Saturday, March 10, 2012

Attribute Spotlight - QB Deep Accuracy (DAC)

    Let's talk about NFL/Madden quarterbacks.  Everyone loves the deep ball in Madden and in the NFL.   What you will see in this blog is that the Madden attribute for deep accuracy (DAC) is WAY off for some QB's and basically neglected for others.  To be honest, some of the DAC attributes are accurate based on my research, but I believe this to be mostly a coincidence since the other numbers make no sense and don't appear to be based on anything other than general perception.  Perception may play a role in more subjective attributes like spectaclar catch (SPC), but it should have very little to no impact on a more objective attribute like DAC.

    I believe that with more accurate DAC attributes in Madden, the gamer will have a more true to life experience.  Can't DAC in Madden be based off of statistics?  I believe it can and should be based off of statistics.

    I had to educate myself on the deep accuracy (DAC) of NFL QB's before I could provide any credible analysis.  The only place I could find that provides deep accuracy statistics for the NFL is They have only been providing this statistic for the last 4 seasons, so that is what I had to work with.  PFF defines a deep pass as a targeted pass of 20 yards or more.  For computing the deep accuracy of the QB's, PFF included attempts, completions, and drops of 20 yards or more.  You may ask why drops?  I actually like it.  PFF defines a drop as a "catchable" pass and they apply the same rules to all players equally.  By having drops included in the deep accuracy completion percentage, the QB doesn't get penalized for having poor receivers.  All the QB's are graded equally and for that reason I believe the deep accuracy statistics on to be very credible.

   PFF has their statistics broken down by season for each player, but DOES NOT compile averages for the League as a whole, they just go by individuals.  And while they have four seasons worth of statistics, they don't supply 3 or 4 years averages.  So, I caculated the deep accuracy of all NFL QB's to come up with yearly totals for the entire league, as well as 3 and 4 year averages.  By doing this, I can compare each QB to the NFL average to see how each individual stacks up before assigning the DAC attribute ratings.

    Please note that I only used QB statistics for the NFL totals, I excluded any passes made by other position players such as:  RB, WR, TE, FB, K, or P.  This way QB's are only rated against other NFL QB's.

    Here are the NFL totals for each of the last four seasons, plus the 3 and 4 year averages:

    You can see here that there has been a gradual trend upward since 2009 in deep accuracy completion percentage, but when you go back to 2008 it's very close to what we see in 2011.  Even though we only have four seasons of deep accuracy statistics, I believe you can still determine a benchmark.  For Madden attribute ratings, I like looking at the big picture.  For that reason, I give more weight to the last three years than the current season.

    In Madden 12, there are 125 Real NFL QB's (there are an additional 8 fake QB's that are EA employees and one U. S. Marine female who won a contest).  The average DAC of the real QB's in Madden 12 is 69.  Know that the QB's with lower OVR's tend to have poor deep accuracy regardless of reality and that brings down the overall average.  While 69 is close to what I came up with (73), I believe it's a little too low to be used as benchmark average.  The DAC attribute range in Madden 12 is 47 to 96.  The most frequently assigned (mode) DAC rating is 66 (9 times).

    The following table shows some of the Madden 12 QB's and their current DAC attribute ratings.  I selected both popular and unpopular QB's for this table.  Each QB was compared to the NFL average for that column.  In the end I came up with my suggested DAC attribute ratings for each of these QB's.

Blue = Better than the NFL average.

Red = Worse than the NFL average

M12 DAC = This is the current DAC attribute rating as of the final roster update in Madden 12.

** Please remember per ProFootballFocus, deep passing statistics are "derived from passing attempts targeted 20 yards or more downfield."

This table is sorted by highest deep accuracy from 2009 to 2011:

 My observations:

- Notice both Peyton Manning and Tom Brady and below the league average for the last 3 seasons and Brady is actually below the average for the last 4 seasons.  Both of them have missed a season in the last four years as well.  Yes, it's very strange that Manning and Brady have the exact same numbers over the last three years, but I triple checked it and it's true.

- I came up with a 73 Deep Accuracy rating for the average QB when looking at the last 3 to 4 years in the NFL.  Both Peyton Manning and Tom Brady should be closer to the average, instead they have been overrated in the game due to general perception.

- Tim Tebow is flat out horrible, but due to TebowMania he is also vastly over rated.  He currently has an 80 DAC, I would rate him a 65.

- Carson Palmer was above average this year, but he is vastly overrated when you look at the entire picture.  Even if I rated him only on this season he would have been a 79 or 80 in DAC, not his current 86.  When looking at the entire picture, I rated him a 69.  The general perception is that Palmer has a very strong arm and is a great deep passer.  Well, his arm may be strong, but his accuracy is nothing special.

- In no way should Eli Manning have the highest DAC in Madden 12.  While Eli is good, I have him at an 88.  During the NFL playoffs, Eli's DAC attribute went up 10 points.  Now let's take a look at his playoff numbers for passes targeted 20 yards or more downfield:

14 attempts, 6 completions, 0 drops, for a completion percentage of 42.9%.

That percentage is lower than his 2011 regular season statistics and his three and four year averages.  I am sorry, but the ratings folks should be embarrassed by his current DAC attribute.  It's simply not based on reality, regular season or post season.  While he completed some big passes in the playoffs, that shouldn't be a reason to increase his DAC 10 points.  Isn't Eli's "clutch" trait suppose to effect those big-time passes during crucial situations?

- I think it's clear that the best deep passer in NFL is Drew Brees.  He has been very consistent over the last four seasons, but doesn't appear to get the credit in his DAC attribute.  While an 87 is good, I gave him  the highest DAC with a 95.  Again, I put more weight into consistency over the last three to four years.

- While Aaron Rodgers was better than Brees in 2011, he has never come close to that number in the past.  I gave Rodgers a 90, which is the second highest DAC that I gave out.  If Rodgers' proves that 2011 was not a total outlier, than his DAC will increase in future Madden titles.

- Matt Schuab ended up with an 88 because of his slight regression this year before he was hurt.  Based on his last 3 to 4 years he is very underrated in DAC and earned the increase from 79 to 88.

- What about incoming Rookie QB's?  How should their DAC be determined?  As with all rookies, the college statistics are not as good or as in depth as the NFL statistics.  This is when a ratings team would have to get together to look over all of the predraft analysis and scouting reports before assigning the DAC attribute ratings.  A ratings team could then come to a consensus on what the the DAC rating for that player should be.  I think the DAC rating for rookie QB's should be conservative to begin with, but can be tracked throughout the season.  Rookies might require a little more adjustment than a veteran as they accumulate more NFL statistics.  By the end of the season if they started a significant number of games, the ratings team would have a better idea of how the rookie performs on the NFL level.

Example - Blaine Gabbert, 2011 Deep Accuracy Rookie Stats:

                36 attempts, 9 completions, 1 drop, for 27.8%.

    That is absolutely horrible.  I would assign Gabbert 51 DAC attribute rating at the end of the year.  Over the next couple of seasons he could drastically improve that as he accumulates more (better) stats.

- What about backup QB's with little experience?  Most of them would have there rookie ratings still.  You could also use what statistics they have accumulated to fine tune the DAC with the ratings team.  Veteran backups who use to be starters will have data to go off of.  Regression will show in those numbers as well as the fact that they are no longer considered a NFL starting caliber QB, such as McNabb.  McNabb was at 28.6% in 2011 until he was benched.  His three previous seasons were: 38.3% (10'), 43.5% (09'), and 39.7% (08').  He was only above the NFL average once over the last for years and is showing clear signs of regression.  His DAC should reflect that regression, by using statistics and discussion with the ratings team.  I would give him a 69 DAC attribute going into Madden 13.  

Final Remarks:

  In general I don't think a QB's deep accuracy should change much during a season, definitely not the amount we saw with Eli's increase during the playoffs.  This goes back to having a broken OVR formula.  Because DAC is the 5th most important attribute when it comes to the QB's OVR, it's often used to increase or decrease the OVR instead of representing reality.

    If EA Madden Football wants to be true to the real NFL game, than statistics such as deep throwing accuracy should be utilized when assigning attribute ratings.  Having more accurate player attributes will produce more realistic gameplay, while rewarding gamers for utilizing the player strengths not their weaknesses.  The underrated QB's like Schuab would become more effective in the deep passing game just like in real life.  The overrated QB's like Brady and Manning would be a little less effective in the deep passing game, just like real life.

    It's time to get rid of the OVR formula and make it a stand alone number or ranking.  It's time for statistics to play a larger role in attributes that are more objective.  The DAC attribute should be almost entirely objective, with a little subjectively in the fine tuning.  Subjectivity and perception SHOULD NOT be the biggest factor in this attribute.

    Stay tuned for more suggestions and projections on Madden 13 player ratings and attributes.  Follow me on twitter @mannmicj to receive updates on my newest blog posts.




  1. great write-up... great understanding and correct usage of the stats

  2. Send this to Ea sports and the madden team right d**n now. Its about time someone with some sense showed in writing and with a proven formula what we all already have known for some time now.