Sunday, May 6, 2012

Attribute Spotlight - Speed (SPD) Rookie RB's and CB's

UPDATE - See my blog on the Carry (CAR) attribute for QB's (Sack fumble % included): http://maddenmanniac.blogspot.com/2012/07/look-at-madden-12-qb-carry-car.html  See who is the worst QB at protecting the football.


ATTENTION - As of now Madden 13 will not give gamers the ability to hold fantasy drafts, online or offline.  If you would like to let your voice be heard by voting and commenting on the "fantasy draft" topic, please go to this link  http://www.easports.com/gamechangers/ideas?campaignId=27782
------------------------------------------------------------------------------------------------------------

I have been putting off this blog for sometime.  I wanted to make sure I covered all my basis and could deliver the information accurately.  So here it is.

The Madden attribute of Speed (SPD) is considered by many to be the single most important attribute in the game.  For years, SPD dominated Madden (bombs to Randy Moss and running tosses with Chris Johnson), it was the first attribute you looked for when setting your lineup. That said, in Madden 12 it seemed speed was not quite as powerful as it once was.  Don't get me wrong it's still important in the game, but offensive players no longer blew by the defenders with ease.

Here is a nice article by ESPN Blogger Jon Robinson: http://espn.go.com/blog/playbook/tech/post/_/id/408/rookies-envision-their-madden-characters  He interviewed 9 incoming rookies about their Madden characters, 5 out of them specifically mentioned their speed rating.  

For some time, the general assumption in the Madden community was that the SPD attribute was determined by 40 Yard Dash times (combine and pro day).  Recently, there have been claims by EA's top ratings person (Donny Moore) that "game speed" is more important than 40 time.

Something to also keep in mind w/ SPD ratings/40. We use game speed. Not, running in shorts & no pads speed. It factors in, but not only

So funny after the draft, everyone sending me 40 times (some legit, others questionable) for only the players that go drafted on their team

That was after tweets like this:

Remember the name Nick Taylor. Former FIU Point Guard signs 3 yr deal with Vikings to play CB. Ran 4.27-4.33 at Pro Day. #95+SPD

@JORO12 seeing low 4.41 being reported on some stopwatches for Mr. @JustBlack81. Low 90's range is in the ballpark for his SPD#madden13

Last May, Donny said this:
@Begasu26 40 time does not equal = Madden speed rating, that is the quickest answer




Ok, so if you are confused don't worry, you are not alone.  It seems like two different messages are being put out here.  No wonder fans send Donny tweets about 40 times.  Well, I set out to find the TRUTH and let the numbers speak for themselves.  I am not trying to bash Donny or EA here, but this is the type of information fans are currently receiving.  One of the main things I have asked for in Madden ratings/attributes is consistency, these tweets are far from consistent.

What you are going to see in the following data, is that 40 times have an incredibly strong effect on a players initial (Rookie) Madden SPD attribute.  You will also see that the effect of 40 times on the SPD attribute lessens after the rookie season.  This must be the "game speed" Donny is talking about.

BTW - I could not find a "game speed" calculator anywhere and I found no data available to advise someone how to determine "game speed".  <---This is truth and sarcasm mixed in.

To me, saying "We use game speed", is just another way of saying we just guess.  In most instances, that is the guess of only one person.  Talk about a subjective (bias) way to rate players.  My guess is... it would take a room full of skilled NFL scouts to accurately determine "game speed" and those scouts would have to watch every player in every game to consistently and objectively assign those game speeds.

I hate to break the news to all of you, but Donny doesn't watch every player, every play, of every game. It's not his fault, it's impossible for one person to do that.  That is were 40 times come into play.

Here is a link to NFL.com Mike Mayock discussing combine drills (including the 40):
http://www.nfl.com/combine/workouts

Click on the drop tab in work out drills to listen to him about the 40 yard dash. Here is a blurb from the webpage:
   "40-yard dash
The 40-yard dash is the marquee event at the combine. It's kind of like the 100-meters at the Olympics: It's all about speed, explosion and watching skilled athletes run great times. These athletes are timed at 10, 20 and 40-yard intervals. What the scouts are looking for is an explosion from a static start."

The great thing about using 40 times to determine initial Madden speed, is that it's the most objective way for one person to do it.  These times help reduce bias when one individual is doing most of the player rating.  Every player is on the same page at the NFL combine, the same set of procedures are used to time all players.  ProDay times are the 2nd best source in my opinion, since not every player is at each school's ProDay, but it is the best available data for non-combine invitees.  Unofficial reported times for players who didn't go to the combine or have a documented ProDay are the least reliable in my opinion.

Here is my process for data collection and interpretation:

1.  I collected the initial (1st Madden 12 roster) SPD attributes for all rookie RB's and CB's in 2011, plus I collected their final (Last Madden 12 roster) SPD rating if there was any change.

2.  I then collected all of their 40 times.  If they participated at the NFL combine, I used that 40 time.  If they didn't run at the NFL combine I used there college ProDay time.  Otherwise, I noted the only unofficial reported time I could find on the internet.  My primary references were ESPN.com, NFL.com, and NFLDraftScout.com.

3.  I created a spreadsheet for each position and transferred the data to a scatter plot to determine if there was any correlation between Madden 12 rookie SPD attributes and their 40 times.  I used initial Madden 12  SPD attributes for the scatter plot.

4.  I repeated steps 1 thru 3 for all 2nd year RB's and CB's in Madden 12 for a later comparison.

5.  In the end I used this information to predict the Madden 13 Rookie speed (SPD) attributes for every RB and CB the went to the combine, was drafted, or signed as an undrafted free agent (UDFA) this off-season.

Here is the data, enjoy:

M12 SPD (O) = Original Madden 12 roster speed


M12 SPD (F) = Final Madden 12 roster speed (if blank, SPD did not change from original SPD)


* = Not on original Madden 12 roster, this is final Madden 12 roster speed






Here are the scatter plots:

Just a quick glance at these graphs and you can see the relationship between Madden 12 rookie speeds and 40 times.  I inserted a trend-line and the R^2 (R squared) value to show how strong the correlation is.  The closer the point is to the line, the stronger the correlation.  It's widely accepted that any R^2 value over .80 is a strong correlation.  The closer the R^2 is to 1.0, the more accurately you can predict future SPD ratings.  

You can see all three of these graphs have an R^2 over .80.  If you are interested, here is more information on this type of statistical anaylsis:  http://mathbits.com/MathBits/TISection/Statistics2/correlation.htm

The bottom line, is that in regards to rookie RB's and CB's there is a strong relationship between the assigned Madden 12 speed (SPD) and their 40 times.  


Maybe the folks at EA are not aware of their own tendencies, but in this case the numbers don't lie.  I am actually happy they use something more objective like 40 times, it beats the alternative of one person guessing "game speed."  Since most of us can now agree that EA uses 40 times as the primary factor for rookie SPD attributes, why won't Donny Moore just admit it?  Why deny it?  

I say, keep sending Donny legit 40 times for the incoming rookies, because they clearly have a significant impact on a rookie's initial Madden speed.

* Please note - While there is a strong correlation between M12 speed and 40 times (for these positions), I am not claiming to be able to predict future rookie speeds with 100% accuracy.  These R^2 values indicate that over 80% of the time, SPD can be explained by looking at the 40 time.  Also, this blog only pertains to rookie RBs and CBs (I plan on doing additional positions in future blogs).

------------------------------------------------------------------------------------------------------------

Well, that explain rookies.  What about after their rookie year?  Well that is when the so-called "game speed" takes over.  Or in other words, if your performance is good, you get rewarded with a SPD boost, if you are bad you get penalized with a SPD decrease.

Remember, SPD is the top factor (tied with AWR, CAR, BCV) in determining a RB's overall (OVR) Madden 12 rating.  Therefore, as player does better throughout his career he is likely to see an increase in speed regardless of 40 time.  The opposite is true as a player struggles, SPD is likely to be decreased regardless of 40 time.  It's not about "game speed", it's about performance and manipulating the OVR formula (which of course is a horrible formula, that is why you see low CAR attributes even though a RB doesn't have a fumbling problem).  Granted, you can't use the 40 time forever.  Players do age and will naturally slow done.  How much they slow down should be determined by a group of people, not one person.  Again, the goal should always be to reduce bias and be as objective as possible when assigning player ratings/attribues.    

For CB's, SPD is the 2nd most (tied with ACC) important attribute in determining the Madden 12 OVR rating.  MCV is the most important.  So speed will also be manipulated for non-rookie CB's to either increase or decrease OVR, regardless of their actually speed.  "Game Speed" when it comes to Madden attributes is only a myth.  Is anyone at EA really confident enough to say they can accurately determine a player's speed by watching a live game? (Even if they claimed they could, are they really watching every player, every play, and every game to objectively determine that?  NO, they are not.)  Let's call it what is it.  It's a way to reward or penalize players.

Again, let's look at the data.  Here are the 2nd year RB's and CB's in Madden 12:

* Please note - for Dexter McCluster I used his ProDay time since his Combine was viewed by many as a disappointment.  This is also what I believe Donny Moore did based on my research.



* Please note - for Joe Haden I used his ProDay time since his Combine was viewed by many as a disappointment.  This is also what I believe Donny Moore did based on my research.  Not to mention, Josh Looman at EA is huge Browns fan and I'm sure he had an influence on Haden's SPD attribute (he had a horrible 4.62 combine time, 84 SPD range).    

Here are the scatter plots:



A quick glance here, shows a drastic difference in the the Madden 12 speed and 40 time relationship.  The RB sample size was clearly smaller (just the nature of the position in the NFL), but still shows a correlation between SPD and 40 time.  That correlation can no longer be considered strong, but it's there.

The CB's have a larger sample size, but more of them are further from the trend-line.  You have to see through the clutter and look at the R^2 values.  The 2nd year CB's have very little correlation with their 40 times.

Combined, the CB's and RB's still show a correlation, but it's considered weak.

What does all of this mean?  Basically, performance will have a bigger impact (increase or decrease) on SPD with 2nd year players than reported 40 times will.  If you look at the spreadsheets, I believe you can safely make that assumption.  This is what EA calls "game speed".  I call it a more subjective (bias) way of rating player attributes, especially when one person is doing most of the rating.

When Madden 13 comes out, I will revisit the speed attributes of all of the players in this blog.  My guess is the biggest speed adjustments for non-rookies will take place when going from one title to the next (i.e. Madden 12 to Madden 13), that would be EA's opportunity to make those speed adjustments with fewer fan noticing.  In season speed decreases are not a welcome thing by most Madden gamers.

There is currently no indication that EA is going to adjust (fix) the current broken overall (OVR) formula, so comparing M12 attributes with M13 should be credible.

To wrap this section up.  Donny is both right and wrong.

He is wrong when he says, "40 time does not equal = Madden speed rating", because it clearly does (in most cases) in regards to rookies.

He is right when it comes to non-rookies, yet there is still a correlation (albeit weaker).

In the absence of a complete Madden ratings team (this would be a huge undertaking, with at minimum of one skilled ratings representative for every NFL team), I would prefer the 40 time be dominant factor in regards to player SPD (noting that age will catchup to most players and eventually reduce their speed).  Sorry, but using so called  "game speed" factor to determine player SPD, won't provide consistent and accurate attributes.

Please, anyone who can tell me the true definition of "game speed" please let me know.  While your at it, please tell me were I can get fully educated and certified to assign a "game speed" number to every NFL player.  I simply haven't be able to find it.

I think there is a reason why the 40 time is included at the NFL combine.  If it wasn't important, it wouldn't effect the draft position of so many players every year (Yes, other teams besides the Raiders look at combine numbers).
------------------------------------------------------------------------------------------------------------

Here is what you have all been waiting for.  Below is my projected Madden NFL 13 Rookie speed (SPD) attributes for all RB's and CB's that participated at the combine, were drafted, or reportedly signed as an UDFA.

I included draft position, because I believe (and have shown in my Rookie ratings blog series), that Donny will overrate rookies based on draft position.  I believe he will go with a lower ProDay 40 time for the higher draft picks if they disappointed at the combine.  I do not claim that these projections are 100%, but I do believe they will be very close to the Madden 13 speed attributes for rookie RB's and CB's.  Also, I wouldn't be surprised if Donny continues to deny the impact that 40 times have on rookie speed ratings.

After the projections, I will discuss the 2012 Madden Ultimate Team speeds that some of these players received on draft night.  Some people think the MUT ratings might indicate Madden 13 ratings/attributes.

* Please note - The formula for RB's is slightly different than that of CB's based on my research.  There may be some small differences in the two positions, even with identical 40 times.  It's a very small difference, but this is based on current EA speed rating tendencies.

RB's First:



  - David Wilson received a SPD of 93 in MUT12.  I projected an 89 based on his combine time.  He ran a 4.4 at his ProDay which would have resulted in a 93 using the same formula.  This is an example of the 1st rounder getting the benefit of the doubt.

- Doug Martin received a SPD of 87 in MUT12.  I projected an 87 as well.  He only ran at the combine, there is no ProDay 40 time to give him the benefit of the doubt.  I guess his "game speed" and combine time just happen to matchup.  Since he was drafted by Tampa Bay, I have no doubt EA will quickly increase his speed if he performs well.  Many folks at EA are Tampa Bay fans.  Keep an eye on this one.

- LaMichael James received a SPD of 91 in MUT12.  I projected a 91 speed as well (combine 40).  He ran a 4.41 at his ProDay, but slipped to the 2nd round which may explain him not getting the benefit of the doubt.  A 4.41 would have resulted in a 92 SPD using the same formula.  If he has success early, Niner fans will scream for a SPD increase and will likely get it.  Niners fan pound Donny Moore on twitter and he gives in to them more often than not.

- Trent Richardson received a SPD of 90 in MUT12.  I projected an 89 speed based on his ProDay time.  He didn't not run at the combine.

Here are the projected speeds (SPD) for the Madden 13 rookie CB's:


- Stephon Gilmore received a 94 SPD in MUT12.  I projected a 92 based on his combine time.  He did not have a recorded ProDay time.  I found a unofficial reported time of 4.38 (here: http://www.buffalobills.com/news/article-2/Bills-go-with-CB-Gilmore-at-10/e4577261-4e74-4136-9ea9-f915156512f4 ) ,which would have been a 94 using this formula.  He was the 10th pick in the 1st round, so again he was given the benefit of the doubt by EA.

- Dre Kirkpatrick received an 89 SPD in MUT12.  I projected a 88 based on his combine time.  An unoffical reported time of 4.39 will come out if research Kirkpatrick on the internet, but it has no real credibility.  EA correctly disregarded that number, but did give him a slight bump for being drafted in the 1st round.

- Morris Claiborne received a 92 SPD in MUT12.  I projected an 88 based on his combine time.  He ran 4.43 at his ProDay, which would have resulted in a 91 SPD using this formula.  EA gave Claibourne the benefit of the doubt big time.  This is overrating based on his draft position, he was the 6th overall pick and oh by the way, was drafted by the Cowboys.  Let the Cowboy bias conversation begin.  EA makes it too easy for fans to see the clear bias in favor of the Dallas Cowboys.  Come on EA, really?


And just for the fun of it.  Let's do RGIII.  I have not studied the 40 times of rookie QBs or the tendencies of Donny when rating their speed, but let's put his 40 time into the CB formula.

Robert Griffin III received a 93 SPD in MUT12.  His combine 40 was 4.41 which would have resulted in me predicting a 92 SPD based on the CB formula.  RGIII was the 2nd overall pick in the draft and had a rumored unofficial 40 time of 4.32.  To be fair, RGIII was also a very accomplished high school track athlete who set several state records in Texas.  There is a lot of information out there that would back up a 93 SPD rating.
------------------------------------------------------------------------------------------------------------

That's it guys.  I know it's a lot of information and I have no problem if you just skipped to the end to see the Madden 13 speed projections.  I plan on doing more positions in the coming weeks, but this kind of blog takes some time to put together.  I need a break.  I would like to thank my loving wife for her support and encouragement to follow my passion.  She is very understanding, while I juggle my real job, family, and this passion.  I would also like to thank all of you that follow this blog, you drive me to do better and dig deeper.  The blog went over 10,000 hits yesterday, I am still amazed.  Next stop, 100,000 hits.

Have a great week.  You can follow me on twitter @mannmicj

Also, don't forget to tweet Donny Moore all of your rookie 40 times (they are important).  You can tweet him @Donny_Moore or you can tweet @EAMaddenNFL

6 comments:

  1. "EA makes it too easy for fans to see the clear bias in favor of the Dallas Cowboys. Come one EA, really?"

    Stop Crying... waahhhhh waahhhhhh omg those cowboys.

    Good article though

    ReplyDelete
    Replies
    1. I am just stating the obvious. Here is a blog I did after mid season on Cowboy bias: http://maddenmanniac.blogspot.com/2011/12/ea-madden-ratings-dallas-cowboy-bias.html

      I respect your opinion and my guess is you are a Cowboy. Like I said, it's too obvious to ignore. It's been that way for years. Thanks for sharing your thoughts and reading the blog.

      Delete
  2. Wrong on josh Robinson, he was the fastest clicked at the combine at 4.28 40 time

    ReplyDelete
    Replies
    1. Actually, I am not wrong. At the time of the combine the numbers are unofficial. The official combine numbers can be found here:

      http://espn.go.com/nfl/draft/combine/_/id/92

      and

      http://www.nfl.com/combine/top-performers#year=2012&workout=FORTY_YARD_DASH&position=QB-RB-WR-TE-S-DL-LB-CB-OL-SPEC

      You will see that both of these resources have Robinson as 4.33. Thank you for reading my blog.

      Delete
  3. i just read today that stanford routt fabian washington and darrius heyward bey all clocked 4.28 yet bey has a 97 washington has a 98 and routt has a 95,why?they should all be 98,and demarcus van dyke has a slower 40 speed but a faster madden speed that routt,why?its arbitrary to me.bo jackson ran a 4.12 he should have had 100 speed but he didnt lol.and chris johnson clocked faster than desean jackson yet still both got 99?oh well ill drop you all a little tip on madden,4-6 ss blitz or normal quarter defense are the best for getting sacks,not any nano crap.i got james harrison lamarr woodley and julian peterson all over 50 sacks 8 games into my current season.one thing is u have to have a good mix at the right positions of high rating and high speed.i sub peterson in at de for lamaar houston for oakland in the normal quarter defense and he kills the rt and with help of the blitzing fs 2 he gets in the backfield alot.same for the steelers or any other team you use.i got to the point now thru experimenting with player combos that i have 6 shutouts this season and average maybe 10 pts allowed per game.or less.i just know i only gave up 20 pts once in 16 games and 6 shutouts in 16 games.all other games i allowed maybe 10-17 pts.plus i score 50+ per game .

    ReplyDelete