Madden: A Look Inside for CFM Play

[khaliib]

Whether the discussion is about Player Ratings, Game Sliders, Draft Classes or CFM in general, there's always a cloud of mystery about what is what in the game and much is left for us to come up with our own theories.

So for the past two months, I've been digging through the game install of the 360 version to find something that would bring some clarity into these mysterious areas and their structure/function within CFM.

This is not intended to create a debate thread or challenge individuals conclusions derived from testing dealing with ratings, sliders or otherwise.

Just providing some information that helped me and hopefully bring a some clarity to help improve someone's else CFM experience.
Some notations may/may not already be known by some.

(I can/will not post any of the data pulled to support what's being noted, but will just give a synopsis of the information read)


Sliders
A) Game Difficulty Levels
1) Rookie, Pro, Alll-Pro, All-Madden
- the last change/edit the gamer leaves from is the setup the game will use.
- ie edit Pass Block, thus making it a "Custom Set" or switching from All-Madden to All-Pro, making it a default setup.

2) Default difficulties sets every Hum or AI slider to a specific level.
- All-Madden 25 (range consist of 21-25)
- All-Pro 50 (range consist of 46-50)
- Pro 75 (range consist of 71-75)
- Rookie 100 (range consist of 96-100)
**Unknown why there is a 21 pt difference between each level.

3) Some sliders from previous are no longer excessable, but remain although hidden or may have simply been renamed
- "Broken Tackles" under Rushing
- "Pass Rush" under Pass Defense (renamed to Pass Coverage)
- "Reaction Time" under Rush Defense
- "Block Shed" under Rush Defense

- If we select Pro, every Hum slider, including the hidden ones, are set at "75" until it is altered
- Like wise, on All-Madden, every Hum slider is set at "25"
-This is important because the hidden sliders remain at these levels and AI Sliders that would counter "must" be altered to balance/compensate
(i.e. Pro = Hum Broken Tackles = 75, so AI Tackling must be 75 to balance, or higher to give AI an advantage )
- This is the pattern for every level

**Rookie Level is actually playable/challenging with an edite default roster that's in-line with the CFM Draft Class rating min/max structure coupled with this information

**Difficulty Challenge actually lies in the "Speed" the game is played at
- ie slower allows you see, processs and react easier, while faster speeds minimizes the gamers time

4) Sliders shift/alter the ratings +/- that they are meant to impact
(ie WR Catching = Catch, Spec Catch, CIT, Route Run)


Draft Classes
1) There are 30 different Draft Classes
- each class has pre-made/defined players made
- every time a new CFM is started, every player's stat (i.e. height, weight, build, ethnicity etc...) in that class is manipulated to produce a different player

2) Colleges impact Ratings/Back Stories/Dev Trait
- Designated as Large, Medium and Small
- players from Alabama (large) type, tend to have a lot of draft "Bust" mingled with players that start with great ratings
- small schools have some hidden gems in them that look like scrubs in the beginning.


CFM
1) Created Coach's impact AI gameplay/decisions
- adjusting Run/Pass ratios and Aggressiveness then setting to Auto Pilot (for non-user controlled teams) allows gamers the ability to edit the AI gameplay of each team
- chosen Back Story will impact drafting, free agent signings etc... to meet the goals of the chosen Back Story.


Last but not least,

Player/Ratings
1) Ratings function on 1 or 3 levels
- Animations
- Die Roll/Win Loss
- Modifier

2) Animations
- Speed
- Agility
- Catching
- Trucking
- Release
- Spectacular Catch
- Stiff Arm
- Spin
- Juke
- Impact Block
- Tackle
- Press
- Play Action

3) Die Roll/Win Loss
- Strength
- Jump
- Carry
- Run Block
- Pass Block
- Power/Finess Moves
- Block Shed
- Hit Power
- Catch In Traffic

4) Modifier
- Awareness
- Acceleration
- Route Run
- Ball Carry Vision
- Elusiveness
- Play Rec
- Zone Coverage
- Man Coverage
- Pursuit
- Short, Mid, Deep Pass Accuracy
- Throw On Run

5) OVR has two functions
- Depth Chart positioning
- Contracts (length, amount)

6) Awareness Functions (has nothing to do with AI awareness, might be miss labeled)
- Is the actual OVR of a player (might be mislabled)
- the higher this is, the higher the #5 is.
- Players in Draft Classes depth chart positioning are in-line with this rating
(ie the higher the rating, the higher they are on the depth chart)

- Modifier of all ratings
(Players will have their high ratings modified by having low AWR)
- as it increases the closer to that rating the AI controlled player will function
(ie Catch = 85/AWR = 40, player will drop passes even though the rating is high)


7) Players are "Tiered" as 4 Types (based on ratings)
- Elite
- Good
- Base
- Poor

8) Min rating level to be classified as "Elite" in that area (differs per position)
(Half Back)
- AWR = 65
- ACC = 88
- SPD = 88
- AGI = 88
- CAR = 76
- ELU = 88
- JUK = 90
- Stiff = 84
- TRK = 88
- SPN = 84
- BCV = 82

**Same is done for other Tiers and all positions

9) Some missing Traits but are in the game
- Initiate Contact
- Jump Route
- Deep Ball
- Pump Fake
- Tackle Low
- Wrap Up

Added 12/9
A) Auto Sprint is on behind the LOS and on Kick Return.
- As soon as the ball carrier crosses LOS, the initial sprint is slowed

- Since there in no LOS on KR, AI returners consistently speed burst pass their blockers.
- this is critical because players need to be interacting in a blocking animation for AI ball carriers to recognize running lanes.
(this is why AI returners run away from the blocking, because it thinks there is a lane with no blocking interaction occurring)
*** (may/may not be a bug)

B) Ratings and/or Traits per position that drive AI Intellegence (i.e. Awareness)
1) QB
- Elusiveness = Tackle/Blitz Avoidance
- BCV = Open lane recognition as a "ball carrier"
- QB Aggressiveness Trait + Player Type (not Tendancy) = Attempted Pass Types
Aggressive Trait = old Force Pass Trait = how much radius needed between WR and Defender recognition
Player Type = i.e. Strong Arm = attempts mostly deep routs with DAC as a modifier upon Aggr Trait
- Sense Pressure = radius to declare to pass
** I'm digging to see if this has an impact on over riding PBK.
ie Ideal = pocket to be held longer.
ie Oblivious = pocket breaking down faster.

2) Ball Carrier
- BCV = Open Lane Recognition
- Elusiveness = Tackle Avoidance

3) WR/Route Runners
- there currently is nothing structured to represent intelligence for route runners, as routes are ran with no respect to how the route runner is being covered.
The closest we can get is...
- Route Run = getting open

4) Blocking (getting there, but not quit yet)
- Pass/Run Blocking = Breach Radius Rec
Higher ratings = greater radius causing blocker to block a free defender that breaches the radius.
They will even slide/warp towards that defender (really evident on PAT)

5) Pass Defense
- MCV = coverage radius to route runner
- ZCV = recognition of ball leaving QB hand
- Jump Rating = Die Roll/Win Loss for INT/Swat/Knock Down
**this is why DB's and LB's are rated so high in draft classes, while most DL are not (JJ Watt carries a very high rating here because he's known for batting down many passes)

**May be the reason (outside of Route Run) why Slants are so open and easy.

6) Rush Defense
- Play Rec = Run Support (just like ZCV is for pass def)
Higher ratings will cause instant "ball carrier" recognition
***Difficulty Level impacts/modifies Rush Def "Reaction Time" Slider because is set at that difficulty level


B) QB min for "Elite" status
- Play Action 62
- BCV 88
- Elus 78
- Juke 86
- Acc 88
- Agil 88
- SPD 82
- THP 92
- SAC 82
- MAC 75
- DAC Unknown, went to gibberish

**There are minimums with Luck designated before them.
Don't know if these are placeholders for A. Luck type of QB's or what.
- Luck SAC 85
- Luck MAC 80
- Luck DAC 75

WR min for "Elite"
- ACC 92
- SPD 93
- AGI 91
- AWR 56
- CAT 80
- CIT 77
- RRun 70
- REL 76
- SCAT 80
- SPN 82
- Juke 90
- Jump 91 (this must be reduced for lower tier WR to minimize win/loss modifier against better defenders)


C) Rookie Level is the only level that removes the "instant" direction changes of AI defenders because the Rush Def "Reaction Time" (set at 25 for AI) is modifying defenders AGI and Pursuit ratings that are generated so high.

12/12/15 Update
A) Play Recognition and BCV (needs a closer look/testing for "true" impact in CFM)
1) Every position has these two listed for AI Intelligence

2) Looking at Draft Recap immediately after draft has concluded, "every" rating is listed for the players except "Play Rec".
- Every player has a ratings of "0" until the start of the 1st game, then players are given a value.

B) Combine Stats
1) Broad and Vertical Jump are both designated for determining "Jump" rating

12/14 Update
A) Injuries (as a Modifier)
1) Players ratings are "negatively" affected and do not come back at the same levels.
- how much "weight" applied to decrease and/or what specific ratings are targeted are unknown.

B) If you want a greater twist added to your CFM, increase the Injury Slider
- types of injuries are expanded
- Toughness and Injury Ratings need to be considered when drafting a player that might be a starter.

C) Trainers are "VERY" important dealing with injuries
- they dictate "how much" if an decrease a player will suffer once returned to play.
- not many "Elite" level Trainers to minimize negative impact on ratings.
- signing an Elite Trainer will make it difficult to sign an Elite Scout or better coach (for owners) at the same time.

1/4/16 UPDATE
A) Player Types
1) Players are not designated according to their natural "Player Types" base off of type rating levels
(ie Mack listed as "Balance DE" after being moved to DE)
- Must be switched, saved, then switched back for position change to occur.
- NOT ALL players can be done for some unknown reason.
- this can be done anytime during CFM

2) Every player has a "Trade Status" designation that affects Trading.
- None
- Untradable
- Tradable

3) Players not meeting their goals due to not being played/utilized, will "Hold Out" to force a trade or release from the team.
- Both Human and AI

B) Coach's
1) Every coach has these ratings but I didn't see anything that noted their impact on players (must test)
- Chemistry
- Knowledge
- Motivation
- Performance Level
- Work Ethic

2) These are known
- Team Building = how aggressive they go about building the team (i.e. Normal, Through Draft, Through Free Agency)
- Trading Tendency = their trading philosophy
(i.e. Trades Up, Trades Down, Does Not Trade)

1/19/16 UPDATE
A) Roster taken into CFM is "crucial" towards integrity/fluidity of the Mode.
- If rerated rosters consist of players rated too high or too low, CFM mechanisms that work through draft classes are derailed.

1) Draft Classes
- Most OVR's (see above for OVR purposes) fall between 58 and 80'ish, with most ranging in the mid-60's.
- With high reratings, these players (a large portion of each draft class) will not be utilized.
- With low reratings, some of these players (with superstar dev trait) will quickly overtake/surplant some of the star players as they begin to regress within 2-3 yrs.
- 1st/2nd/3rd round players that would/should get playing time (AI wise) are pushed further down on the depth chart, ultimately regressing faster due to less playing time and not meeting gameplay goals.

2) Contracts and Trades
- With high reratings, players from default roster will negatively impact trades due to the built in trade values structured within a players OVR rating.
- As player OVR increases, contracts will increase causing the AI to avoid many of these overvalued player/contracts.
- Unlike the Human player, the AI may hold on to these players/contracts (AI to AI trades mainly), no matter their OVR, while cutting better players/contracts in order to meet 53 man roster maximums.

3) Regression/Retirement
- Highly rated rosters will have players filling roster spots (#2 above) due to larger than normal ranges for ratings to regress from, while their OVR still remains higher than in coming players from the draft.
- Higher OVR's may delay retirement, even if they're lower on the depth chart.

4) Player Disparity
- Draft Classes produce very few players above 80 OVR (again see above for OVR info) as a way to create separation between perceived top players of each class.
- Default and High rerated rosters narrow the disparity by having many players grouped closely together.

B) Specific Ratings not in-line with Draft Classes, that will change "Gameplay"
1) Play Recognition (DL/LB/Safeties ONLY)
- Draft Classes generate up to 80 for these positions, but default roster(s) has it lowered.
- Since draft classes can come in at 80, editing established players (> 2-3 yrs experience) for these positions to be above 80 will greatly increase "Defensive Challenge" by AI unit as a whole.


So much more, will update as I come across information.
Hope this info helps.
I'll try and answer any questions, but will not go into great detail of what's there.

[bcruise]

Interesting stuff here, very detailed and something that seems like it would be a whole lot of trouble to make up. I can definitely believe the stuff about the difficulty levels and sliders, given how unbalanced Rookie and All-Madden are.

A few comments:

It's entirely possible that some of what was found in the code isn't actually active in it anymore. So even if all of this is actually still in the code, there's no guarantee that it's ALL having an effect on gameplay. That said....

1)

Quote:

4) Sliders shift/alter the ratings +/- that they are meant to impact (ie WR Catching = Catch, Spec Catch, CIT, Route Run)

So WR Catching affects Route Run...good to know. Can you break down all of the sliders like this? Or, if not, at least the Pass Coverage/Pass Reaction as those are the most baffling of all of them. I do understand that this could potentially step on the toes of slider makers a bit, but it really is important info and not something EA has been particularly forthcoming with. At the end of the day I think we all just want a better playing game, and understanding how the sliders work is fundamental to that. Would love to hear Armor/Josh/King et al's take on that.

2)

Quote:

6) Awareness Functions (has nothing to do with AI awareness, might be miss labeled) - Is the actual OVR of a player (might be mislabled) - the higher this is, the higher the #5 is. - Players in Draft Classes depth chart positioning are in-line with this rating (ie the higher the rating, the higher they are on the depth chart) - Modifier of all ratings (Players will have their high ratings modified by having low AWR) - as it increases the closer to that rating the AI controlled player will function (ie Catch = 85/AWR = 40, player will drop passes even though the rating is high)

I find this fascinating, and believe it completely. Fits right in line with that old chart we saw in the past about how ratings were weighted, and AWR was decently high on almost every position's list:

http://fivethirtyeight.com/features/madden/# (a ways down the page)

So do you have any idea what AWR rating a player would have to have to play to a "true" 85 Catch, to use your example?

3) And, because I just have to ask....any evidence of scripting/comeback nonsense in there?


Thanks for the research.

[DerkontheOS]

Very interesting. I would like too know more.

I ****ing knew awareness matter for user controlled players.

[SkyWalker]

Wow. So awareness does play a role in whether or not a catch is successful. I wonder how awareness affects running backs. With high awareness, will they break more tackles?

[bcruise]

Quote: Originally Posted by SkyWalker

Wow. So awareness does play a role in whether or not a catch is successful. I wonder how awareness affects running backs. With high awareness, will they break more tackles?

If it truly modifies all ratings, then yes. It would be the most important rating in the game (across all positions), hands down. But since it's a modifier and doesn't really have an effect of its own, you'd still need high ratings in the other specific categories for it to mean anything.

Thinking about it like this puts a whole new spin on what I think about late-career progression. If a veteran's physical ratings are still high enough, increasing their awareness could have just as much or more of an effect as trying to maintain their individual "football skills" ratings. Thinking of awareness as "league experience" is a lot easier now, knowing this.

[Robo COP]

Quote: Originally Posted by bcruise

It's entirely possible that some of what was found in the code isn't actually active in it anymore. So even if all of this is actually still in the code, there's no guarantee that it's ALL having an effect on gameplay. That said....

I would assume the draft class info he posted would be included in this, correct?

[bcruise]

Quote: Originally Posted by Robo COP

I would assume the draft class info he posted would be included in this, correct?

Possibly. But with this line:

Quote:

- every time a new CFM is started, every player's stat (i.e. height, weight, build, ethnicity etc...) in that class is manipulated to produce a different player

I don't think it would matter if the draft classes are actually still pre-made - you probably wouldn't be able to figure out who was who.

In other words, even if this is true I don't think we have to worry about draft guides and such anymore. It still fits with what EA was telling us about the classes being randomized.

[BlackBetty15]

i posted this to the right dev twitter. Simply saying there might be some things even THEY dont know about their own game and to take a look. Who knows?