap shoot in which a large majority of success stems from luck, but at the same time that doesn’t negate the value of specific draft strategies.  The idea is comparable to blackjack, where the outcome of any single hand is determined almost solely by “luck.”  A great blackjack player might win perhaps one out of 100 hands more than an average player.

In a sport like football where the competition is so stiff, though, very small advantages equate to big success.  It is the difference between a DeMarcus Ware and a Bobby Carpenter, for example.  One draft selection can lead to a monumental difference in production.

“VORP is a baseball term and should not be applied to NFL draft strategy.”

I simply use VORP (value over replacement player) as a label.  Change it to ‘position scarcity’ if you would like.  The idea is the same: short-term bypassing of the BPA can lead to overall greater value, particularly in the draft’s early rounds.

“Too many things can happen in the future which make using VORP now usless.”

The same thing can be said for any draft strategy.  We can only work with the information we have at hand.  What if Tony Romo tears his ACL in camp?  There are a lot of things in the future that could alter the efficiency of past decisions, but the best way to maximize the opportunity for future success, and the only way, is to use all present information.

“VORP assumes that each draft is an independent variable and the potential of future drafts has no bearing on the draft this year, which simply isn’t true”

Again, all present information can and should be factored into rankings.  I’d say the ability for any organization to effectively evaluate the positional talent for a draft which is over a year away, however, is small enough that it should have little to no bearing on current decisions.  In addition to not knowing draft slots, teams don’t have a grasp on which prospects will enter the draft, which ones might get hurt, how they will perform in their final collegiate seasons, and so on.  Bypassing a talented player at a position of need which VORP draft strategy suggests to select because the subsequent draft might be full of talent at that position seems like a poor philosophy regardless of the draft technique used.

“The draft is an inefficient market, devaluing the use of VORP.”

VORP’s value would certainly be decreased if the draft was an inefficient market, but for the most part it is very efficient.  Yes, teams hit and miss all the time, but that doesn’t make the market remarkably less efficient.  When we take a large sample size of drafts into account and analyze the success of players based on draft slot, we see a very specific downward curve.  Whether we compare draft spot to years as a starter, Pro Bowl selections, or career approximate value, higher picks tend to perform superior to lower ones: http://www.advancednflstats.com/2009/04/career-success-by-draft-order.html

In short, although there are random deviations (which are to be expected), teams are generally efficient in the draft market.

Like Dallas Cowboys Times on Facebook

Follow Dallas Cowboys Times on Twitter



Is Jason Garrett Too Aggressive? Actually, Stats Say He’s Not Aggressive Enough

Jonathan Bales

One of the things which irks me most during the football season comes on Mondays after a loss, when all of the Dallas-area analysts come out of their holes and (usually) discuss why Jason Garrett’s “wild” and ultra-aggressive play-calling cost the Cowboys a victory.  How can the ‘Boys win football games when their coach is going for it on 4th and 4?

I agree with the writers at ESPN, Dallas Morning News, and the other stupendous media outlets that the level of aggressiveness in Garrett’s play-calling and decision making needs to change.  Specifically, Garrett needs to become (much) more aggressive.

One aspect of what is generally considered aggressive play-calling is going for it on fourth down, although this can be somewhat of a mislabel.  In many situations, going for it on fourth down (say, 4th and 3 at the opponent’s 40-yard line) is far and away the correct call, so labeling it as ‘aggressive’ is a bit of a misnomer.  It is aggressive only insofar as it flies in the face of “conventional” football wisdom.  Don’t confuse being aggressive with being risky.  In reality, punting the ball in such situations is almost always the true risky play.

I often discuss how Garrett’s unjustified fourth down punts cost Dallas points (and wins).  I recently came across some numbers which show that Garrett is not only not as aggressive as people believe, but also that his decisions are decreasing the team’s win probability in a significant way.

Over at Football Outsiders, the 2011 Aggressiveness Index is up.  There, you can see how often NFL coaches went for it on fourth down while in opponent’s territory this season.  Garrett’s rank: 25th. . .not as aggressive as you thought, I am assuming.

Of course, game situations can alter coaching strategy, particularly with a somewhat limited sample size of fourth down attempts in a single season.  Garrett isn’t off the hook just yet, though, as Advanced NFL Stats recently displayed some fourth down numbers as well.  Unlike at Football Outsiders, ANS takes game situations into account.  Brian Burke writes:

But because every situation is unique in terms of distance, time, and score, we can’t make any judgments about the aggressiveness or timidity of any coaches yet. The next table totals all the WP forfeited by each team on 4th downs when the numbers said go but the coach said kick. Also listed are the total number of 4th down go opportunities as well as the WP forfeited per opportunity. The higher the WP Forfeited number, the greater the sum of the 4th down errors.

According to the numbers, Garrett forfeited 0.74 wins in 2011 due to poor decision-making on fourth down–the seventh-worst mark in the league.  The teams worse than Dallas are Arizona, Denver, Seattle, St. Louis, Cincinnati and Cleveland. . .hmmmm.  Among the league leaders (i.e. best fourth down play-calling) were San Diego, Detroit, New England and Green Bay. . .another hmmmm.

On top of that, the Cowboys had only 36 fourth downs all year when they “should have” gone for it, making the win probability Garrett forfeited per opportunity (.021) the second-worst mark in the NFL.  So for those who are claiming the Garrett and the Cowboys need to “cool it” with the aggressive play-calling. . .you’re simply mistaken.  In all practical terms, Garrett was actually the second-most conservative coach (in terms of fourth down decisions) in the entire NFL in 2011.

Like Dallas Cowboys Times on Facebook

Follow Dallas Cowboys Times on Twitter


Should the Cowboys Make a Run at G Carl Nicks?

Jonathan Bales

In the past week or so, I have written extensively on the Cowboys’ offensive line.  In my last post (a look at some interesting offensive statistics from 2011), I hit you with these numbers:

  • Montrae Holland checked in at 20th among all guards in terms of pass protection efficiency, allowing some sort of pressure on 2.5% of pass plays.  This confirms my thought that Holland was very underrated this year. Kyle Kosier was 33rd at 3.2%.  He was just a league-average guard in 2011.
  • Phil Costa was 29th in the NFL among centers with 2.7% pass protection efficiency.  He really shouldn’t start in 2012, although he probably will.

It is pretty clear the interior line is in disarray in Dallas, and something needs to be done to fix it.  I suggested moving Doug Free to right guard and drafting a right tackle in the first round (with Tyron Smith obviously kicking to the left side).  There are some pros and cons to that plan, but I like it because it instantly upgrades two spots.

Others have suggested the Cowboys might make a run at impending free agent guard Carl Nicks, though.  Although Jerry Jones has refrained from signing big-money free agents since Jason Garrett has taken over as head coach, this is one I actually believe the ‘Boys should jump all over.  Here’s why.

Nicks will command a hefty contract, but guards are continually underpaid in the NFL.  He won’t garner nearly as much money as an elite left tackle, but his impact (for Dallas, especially) isn’t that much less than his tackle counterpart.  We saw how much a weak interior line can affect an offense in 2011.  Don’t let it happen again in 2012.

Nicks was the No. 2 ranked guard by Pro Football Focus, yielding only eight pressures all season.  He had the second-highest pass blocking efficiency in the NFL, allowing a sack, hit or pressure on just 1.4% of pass plays.  Nicks is a dominant run blocker as well.  Saints running backs averaged a ridiculous 5.96 yards-per-carry when Nicks was at the point-of-attack this season.  Compare those numbers with the Cowboys’ interior linemen (above).  Dallas backs averaged less than four yards per carry when running behind Holland in 2011, and he’s a player whose run blocking I praised as solid.

Personally, I don’t think the acquisition of Nicks means the team should automatically forget about switching Free’s position.  A tackle-to-guard transition might not seem as appealing with Nicks in town, but an offensive line of Smith, Nicks, Kosier (who can play center), Free, and a rookie right tackle looks pretty damn good to me.  Throw in Holland and Phil Costa as backups, and you’re all set.

Either way, Nicks is a player who the Cowboys should seriously consider.  He will demand a pretty penny, but guards are repeatedly undervalued.  He’s a player on whom to break the bank this offseason.

Like Dallas Cowboys Times on Facebook

Follow Dallas Cowboys Times on Twitter


Dallas Cowboys 2011 Recap: Interesting Offensive Stats

Jonathan Bales

I am going to begin my 2012 Draft coverage early this year, and you can expect it to be superb. . .as per usual.  Between those articles you can also expect to find stat analysis of the Cowboys 2011 season.  Below, I have pasted some interesting numbers from both Pro Football Focus and my own Excel spreadsheets.  Similar defensive statistics to come.

Tony Romo

  • Romo finished the season fourth in the NFL in passer rating at 102.5, behind only Aaron Rodgers, Drew Brees and Tom Brady.  That includes a 104.4 rating in the fourth quarter. . .not bad for a “choke artist.”
  • Taking away drops, spikes and throw aways, Romo’s completion percentage was 73.5%.
  • On deep passes of 20+ yards, Romo completed 54.8% of his attempts.  That was second in the league to Aaron Rodgers, but only 11.9% of Romo’s passes traveled that long–good for only 13th in the league.  He threw 13 touchdowns and only two picks on deep throws.  I’ve been saying for years the Cowboys would benefit immensely from more deep passes.
  • Romo was under pressure on 30.7% of dropbacks, which was 13th highest in league, but completed 56.7% of his passes in these situations.  That was second-best in the NFL to only Drew Brees.

Dez Bryant, Miles Austin and Laurent Robinson

  • Cowboys quarterbacks had a 110.8 passer rating when throwing to Dez Bryant, which was the 16th-highest of any receiver in the NFL.  Romo threw three of his interceptions when targeting Bryant.
  • I have seen some criticisms of Miles Austin lately, even from “expert” Dallas-area writers.  Don’t listen to it.  Austin’s only problem has been staying healthy, as Romo posted a 117.8 rating when throwing to Austin, including zero interceptions.  That rating is good for 11th among all receivers.  Austin is an elite wide receiver who will have a monster 2012 season if he stays on the field.
  • Puzzling to me are Austin’s drops.  After a 2010 season in which he struggled with dropping passes, Austin let four more get through his hands this season.  That isn’t an enormous amount, but it was 8.5% of catchable passes and good for just 37th in the NFL.  I think this is a small sample size at work, though, as just one less drop would shoot Austin up to 23rd.
  • Meanwhile, Bryant tallied only one drop all season–second-best in the NFL of any receiver who played 25% of his team’s snaps.  Only Golden Tate caught every pass possible.
  • Laurent Robinson caught 58.8% of deep passes (20+ yards) thrown his way, good for third in NFL.  Austin was 10th at 50.0%, and Bryant 29th at 36.8%.  These numbers are misleading, as Robinson is very rarely the first read on plays.  If he is thrown to, chances are he’s fairly open.  Bryant gets balls in double-coverage, and so we’d expect his deep catch rate to be lower.  Larry Fitzgerald, for example, was just 24th in the NFL at 41.2%.
  • Robinson tallied 2.18 yards per route–the top number on the Cowboys.

DeMarco Murray and Felix Jones

  • DeMarco Murray and Felix Jones were both solid at avoiding defenders in 2011, tallying 3.01 and 2.98 yards-after-contact/attempt.  Those rates were 10th and 11th in the league.
  • Murray had 36.8% of his yardage come on runs of 15+ yards, which was the 12th-highest rate in the NFL.  Jones was 31st at 26.4%.  Again, this stat can be misleading.  While you always want big plays, a really high “big run rate” can be an indicator that a running back will regress to the mean the following season, rushing for fewer big plays and seeing a decrease in both total yards and yards per attempt.  Murray and Jones are both breakaway players, and I’d expect both of them to be around 35% in any given season.  As an example of how much these numbers can fluctuate, consider that Jones saw 44.0% of his yards come on big plays in 2009, compared to just 15.3% last season.
  • Murray and Jones were 24th and 26th, respectively, in catch rate at 89.7% and 89.2%
  • Murray and Jones both need to improve in pass protection.  Jones allowed a pressure, hit or sack on 6.3% of snaps he was in pass pro.  This was just the 41st-best mark in the NFL.  Murray’s 9.7% number came in at 62nd in the league.

Jason Witten and Martellus Bennett

  • Jason Witten dropped 3.61% of balls thrown his way (three total), good for 10th in the league.
  • 13.5% of Witten’s snaps came in the slot.  That was just the 17th-highest percentage for tight ends, and the rate was well behind the top 10 (all of whom played 25+% snaps in slot).
  • Witten was 12th in yards per route at 1.69.
  • Witten blocked on only 9.4% of pass plays, well below his rate in past seasons.  He was 18th in the NFL with 3.9% of snaps resulting in a pressure, hit or sack.  Martellus Bennett was 17th, with 3.8% of his snaps resulting in some sort of pressure.  It confirms the notion that Witten and Bennett are similar in pass protection (although Bennett is far superior as a run blocker).  Bennett blocked on 20.1% of pass plays.

Offensive Line

  • The entire offensive line was 14th overall in pass blocking efficiency, allowing a pressure, hit or sack on 18.5% of pass plays.
  • Tyron Smith was the league’s 14th most efficient tackle in terms of pass protection, allowing a pressure, hit or sack on just 4.0% of pass plays.  Free was 48th with 6.3%.  He also allowed 10 sacks, which was sixth-worst in the NFL.
  • Montrae Holland checked in at 20th among all guards in terms of pass protection efficiency, allowing some sort of pressure on 2.5% of pass plays.  This confirms my thought that Holland was very underrated this year. Kyle Kosier was 33rd at 3.2%.  He was just a league-average guard in 2011.
  • Phil Costa was 29th in the NFL among centers with 2.7% pass protection efficiency.  He really shouldn’t start in 2012, although he probably will.
  • Bill Nagy allowed pressure on 4.1% of pass plays, good for 41st in the league.

Like Dallas Cowboys Times on Facebook

Follow Dallas Cowboys Times on Twitter


Dallas Cowboys vs. New York Giants, Week 17: How Dallas Can Win the NFC East

Jonathan Bales

In addition to my article for the Times on how Dallas can beat Cover 2 Man Under this weekend in the Meadowlands, I also did a piece for the Dallas Observer.  Head over there to check out my DOs and DON’Ts for Dallas. Along with more analysis of the coverage which irritated Dallas in Week 14, I add a full game plan for the ‘Boys.  Here are some of the highlights:

  • Don’t blitz often.  Eli Manning’s passer rating against the blitz is very comparable to that when four or less defenders rush him, but the Cowboys do not have the talent in the secondary to deal with a blitz that fails.  The team should be in the business of playing aggressively while still allowing for a chance to win the game late, and yielding quick scores due to unsuccessful blitz attempts won’t help.
  • The ‘Boys should mimic the Giants’ Week 14 game plan by playing a lot of Cover 2 Man Under.  By keeping everything in front of them, the defense can maximize their chances of halting Hakeem Nicks and Victor Cruz and force either a tight end or a running back to beat them.  Although Brandon Jacobs wore down Dallas in the teams’ last meeting, Ahmad Bradshaw is the more likely of the two to give Dallas fits this week.
  • The Giants pass a lot out of double-tight formations, so the Cowboys cannot sell out to defend the run when they see the look.  The G-Men used a double-tight set 34 times in Week 14, so the ‘Boys better be ready for it.
  • The Cowboys, on the other hand, do tip their play calls via their formation, personnel package, or down-and-distance.  Jason Garrett could benefit from being a bit less predictable this week.  Garrett’s predictability could be utilized to get the ball downfield with play action. But since 2009, Dallas quarterbacks have thrown for 20-plus yards on only 8.7 percent of play-action passes. And in two-plus years of passes, Garrett has called a play-action pass only eight times with 1-4 yards-to-go for a first down — the situations when faking a run would actually work. Instead, he’s called for a play-action look on 11 plays with 20-plus yards-to-go, when showing a running play is either an obvious decoy or hopeless.

For additional analysis, head over to the Observer and leave your comments there.

Like Dallas Cowboys Times on Facebook

Follow Dallas Cowboys Times on Twitter



Assessing Pivotal Plays in Cowboys-Giants Week 14 Matchup

Jonathan Bales

One of my favorite things to do when I watch the Cowboys’ games at home is track the team’s win probability throughout the night using Advanced NFL Stats‘ win probability charts.  Using a model which takes the down-and-distance, score, and time into account, ANS is able to determine the probability of a team winning a game at any point in time.  This information doesn’t stem from estimates, but rather years of NFL data.

It is always fascinating to see how certain plays can influence a team’s chances of winning.  Punts, for example, often result in a fairly significant drop in win probability because giving away possession is generally detrimental to a team.  Near the beginning of games, it takes a huge play to swing win probability in a major way.  A 4th and Goal defensive stop while up six points with 45 minutes to play might result in a big bump in win probability, but that same play would be much larger–perhaps from around 50% to 100%–if the play was the final one of the game.

Using the graph from Sunday night’s game, I thought it would be fun to take a look at which plays affected the Cowboys’ win probability most significantly.  Below, you can see the chart, along with 10 plays (or short sequences) which I have labeled as the most important. . .


Play 1: 64-yard pass to Hakeem Nicks on 3rd and 7 at NYG 32; 53:44 to play

  • Cowboys’ WP (win probability) drops from 42 to 25 (-17%)

A 17% drop in win probability in the middle of the first quarter is a big one.  This wouldn’t have been much higher even if Nicks scored.  Poor coverage by Alan Ball.

Play 2: 26-yard gain by Felix Jones on 1st and 10 at NYG 42; 48:19 to play

  • Cowboys’ WP jumps from 42 to 52 (+10%)

The touchdown pass to John Phillips put Dallas on top, but Jones and a subsequent defensive holding penalty put the ‘Boys in position to score.

Play 3: Felix Jones fumbles on 1st and 10 and ball recovered by NYG at DAL 14; 31:38 to play

  • Cowboys’ WP drops from 57 to 38 (-19%)

This had the potential to be devastating to Dallas but they made the best of the situation by holding New York to a field goal and kicking one of their own before halftime.

Play 4: 47-yard touchdown pass to Mario Manningham on 3rd and 5 at DAL 47; 19:39 to play

  • Cowboys’ WP drops from 70 to 45 (-25%)

Yielding a 47-yard score on a crucial 3rd down due to a broken coverage is heartbreaking.

Play 5: 74-yard pass to Laurent Robinson on 3rd and 10 at DAL 20; 13:17 to play

  • Cowboys’ WP jumps from 38 to 66 (+28%)

Other than the blocked field goal to end the game, this pass to Robinson was the most important one of the contest for Dallas.

Play 6: 15-yard completion to Mario Manningham on 4th and 3 at DAL 37; 8:24 to play

  • Cowboys’ WP drops from 78 to 56 (-22%)

Here, you can see how game situation affects win probability.  The pass was only 15 yards long, but it came on a crucial 4th and 3 with just over eight minutes left to play.

Play 7: Sean Lee interception on 3rd and 9 at DAL 21; 6:50 to play

  • Cowboys’ WP jumps from 62 to 89 (+27%)

I actually thought this would be more valuable to Dallas, but the fact that it came on a difficult 3rd and 9 (when New York’s chances of converting were low) likely affected the jump in WP.

Play 8: Cowboys’ three-and-out; 2:20 to play

  • Cowboys’ WP drops from 88 to 67 (-21%)

Romo’s infamous incompletion to Austin hurt Dallas in a big way.  If you assume Romo hits that pass 90% of the time and Dallas’ wins 99% of games following a completion, the actual dip in WP would be closer to -31%.

Play 9: Holding on Abram Elam and 18-yard completion to Jake Ballard on 1st and 10 at DAL 19; 1:21 to play

  • Cowboys’ WP drops from 49 to 27 (-22%)

The holding penalty on Elam has been overlooked.  DeMarcus Ware’s offside penalty was also costly, but the full extent of it isn’t factored into the WP chart because the errant snap and loss by the Giants isn’t reflected in the play-by-play.

Play 10: Blocked FG; 0:06 to play

  • Cowboys’ WP drops from 44 to <1 (-43%)

And the Cowboys’ playoff chances drop from potentially around 90% with a win to now around 40% .

Like Dallas Cowboys Times on Facebook

Follow Dallas Cowboys Times on Twitter



Dallas Cowboys vs. New York Giants, Week 14: How a Loss Affects Each Team’s Playoff Hopes

Jonathan Bales

For those who missed it (everyone, I’m assuming), I had an article published on The New York Times’ football blog The Fifth Down titled Keys to Defeating Dallas, From a Cowboys Writer.  You can (and should) click on the link to read it.

Before I receive the inevitable backlash for writing such blasphemous content, let me point out that this article is not much different than my traditional “DOs and DON’Ts for Dallas” weekly post.  Instead of explaining what Dallas needs to do for a win, however, I simply flipped the script.  “Protect Doug Free” would have been sound advice for Dallas this week, and you will find a corresponding “Attack Doug Free” bullet point in my New York Times piece.

You can find all of my thoughts on the Week 14 matchup in that post.  Here, I wanted to take a look at the broader picture, assessing both the Cowboys’ and Giants’ odds of making the playoffs after the contest.  There are a number of ways to do that.  In this article, I will conduct a schedule analysis, estimating the odds of each team winning their remaining games and determining how that relates to the subsequent importance of this particular matchup for each squad.

In the beginning of the season, the importance of the strength of a team’s schedule is very overblown, and for a variety of reasons.  First, when comparing the schedules of two division opponents, the 14 games which are not head-to-head contain just two dissimilarities, i.e. the Giants and Cowboys had just two opponents which differed.  Secondly, the constant talent flux in the NFL makes preseason predictions in regards to a team’s strength of schedule almost useless.  Who wold have thought the Raiders, Bengals, Bills, Lions, etc. would not be “easy wins” in 2011?  Lastly, the overall strength of a schedule tends to even out over the course of a 16-game season.  Simply put, preseason strength of schedule sucks as a method by which to analyze a team’s playoff hopes.

By Week 14, however, a team’s remaining schedule has a huge influence in their odds of making the playoffs.  Over the course of four games, it is certainly possible to have three or four strong/weak opponents, meaning the schedule doesn’t necessarily have time to “even out.”  Plus, we already know the relative strength of each NFL team, so estimating each team’s win probability for a single game is far easier.

With all of that said, let’s take a look at the Giants’ and Cowboys’ remaining schedules, along with the probability they win each game. . .

While the probabilities I used are “just” estimates, they are well-researched, solid estimates using advanced statistics gathered from Advanced NFL Stats.  Using their Generic Win Probability (the odds a team will win a game against a league-average opponent at a neutral site), we can estimate the chances of Dallas and New York winning each remaining game by combining the teams’ GWP and factoring in home field advantage.  Thus, while I am “guessing” the odds of wins for the teams in their remaining contests, those probabilities are likely not too far from reality.

If we give each team a win total that directly correlates with their probabilities, i.e. the Cowboys’ 54% chance of winning this week’s matchup equates to “.54 wins,” you can see the Cowboys’ theoretical win total for the rest of the 2011 season is 2.37, while the Giants’ is 2.40.  That difference is obviously not statistically significant, meaning the small errors which are undoubtedly contained within my win probability projections are surely enough to “make up for” this difference.

In other words, the Cowboys and Giants can be expected to win the same number of games moving forward.  Note that this conclusion does not mean the teams will win the same number of games, but rather that their chances of doing so are the most likely of all possible outcomes.

So what does this all mean for their playoff chances?  The incredible similarity between the teams and their win probability moving forward means that each’s respective chances of making the playoffs, even after we factor in the schedule, are roughly the same as if they were both of league-average quality.

Think of it this way: if a team of the same quality of the Colts was 7-5 (like Dallas) and one with a talent level comparable to that of Green Bay was 6-6 (like New York), we’d still expect Green Bay’s odds of making the playoffs to be greater due to their far higher win probability in each game.  The role of chance in that situation is less powerful than in the case of the ‘Boys and G-Men.

In the latter comparison, we have two teams whose win probabilities for the remainder of the season are roughly equal.  This means that the chances of the Giants making the playoffs, taking their one-game deficit into account, are roughly the same as if we flipped a coin to determine the outcome of each contest.  Perform 10,000 coin flip simulations to allow the numbers to regress to the mean, and you will be left with each team’s playoff chances.

Luckily, we have computers which can simulate seasons, so we don’t need to waste time flipping a coin.  Over at NFL-Forecast.com, they have determined each team’s chances of making the playoffs based on thousands of simulations with the GWP provided by Advanced NFL Stats.  If you head over there, you will see the odds of Dallas winning the NFC East are around 64%, with their overall playoff chances adding up to nearly 69% (meaning there is little chance they earn a Wild Card spot).  The Giants’ odds are 35% to win the division and 36% to make the playoffs.

In terms of particular scenarios, the Cowboys will win the division if they beat the Giants in both of the clubs’ remaining games.  If the Giants perform the same task, they will win the division as long as they do not lose both of their other two games or the Cowboys do not win both of their other games.  Simply put, if one squad wins both head-to-head games, they are effectively the NFC East champs.

If the more probable splitting of head-to-head games occurs, the Cowboys are in with one more win.  Let me break it down for you.  If the teams split and the Cowboys lose to Philly and beat Tampa Bay, for example, they would finish 9-7 and possess the same record as the Giants if they beat both the Redskins and Jets. Then, intricate tiebreakers would come into play.  As shown on NFL.com, those are:

  1. Head-to-head (best won-lost-tied percentage in games between the clubs).
  2. Best won-lost-tied percentage in games played within the division.
  3. Best won-lost-tied percentage in common games.
  4. Best won-lost-tied percentage in games played within the conference

The first tiebreaker would obviously end in a tie following a split.  If the Cowboys lose to either the Eagles and Bucs (and the Giants win against the ‘Skins and Jets), the teams would finish 9-7 and both have a division record of 3-3, meaning the third tiebreaker would be a factor.  In common games (all but two, remember), the Giants and Cowboys would again be tied at 9-5 (including their head-to-head split).  Thus, the fourth tiebreaker is needed.  Here, the Cowboys have the advantage, as they will have a superior conference record following a split, no matter how the other games shake out.

Let’s recap. . .

  • If the Cowboys sweep the head-to-head matchups, they are in the playoffs.
  • If the Giants sweep the head-to-head matchups, they are almost assuredly in the playoffs.  They would miss out if they lose both remaining games and the Cowboys win both.
  • If the teams split, the Cowboys are in if they win one of their other games.  If the Cowboys lose both of those games, the Giants are in if they win both of their remaining contests.

For Dallas, all of the math equates to one simple conclusion: win this week, and you will almost certainly be NFC East champs.

Like Dallas Cowboys Times on Facebook

Follow Dallas Cowboys Times on Twitter


Protected: More on Jason Garrett’s Late-Game Clock Management

This content is password protected. To view it please enter your password below:


Tony Romo Directional Passing: Why Dallas Should Throw Deep More Often

Jonathan Bales

I recently assessed Tony Romo’s direction-based passing thus far in 2011. . .

You can see Romo has been a bit superior when throwing to the left and middle portions of the field, but not enough to draw any statistically significant conclusions.  Equally effective passing to all portions of the field has been Romo’s M.O. over the course of his career, as he has never really racked up a majority of his touchdowns, completions, or yards in any given section of the field.

The fact that his efficiency has been relatively equal all over the field is somewhat unusual, as most quarterbacks are more accurate when throwing to the right side of the field (right-handed quarterbacks, anyway).  I suppose Romo’s lack of increased success when throwing to his right is a byproduct of his game–one which emphasizes buying time for receivers to get open as opposed to extremely accurate throws.  As I have explained over the years, I would rate Romo in the bottom half of the league in terms of pure accuracy, but his completion percentage is always good because he has the ability to buy time in the pocket.

Ultimately, it doesn’t really matter how you complete passes.  Romo’s game is not any better or worse than another style of play, but it is likely the reason we see such an even distribution of passing efficiency, regardless of field sub-section.

Perhaps more interesting than Romo’s directional passing are his numbers when broken down by pass length.  According to Pro Football Focus, Romo has racked up a 122.8 passer rating when throwing 20+ yards in 2011.  This includes a 58.1% completion percentage, 11 touchdowns, two interceptions, and a ridiculous 20.2 yards-per-attempt (yes attempt, not completion).

Let me start by acknowledging there is likely somewhat of a selection bias at play here.  Romo is far more likely to hold onto the football when a deep pass isn’t open as compared to a short or intermediate route.  Plus, no throw-aways are traveling 20+ yards, so his numbers on deep passes aren’t hindered by “give up” plays.

Nonetheless, Romo’s success on deep passes has been a trend over the years.  I have detailed twice (here and here) how important deep passes can be to an offense, along with why the Cowboys need to attempt far more of them.  In those articles, I noted quarterbacks as a whole seem to have more success when they throw deep more frequently. . .

You can see interception rates actually decrease for quarterbacks who throw a lot of deep balls, while both average yards-per-attempt and success rate increase.  We find similar numbers in 2011.

So why don’t NFL offenses air it out more?  As Brian Burke explained earlier today, NFL coaches are risk averse:

A risk-averse mindset is echoed in an old football saying attributed to Texas coach Darrell Royal and still often repeated: “Three things can happen when you throw the football, and two of them are bad.” Coaches tend to classify outcomes as good things and bad things, and then count them up. I’m sure coaches obviously know that interceptions are a lot worse than an incompletion or a one-yard stuff. But they don’t accurately account for the payoff and frequency of each possible outcome, and this is reflected in how infrequently they call for deep bombs.

In addition to all of the stats we can measure from deep passes, offenses gain an even greater advantage due to defensive penalties on these throws.  As Burke points out, even if we list every sack in the league as being the result of a deep pass attempt, that advantage still exists.  NFL teams need to throw deep far, far more often than the current rate.  That seems particularly true for a Cowboys squad with a play-making quarterback and three legitimate deep threats at receiver.

Like Dallas Cowboys Times on Facebook

Follow Dallas Cowboys Times on Twitter