If you look at NFL futures odds, you’ll see that the Cowboys are currently +275 to win the NFC East. That’s the same as the Giants, but significantly worse than the Eagles. One of the reasons Dallas is a team moving in a different direction than Philly is that they ignore the numbers, handing out “bad money” contracts like Witten’s five-year, $37 million deal that doesn’t expire until 2018.
So why is Witten no longer elite? From my B/R article:
Yards Per Target
The first way that we can measure Witten’s efficiency is by looking at what he does with his targets. Here’s a look at the progression of his yards per target. I marked the NFL league average for tight ends with white dash marks.
You can see that Witten was continually around the league average (8.0) in efficiency prior to 2012. That’s not horrible, considering that he always saw more targets than most tight ends and as workload increases, efficiency naturally decreases.
The fact that Witten was around average in efficiency prior to 2012 shows he was an above-average tight end, but probably not as efficient as you might have expected. The truth is that Witten has always been a really good player, but one who’s really been aided by bulk attempts.
Would we have considered Witten a monster receiving tight end if he had played on an offense that didn’t focus on him so heavily? Doubtful.
Again, that’s pre-2012. Look at what happened two seasons ago. In the year in which Witten broke the record for receptions by a tight end, he also turned in a career low in yards per target. And the drop wasn’t a small one; it was magnificent.
In his “career year” of 2012, Witten produced more than one yard less per target than an average NFL tight end. Is it now easier to see why predicting him to disappoint in 2013 was so easy?
Witten’s yards per target improved in 2013 with fewer looks, but his average was still the second-worst and below-average for a tight end. It’s difficult to label a player who is below-average in efficiency as “elite,” especially when they receive the fewest targets they’ve gotten in seven seasons.
Yards Per Route
Of all stats to analyze for receivers, my favorite is yards per route. That’s because it penalizes for not getting open (and thus failing to receive a target).
It’s also good for tight ends because it doesn’t penalize them when they stay in to block. One of the oft-overlooked aspects of Witten’s game is that he doesn’t stay in to block on very many passes.
Last year, he stayed in to block 13.6 percent of the time, according toPro Football Focus (subscription required), which ranked 28th of 35 qualifying tight ends. When the Cowboys pass, Witten is almost always an option for Romo, which will naturally inflate his stats. Yards per route adjusts for that.
Plus, yards per route does a better job than yards per target of displaying Witten’s value.
Even though Witten’s yards per route has declined every year since 2008, he was still well above the league average up until 2013. Part of that is due to receiving a high percentage of the Cowboys’ targets—again, the bulk opportunities obviously help—but Witten was also producing at a higher level a few years ago.
The fact that Witten’s yards per route declined in 2012, when he saw an unfathomable 150 targets, is a really strong sign that his play has been regressing. To produce just above league-average efficiency on a per-route basis despite seeing a target on 23.8 percent of his routes (a high rate) is a good indicator that Witten wasn’t actually playing at a level that his raw totals suggested.
The go-to rebuttal for Witten apologists is “but he’s an elite blocker, too.” That’s probably the case because it’s difficult to quantify blocking ability, and thus difficult to refute. Those who use Witten’s blocking to claim he’s still an elite tight end will tell you to “just turn on the tape.”
Well, Pro Football Focus does turn on the tape, but unlike traditional scouts, it quantifies what it sees. We can use its numbers on tight end blocking efficiency, at least in pass protection, to determine Witten’s value as a pass-blocker.
Here’s a look at Witten’s pass-blocking efficiency since 2007 versus the average NFL tight end.
Witten has adequately blocked his man on 95.1 percent of pass snaps. That’s better than the league average, although the effect perhapsisn’t as great as it appears; the average NFL tight end has pass-blocking efficiency just one percentage point lower at 94.1 percent.
That means we can expect Witten to allow one less pressure on Romo for every 100 pass snaps that he stays in to block as compared to a typical tight end.
But remember, Witten doesn’t remain in to block all that much—an average of just over 61 snaps per year over the past eight seasons. That’s it. So on average, Witten has “saved” 0.61 quarterback pressures per season for Dallas.
Now, can he really overcome a precipitous drop in receiving efficiency to remain an “elite” tight end because of 0.61 pressures per year?
Witten is a good tight end. He’s a great guy, a hard worker, and the type of player you want on your team.
He’s not an elite tight end.
If you’re going to continue to label the future Hall of Famer in that way, you need to somehow show that, despite being below-average in yards per target and yards per route in 2013, Witten still somehow offers enough value to overcome his lackluster efficiency. Presumably, those traits will need to be unquantifiable, like “he plays with such heart” or “he’s a good leader.”
OK, Witten is a good leader. He’s a good leader who produces like an average NFL tight end and is by no means an elite option for the Cowboys anymore.
1. The Cowboys will free up enough cap space to sign one major free agent.
Dallas doesn’t have the best salary cap situation, but it also isn’t as bad as it initially appears. By doing everything it needs to do to trim away the extra fat, Dallas should be able to put itself in a position to sign a fairly high-priced free agent.
2. That player will be defensive tackle Henry Melton.
If the Cowboys do indeed sign a big-name player, Melton is perhaps the most likely. He fits with what Dallas wants in a 4-3 defensive tackle. You could argue that the Cowboys will want to wait to see what happens in the draft, but they really need two defensive tackles. Since signing Melton wouldn’t stop Dallas from drafting someone like Pitt’s Aaron Donald in the first round, the move makes sense no matter what the Cowboys want to do in the draft.
3. DeMarcus Ware will remain in Dallas.
The Cowboys have all the leverage and Ware knows it; he will probably take a pay cut to stay in Dallas.
4. Dallas won’t sign a defensive end.
Assuming Ware remains in Dallas, it’ll have him and George Selvie to start in 2014. There are a bunch of intriguing second-round defensive ends in the draft, and don’t forget that Tyrone Crawford has the versatility to kick outside. Since ends usually cost a pretty penny in free agency, it makes sense for Dallas to keep what it has and upgrade via the draft.
5. The Cowboys won’t re-sign a single one of their free agents.
This really comes down to whether or not Dallas will keep either Jason Hatcher or Anthony Spencer. Hopefully, the Cowboys have learned their lesson from handing out sizable contracts to aging players.
I also explained why I’m “buying” or “selling” certain rumors in Dallas:
The ‘Boys Should Sign a Free Agent RB
There’s perhaps no worse proposition in all of football than signing a veteran running back, but that hasn’t stopped many from speculating that the Cowboys could bring one in. Dallas Morning News listed Ben Tate, Rashad Jennings, and Donald Brown as possibilities.
The problem with running backs is that, outside of a few special talents like Adrian Peterson and Jamaal Charles, they’re extremely replaceable. Part of the reason for that is because they’re so dependent on their offensive lines for production. If the majority of running back success is due to factors outside of his control, why pay for one?
Another reason is that running backs enter the league at near-peak efficiency. If the Cowboys want another running back, which isn’t a bad idea with DeMarco Murray’s contract set to expire after the 2014 season, they should spend a mid-round pick on one. Just not a back who runs a 4.63 40-yard dash at 198 pounds.
And finally, a look at why I’d trade for Dolphins DE Dion Jordan:
Putting on the Pressure
Jordan’s rookie season in Miami was widely considered a bust because he recorded only two sacks. However, considering he rushed the passer only 206 times, according to Pro Football Focus (subscription required) and was able to pressure the quarterback at an elite rate, Jordan’s first year wasn’t as bad as people believe.
Looking at pressure rate, which is the percentage of pass-rushing snaps on which a player hurries the quarterback, we see Jordan was actually really good.
Jordan didn’t play as many snaps as the other rushers, but the fact that he recorded a higher pressure rate than Greg Hardy, widely considered one of the top young pass-rushers in the NFL, shows you something.
So why only two sacks? Well, Jordan got unlucky. There’s good evidence to show that most pass-rushers bring down the quarterback on right around 25 percent of their pressures; that is, for every four times a pass-rusher hurries the quarterback, he typically records one sack. Getting to the passer is a skill, but obtaining a sack once you’re already there is a much more random occurrence.
If you examine Jordan’s sack-to-pressure ratio during his rookie year, you see it’s quite low.
You always want sacks, of course, but pressures are even more important than sacks when predicting future sacks. The fact that Jordan was able to reach the quarterback suggests he’s going to generate plenty of sacks in the NFL but was just unlucky in his rookie year. Based on his pressures alone, his most likely sack total was 4.5, not 2.0.
What About the Money?
One possible concern about trading for Jordan is his contract. The Cowboys don’t have very much cap space with which to work.
However, even as the No. 3 overall pick in 2013, Jordan’s contract is far from prohibitive. Over the Cap has his 2014 cap hit at only $4.7 million, with $16.8 million guaranteed remaining on his deal. In comparison, right tackle Doug Free’s 2014 cap number is $6.5 million.
Plus, the “real” cost of Jordan is his contract minus whatever the Cowboys would need to pay their first-round pick in 2014, since they’d move that selection in order to acquire Jordan.
The Dallas Cowboys are clearly in the market for a difference-making defensive end; veteran DeMarcus Ware is rapidly declining and 2013 breakout player George Selvie is a No. 2 rusher at best. In addition to defensive tackle, there’s perhaps nowhere the ‘Boys can help themselves more than at defensive end.
The key to finding value at any position is identifying predictors of success that other teams aren’t valuing, or at least aren’t valuing enough. At the wide receiver position, for example, most NFL teams pay for speed when they should be more concerned with size. At quarterback, they seek height when it appears as though hand size is more important.
At defensive end, a lot of teams seem to dismiss college production. They think they can spot talent with the “eye test,” frequently overrating defensive ends who look the part but haven’t gotten to the quarterback at a high rate.
One of the traits teams overvalue seems to be quickness. Certainly quicker is better for every player, but even more important is size. Overall, teams do indeed pay for one aspect of size in height. Height is indeed correlated with NFL success for defensive ends, but as with quarterbacks, that might just be because it’s linked to another trait that matters more. For defensive ends, that characteristic is long arms.
Even though many NFL organizations look at arm length, they’re still acting as though height matters more. We continually see tall players with short arms get drafted ahead of short players with long arms. The latter type of prospect might actually offer the most value, since they possess the trait that helps get to the passer (long arms) but not the one that shoots them up boards (height).
In that way, not all lengths of arm are created equally. A 6’2″ defensive end with 33-inch arms is better than a 6’5″ end with the same arm length because the shorter player is more likely to fall (unnecessarily), and thus offer value.
Valuing both college production and arm length, here are five defensive end prospects the Dallas Cowboys should consider in this year’s draft.
Scott Crichton, Oregon State
College Production: 22.5 sacks, 51 tackles-for-loss in three seasons
Arm Length: Unknown
Projection: Early Second Round
Oregon State defensive end Scott Crichton is a really interesting prospect. At 6’3″, Crichton is rumored to have good length for his height. That’s the perfect combination when seeking value at defensive end. It will be really interesting to see how Crichton measures up at the combine; if his arms check in at 33 inches or longer, that’s a great sign.
Crichton was also quite productive in college. Teams will obviously look at sack totals, but tackles-for-loss are just as valuable. They represent the same sort of explosiveness to get into the backfield. Actually, when a defensive end has a ton of tackles-for-loss but a moderate amount of sacks, it could be a sign that he just got unlucky with bringing down the passer, which will improve in the future.
The Cowboys would probably need to pick up Crichton in a trade down from the No. 16 overall pick.
Jackson Jeffcoat, Texas
College Production: 26 sacks, 49.5 tackles-for-loss in four seasons
Arm Length: Unknown
Projection: Second Round
Like Crichton, Texas’s Jackson Jeffcoat is rumored to have “significant arm length.” He won’t drop because of his height, which is about average for a defensive end, but he still seems undervalued with a second-round projection.
Jeffcoat had 26 sacks in four years at Texas, but he was quite dominant recently. He played only a handful of games in 2012, but in his last two full seasons, Jeffcoat averaged 18 tackles-for-loss and 10.25 sacks.
The Cowboys might be able to pick up Jeffcoat in the second round without moving up.
Head to BR for the full article.]]>
Exploiting the Chart
Because teams still use the original trade chart to make moves, the Cowboys can exploit inefficiencies in the way the chart is constructed relative to how players actually perform. That is, they can and should look for areas on the chart where the cost of trading is less than the value of they player(s) they can expect in return.
Looking at historic NFL production, here’s a chart displaying approximate value versus the cost of each pick, both as a function of the overall value. According to the chart, for example, the top overall pick accounts for five percent of the total value, but in reality, the No. 1 overall selection has accrued around just 2.5 percent of the class’s total approximate value.
The cost of the draft’s top picks is massive. Teams trading up into the top 10, especially, need to be extremely confident they’re acquiring a future star, because the cost of the picks is far more than the production that can normally be expected.
You can see that the cost of trading surpasses the expected production up until around the 20th overall selection. At that point, the cost of draft selections on the original trade value chart drops quite a bit, yet there are still quality players in that range.
That trend continues for a few rounds; in terms of historic production, the largest inefficiencies exist from the back of the first round into the third and fourth rounds. That’s the area where the most value exists when you consider the price of each pick.
First-Round Trade Results
This isn’t just a mathematical trick that has little basis in reality. We can look at past first-round trades to see that teams trading down in the round (or out of it altogether) have found a ton of success.
Teams trading down in or out of the first round have accrued the best player 50.9 percent of the time. That’s amazing since the reason a team trades up is because they think they’ll acquire the best player. The numbers suggest that if you believe you’ll get the best player by trading up in the first round, think again.
You might think that because trying to guess the team that will get the top player in a first-round trade is basically a coin flip, it doesn’t matter if you move up or down. However, if it’s really a coin flip, the smart things to do would be to 1) maximize opportunities and 2) do it as cheaply as possible. Why trade up to get something you can probably have later?
Further, teams trading down have acquired 64.4 percent of the total player value in those trades. The numbers are staggering.
Loading Up and Accounting for Fallibility
Again, the Cowboys should be trying to maximize their chances of hitting by loading up on picks where the greatest inefficiencies exist. They actually tried to pull this off last year, trading back from the middle of the first to the back of it, grabbing an extra third-rounder in the process. Whether or not they saw the proper return, they at least tried to make a move that was statistically prudent.
Central to the idea that you should accrue draft picks is understanding your own fallibility. Teams trade up as though they’re sure the player they covet is going to be as good as they believe, but it’s not that black-and-white. There’s a lot of uncertainty that goes into drafting, so teams need to account for that.
Further, teams that trade up are typically selecting a player who is an outlier on their board. If that team didn’t have him ranked much higher than others, he probably wouldn’t still be available. By accounting for the fact that they could be wrong in their assessment of said player, the Cowboys and other teams considering trading up might think again.
The common rebuttal is that every situation needs to be taken on a case-by-case basis and that every player is unique, but that line of thinking is probably what’s gotten teams who trade up in trouble. Every player and situation is unique, sure, but teams drastically overrate their ability to predict the future.
To get the best in a trade-up, teams basically need to think they know something that 31 other teams don’t, which will rarely be the case. Once teams accept that they just don’t know as much as they think they do, it will become easier to recognize the dangers of trading up and the value of maximizing draft picks.
Put another way, consider two dart throwers, one with zero skill and one with perfect skill. The first has a low probability of hitting a bulleye on a given throw, so he wants as many darts as possible. The latter can hit a bullseye every time, so if that’s the goal, he needs only one dart.
NFL teams have been acting as though they’re perfect (or near-perfect) dart throwers. When they realize that there’s just a whole lot more uncertainty involved in the process – when they realize the market itself forces them into being low-skill dart throwers – they’ll act differently, maximizing opportunities and minimizing cost.
Read the full analysis right here.]]>
The Dallas Cowboys could benefit from selecting a potential No. 1 wide receiver in the 2014 NFL draft.
That’s not a popular opinion, as wide receiver is considered a position of relative strength for Dallas, but the Cowboys could be in monumental trouble if Dez Bryant gets injured, or even if defenses commit more to taking him out of games.
Do the Cowboys have positions at which an upgrade would provide more immediate value? Yes. But outside of quarterback Tony Romo, there’s no loss the Cowboys could afford less than that of Bryant; with the current roster, his absence would result in the death of all offensive efficiency.
He’s that important.
Furthermore, wide receivers are undervalued as a whole. As quarterbacks release the ball more quickly in today’s NFL, the importance of the offensive line decreases. Meanwhile, the same wide receiver types dominate and help quarterbacks as much as an elite quarterback can help his pass-catchers. That “type” is big, physical and efficient in the red zone.
Without Bryant, the Cowboys would be in trouble as they approach the goal line. They need to convert yards into points, and players like him do that.
One reason the Cowboys should target a wide receiver in this year’s class is that there are a number of potential No. 1 wide receivers—meaning someone is probably going to drop too far.
Second, NFL teams are valuing the wrong traits in wide receivers, paying too much for speed and not enough for size, youth and college production. That creates marketplace inefficiencies.
I’ve done a lot of research on the importance of height and weight for wide receivers, and there’s a wealth of data that shows they’re more vital than top-end speed.
Everyone agrees that younger is better for any prospect, but NFL teams don’t act like it. Younger players are superior not only because they can play in the league longer, but because they were forced to play older college competition. When a 19-year-old receiver dominates in the SEC, for example, that’s much more impressive than when a 23-year-old does it.
The Cowboys actually fell for this trap last year with their selection of Terrance Williams. The pick wasn’t necessarily a poor one because he still fell pretty far into the third round, but he played much of his rookie season at age 24. In comparison, left tackle Tyron Smith, who has been in the NFL for three seasons, won’t turn 24 until December.
Finally, college production is important for wide receivers. If a guy is going to excel in the pros, chances are that he did it in college. But instead of analyzing bulk receiving stats, it seems more valuable to look at market-share stats, which better predict NFL success. Popularized by rotoViz, market share is the percentage of a team’s receiving yards/touchdowns for which a player accounts.
If a wide receiver has 10 touchdowns on an offense with 30 total receiving touchdowns, his market share of touchdowns (0.33) would be the same as if he had 15 touchdowns on a team with 45 total scores.
Market-share stats are useful for a number of reasons, but one of the most important is that they account for team/quarterback strength. It’s more impressive when a wide receiver dominates on an otherwise poor offense with a lackluster quarterback than when one excels with an elite passer.
So without further ado, here are five wide receivers the Cowboys should consider in 2014. All have an elite combination of size, age and college production.
Allen Robinson, Penn State
Size: 6’3”, 210 pounds
Penn State wide receiver Allen Robinson might be my favorite player in the entire class. With all of the talk about some of the other players at his position, he is the big, young receiver whom no one is mentioning.
You’d like him to be slightly heavier, but he’s far from undersized at 210 pounds, and he has the frame to add some bulk. His final market-share yardage number was through the roof since he had 1,432 yards on 97 catches, but he scored just six times. That might be a concern if he hadn’t scored 11 times in his age-19 season in 2012.
My praise of Robinson is probably the largest you’ll see because I think he’s a less-hyped version of A.J. Green and an inevitable stud at the next level. Dallas would be lucky to land him in Round 2.
Jordan Matthews, Vanderbilt
Size: 6’3”, 209 pounds
At 6’3”, 209 pounds, Vanderbilt’s Jordan Matthews has a very similar build to Allen Robinson. He also had 112 catches during the 2013 season at age 21. His market-share numbers for both yards and touchdowns are through the roof.
Even though he is perhaps an even better option than Robinson, he has a much better chance to fall to Dallas in the second round. Because of how inefficiently NFL teams draft wide receivers, there’s even a chance that the Cowboys could trade back into the late second or early third and still grab Matthews.]]>
How much value does Bailey have over a replacement kicker?
For now, I’m going to bypass the topics of randomness and kicker consistency, instead focusing on how valuable Bailey has been in the past. For the record, giving players contracts based on past play is a quick path to finding yourself in salary cap trouble, as the Cowboys are learning right now. Players need to be paid based on projected future production.
In any event, Bailey has made 89 of his 98 career field goals, which amounts to 90.8 percent. That’s a really high number. Below, I broke down Bailey’s accuracy from different ranges, comparing it to the league average.
Bailey has been more accurate than the average NFL kicker in most areas of the field, particularly from mid-range (38 to 52 yards). His reliability in that area has been really important to Dallas.
It might seem difficult to judge Bailey’s value to the Cowboys in a very concrete way, but we can actually do it fairly easily using expected points. On every field-goal attempt, the Cowboys have a certain expectation of how many points they’ll score over the long run. If your kicker will make a particular kick 50 percent of the time, for example, that expectation is 1.5 points (three field-goal points multiplied by the chances of making it).
By comparing Bailey’s historic production in each range with the league average, we can get a really accurate idea of how many more points he’s been worth to Dallas than a league-average replacement kicker. Here’s how those numbers play out.
During his entire time in Dallas, Bailey has scored 267 points via field goals. Based on the length of those attempts and the average rate of accuracy in each range, the Cowboys could have expected just more than 244 points from a replacement kicker.
That means Bailey has been worth 23 points above average over a three-year period. Playing 48 games, that’s 0.48 points per game, on average.
So the question is whether a mean of 0.48 points per game is worth $23 million over seven years and $7.5 million in guarantees.
Could the Cowboys have spent that money in a smarter way?
That’s a difficult question to completely quantify, but I think the answer is yes. The reason for that comes in the next two questions regarding kicker consistency and sample sizes.
How likely is Bailey to repeat his past stats?
In 2007, the Cowboys drafted kicker Nick Folk in the sixth round. Folks in Dallas thought Folk was the answer at kicker after he made 49 of his initial 56 field-goal attempts, including a number of “clutch” kicks.
After connecting on 87.5 percent of his kicks over two-plus seasons, though, Folk just lost it. He missed an incredible seven of his final 11 field-goal tries in Dallas, and the Cowboys eventually sent him packing.
Folk’s story isn’t that unusual for kickers, whose play is filled with game-to-game and season-to-season variance. Advanced NFL Stats has found that there’s actually a negative correlation between a kicker’s field goals from one year to the next, meaning there’s really no viable way to predict performance; a kicker’s past performance cannot help us predict his future.
That’s scary as it relates to Bailey. Humans are wired to detect patterns in data when they don’t actually exist, so we’re predisposed to believe that Bailey’s past accuracy will be indicative of his future play.
It very well might be, but there’s no way for us to know that. He might continue his streaky play (although long-term field-goal accuracy of more than 90 percent is very unlikely), or he might be the next Folk. We don’t know.
How confident can we be that Bailey’s accuracy is a reflection of his talent?
Related to the topic of kicker consistency is the idea that we can’t necessarily trust Bailey’s past numbers because it’s just a really small sample size. Again, his accuracy might be a close approximation of his true talent, but it also might not be. We haven’t seen enough kicks to really tell the difference at this point.
You might argue that 98 field-goal attempts is a pretty hefty sample size, but that’s not the relevant one. The relevant sample size must include the number of misses and a comparison of that rate to a baseline (the league average).
If I told you to test a population for a disease that scientists think occurs in one in one million people, you’d need to test millions and millions of people to draw a conclusion regarding the actual rate of infection. A sample size of 500,000 people, although it seems large, would be meaningless.
As it relates to Bailey, it doesn’t matter how many field goals he’s kicked; it matters how many he’s missed relative to the average.
Since kickers around the league make most of their field-goal attempts, it becomes more difficult to determine if Bailey’s accuracy is due primarily to his own talent or if it’s just variance.
In any large pool of players, we’d expect performance well above the mean just from chance alone. Maybe he’s just one of those lucky outliers. When you cut down his field goals to include only “clutch” kicks (however you want to define that), the relevant sample is ridiculously small.
Either way, the decision to give a kicker $7.5 million in guaranteed money probably isn’t so wise. The Cowboys are banking on Bailey continuing his past success, but kickers as a whole possess no season-to-season consistency.
As much as I like Bailey and hope he continues to thrive in Dallas, the odds of him continuing to kick at a league-leading level are probably no better than him becoming Nick Folk 2.0.]]>
I really believe most people have a distorted view of how teams should handle the draft. Namely, I reject the prevailing notion that teams should draft the best player available just because he’s the highest-rated player on their board.
In my view, teams should usually (but not always) draft the highest player on their boardat a position of need. The primary reason I believe that blindly drafting the best player available is disadvantageous is because teams overestimate their ability to identify the true best player available. In reality, NFL teams are pretty inefficient at drafting players, but they approach the draft as if their rankings are flawless.
One problem with drafting the best player available is that he’s probably an outlier on your board, i.e. you have him ranked higher than any other team. And while draft grades shouldn’t just be groupthink, there’s certainly value in knowing that the rest of the league doesn’t view a particular prospect like you do.
When you start to factor your own fallibility into the mix, the difference between your best player available and the next-best prospect shrinks. That inflates the value of drafting a player at a position of need.
Teams and fans know deep down that they shouldn’t always draft the best player available. What if the Cowboys’ top player available in the first round this year is a quarterback, but there’s a defensive end ranked one spot behind him? Clearly drafting the quarterback would set the franchise back; I’m just taking that argument and extending it a bit further.
Sometimes the true best player available will be at a position of extreme importance and need. Those situations make for easy draft picks. But other times things won’t align so nicely, in which case Dallas will need to make a choice between its board and need.
The Cowboys should never pass up an extreme value who is ranked very far ahead of other players, but otherwise, drafting the best player available is an overrated draft strategy that assumes infallibility in draft rankings that clearly doesn’t exist.
Verdict: Draft the best player available at a position of need.
Check out the entire post.]]>
The Numbers on Draft Pick Value
I charted the value of each NFL draft pick in regards to both the NFL trade-value chart and their actual NFL production. For actual value, I used Pro Football Reference’s approximate value as a grading tool.
I charted both forms of value in terms of the percentage of overall draft value that each individual pick encompasses. The first overall pick is worth 5.0 percent of the overall value on the trade chart, for example, but has historically accounted for far less in terms of the overall approximate value from his respective draft class.
At locations where the blue line surpasses the orange line, the cost of trading up is presumably too high. You can see that’s the case all the way up until around pick 20. There, the cost of the pick on the NFL’s trade-value chart is representative of how well that player should actually be expected to perform.
The obvious conclusion is that NFL teams are typically paying way too much to move up in the first round. They’re overrating the potential impact of the players selected there, particularly in the top 10. Those players are still expected to be the best in the class, but the cost for a team outside of the top 10 to move there is prohibitive.
First-Round Trade Results
Historically, teams trading down in the first round (or out of it altogether) have found a ton of success.
Amazingly, the teams moving down in first-round trades have acquired 64.4 percent of the total approximate value accumulated by the players involved in those deals, i.e. trading down has been far superior to moving up.
Perhaps even more amazing, the team trading down has gotten the best player in the deal 50.9 percent of the time! I mentioned that stat on Twitter and a reader responded that it simply makes it a coin flip. He was right that it’s a coin flip as to which team will acquire the best player, so the prudent thing to do would be to get that player at the cheapest cost possible. Stockpile picks in the range where production surpasses cost and maximize the probability of hitting on an undervalued asset.
Would you pay $30,000 for a car you can get elsewhere for $28,000? Of course not. Well, NFL teams that trade up in the first round have historically been paying extra for something they could have just gotten later. The cost is too high right now.
Read the whole article right here.]]>
I am of the opinion that every writer, analyst and expert covering any field should make specific predictions for which they should be held accountable. As it stands now, writers in particular sometimes have little incentive to make accurate predictions. With no skin in the game, what’s to stop them from simply making bold claims for the sake of drumming up controversy?
If all writers were forced to make very particular prognostications and then revisit those predictions to see where they went right and wrong, there would be a pretty strong incentive (pride and reputation) to get it right.
Most important, it would help readers understand who knows what they’re talking about and whose content is full of fluff. It’s really easy to look back on past events and analyze them after the fact, but it’s an entirely different endeavor to put yourself on the line in predicting what will take place in the future. It’s easier said than done.
This idea is why I make a number of very specific preseason predictions, as well as scrutinize the predictions after the season.
In 2012, I had what was, at the time, my best year in regards to Cowboys-specific predictions. I analyzed all of them right here, so you can check out my accuracy.
I say “at the time” because I think my 2013 predictions were even more accurate. I’ve listed the majority of them in this slideshow, both good and bad. I actually believe the predictions that go wrong are much more valuable because 1) I can determine if I indeed made a mistake (sometimes you can make a great prediction that simply doesn’t pan out) and 2) I can tweak the models or numbers I use to make predictions to enhance future accuracy.
See the predictions and analysis here.
Also, I just remembered to take a look at my preseason standings/playoff predictions. I posted those at Dallas News.
- Wild Card Round
Saints over Redskins
Cowboys over Falcons
Patriots over Dolphins
Bengals over Titans
- Divisional Round
49ers over Cowboys
Packers over Saints
Broncos over Bengals
Patriots over Texans
NFC Champion: 49ers
AFC Champion: Broncos
Super Bowl Champion: Broncos
Coach of the Year: Sean Payton
Offensive MVP: Aaron Rodgers
Defensive MVP: J.J. Watt
Offensive Rookie of the Year: DeAndre Hopkins
Defensive Rookie of the Year: Dion Jordan
My preseason Super Bowl prediction of Niners-Broncos is still alive. If you want to know the likelihood that my preaseason prediction comes to fruition, you can see the odds right here.]]>
Don’t play the Packers to run on first down.
Green Bay has been a pass-first team with Rodgers at the helm, although they actually run the ball on first down more than you might think.
In the first half—a time when games are “normal” in that they’re still about point-maximization for both squads—the Packers have actually run the ball more on first down than they’ve thrown it. That’s pretty surprising. It’s not surprising that the first down pass rate has decreased even more with Flynn at quarterback.
So why would I suggest that Dallas still play the pass? Risk and reward. The downside of playing the pass and getting gashed by the run might be 15 or 20 yards for running back Eddie Lacy. The downside of selling out against the run and having Flynn show play-action could be a quick deep strike for a touchdown.
Even with Lacy running well, I’d play aggressively against the pass and force the Packers to beat me with the run.
Attack the nickel cornerback.
The Packers have mixed their nickel cornerback strategy this year, rotating Davon House and Micah Hyde. Both have struggled. Below, I charted the yards-per-route allowed by the Packers’ and Cowboys’ cornerbacks.
I like to analyze yards-per-route because it rewards cornerbacks for quality coverage. When Darrelle Revis has such good coverage that he’s not even targeted, he should benefit from that.
You can see that Hyde and House have both been poor—in the same range as Cowboys cornerback Morris Claiborne, who we know has struggled. Meanwhile, cornerback Tramon Williams—one of the more underrated cornerbacks in the league—has given up a yards-per-route figure in the same range as Orlando Scandrick, who has turned in a career year.
If I were game-planning for Green Bay, I’d do everything possible to exploit their weakness in the secondary. Since they typically move cornerback Sam Shields into the slot in nickel situations, I’d leave wide receiver Dez Bryant out wide, using motion or whatever’s necessary to get him matched up on Hyde or House.
Don’t double-team outside linebacker Clay Matthews.
Matthews is Green Bay’s most well-known pass-rusher, but he’s not playing at an elite level right now. Below, I charted the pressure rate for Matthews and fellow Green Bay outside linebackers Nick Perry and Mike Neal, as well as defensive ends George Selvie and DeMarcus Ware, as per PFF.
You can see that Perry has been the best of the bunch, ahead of even Ware. Even Neal—a 6’3”, 285-pound monster for an outside linebacker—has a higher pressure rate than Matthews. The Packers’ star pass-rusher has pressured the quarterback at nearly the same rate as Selvie.
The Cowboys might be tempted to double-team Matthews just because of his big name, but I don’t think that’s a path they necessarily need to go down.
Also at Bleacher Report, I explained why Monte Kiffin’s defense isn’t working:
We can say all day long that every team suffers injuries and you need to respond to them, and while that’s true, it’s not like every team’s injury fate is equal. Some teams will just be more unlucky with injuries than others in a given season, and that obviously hurts their ability to produce. There’s a reason the starters are starters.
This season, 40 players have played at least one snap on the Dallas defense, according to Pro Football Focus (subscription required). Forty!
While the result is all that really matters for Dallas, the team’s ability to achieve the desired result is hampered when players like Caesar Rayford and Jarius Wynn are receiving significant playing time.
1) There’s no pressure.
The top reason that Kiffin’s defense isn’t working in Dallas, hands down, is that the Cowboys haven’t been able to generate much pressure. Take a look at the pressure rates for their top three rushers—defensive tackle Jason Hatcher and defensive ends George Selvie and DeMarcus Ware.
I marked the Cowboys’ wins with an asterisk. You can see the Cowboys’ pressure rates dropped from the beginning of the season to the midpoint—a stretch during which they lost to the Lions and barely beat a poor Vikings team. The pressure rate was at its highest against the Raiders and Giants—both games the Cowboys won.
While the Cowboys have been pretty lucky with takeaways this year, the best way to keep them coming is to get pressure. The correlation between defensive pressure and takeaways is astounding.
There’s not much the Cowboys can do about their injuries, and it’s not like Kiffin is going to dramatically alter his scheme at this point in the season. One things the ‘Boys can do to improve, though, is disguise their looks. They need to do something to create confusion for offenses.
Second, they absolutely need to find a way to get more pressure, even if it means blitzing more. It’s a high-risk strategy, but the Cowboys aren’t good enough to play conservatively and consistently beat good teams. They need to press the issue and just hope the ball bounces their way.
At ABC, I compared Aaron Rodgers, Matt Flynn, and Tony Romo:
There are various ways to measure yards-per-attempt. The blandest form of pure YPA can be a little misleading because it doesn’t account for aggressiveness. Tony Romo began his career extremely aggressively, and his YPA was at its highest. The problem was he was throwing a ton of picks, so it didn’t do all that much good. Recently, Romo has actually played too conservatively—minimizing interceptions at the cost of running an inefficient offense.
Net-YPA factors sacks into the mix. While sacks are frequently assigned to the offensive line, they’re actually more strongly correlated with the quarterback. There’s a reason Peyton Manning has “the best offensive line” wherever he goes; he makes them look like that.
Finally, Adjusted Net-YPA (ANYPA) is probably the most predictive stat we have in football right now. If you’re trying to predict the outcome of a game and you can look at only one stat, it should be ANYPA. That’s because it factors touchdowns and interceptions into the mix, weighting them according to their importance.
Looking at the numbers for each quarterback, you can see Romo is actually last in career YPA. Part of that is because Flynn just hasn’t played all that much, but you can see that Flynn’s numbers drop considerably once you account for sacks, touchdowns, and interceptions. That suggests that if Flynn were to play a less aggressive style of quarterback, his efficiency (in terms of YPA) would plummet.
Note that Romo has been one of the league’s premiere quarterbacks in terms of all three stats. The fact that Rodgers ranks so far ahead of him in two of them is amazing.
To better track the quarterbacks’ ability to lead their offenses without making mistakes, I charted their career touchdown and interception rates.
Not surprisingly, Rodgers has the highest touchdown rate and lowest interception rate. Romo ranks in the middle in both categories. Rodgers’ stats have been so impressive over the years because of the fact that he doesn’t throw interceptions. Again, it’s easy for a quarterback to post a high YPA when he’s being reckless with the ball. Not so easy when you’re throwing picks as infrequently as anyone in the NFL.
Finally, take a look at the completion rate for each quarterback.
Although scheme plays a big role in these numbers, there’s little doubt that Rodgers is an incredibly accurate passer. At this point, that’s probably Flynn’s biggest weakness—and the reason I think the Cowboys will take down the Packers if he’s their starting quarterback on Sunday.
And finally, I analyzed some trends for Dallas through Week 14:
After starting the year on fire in terms of sacks, Selvie doesn’t have a sack over the past month.
His pressure rate has been pretty volatile this year, but he’s been playing okay over the past couple games. One of the main reasons that Selvie hasn’t gotten to the passer quite as much, though, is that he’s just not playing as many snaps. The Cowboys have subbed him out at times; he has only 42 pass-rushing snaps in the past two games, for example—his only two games with fewer than 28 snaps versus the pass.
Let’s take a look at Dez Bryant’s workload.
The asterisks represent the Cowboys’ wins. When Bryant sees over 10 targets per game, the Cowboys are 3-1. When he sees 10 or fewer targets, the ‘Boys are 4-5. Since we know that there’s a possible selection bias—Bryant should see more targets in games the Cowboys are losing since they need to throw to catch up—the numbers are perhaps more significant than they appear.
Finally, here’s DeMarco Murray’s YPC by game.
This is a pretty obvious trend. I bring it up because Murray ran all over the Saints and Bears in recent weeks, yet the Cowboys got annihilated. Murray is averaging 5.2 YPC in losses and 5.3 YPC overall.