FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

[michiganfan8620]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

I'm not trying to make you change your mind or argue the point. I just was curious as to the reasoning behind it. But obviously I know that GMs will take certain guys off their board due to other factors. But you keep going back to the NFL draft...this is about Madden ratings though. Now I get what your saying, but you don't need to take the "if you don't like it go away attitude."

With that out of the way, what I am curious about is strictly what the scouts do. Staying on Ray Rice, for his suspension why would they change his grades/overall at all? Same thing with an injury.

Rather than explain it I have always fancied screenwriting so indulge me here. Hypothetical scenario: After Ray's incident Baltimore released him. Lets also imagine that Bernard's contract expired and the Ravens opted not to resign him. Team X is considering signing one of those players so the GM brings in his scout.

GM: Okay so what do you got for Rice?
Scout: Well we grade him overall at a 26.
GM: what? Yeah he had a bad year but he looked that bad?
Scout: Well no, but he is suspended so we lowered his final grade. Based on film he is a 72 but he beats his wife so we lowered him to factor in the risk.
GM: ...You are a scout. You scout football players. You report to me how good of a football player a guy is. We are aware of his history and myself and the coaches and ownership will take that into consideration. Your job is to report accurate data. How did you determine the 26 anyway?
Scout: Well we added a multiplier that would lower his final grade to a level we felt was sufficient and uh...
GM: It was arbitrary wasn't it? Nevermind, what about Bernard?
Scout: Ah we uh graded him a 32...
GM: why?
Scout: he is coming off surgery.
GM: So you think his ability to run between the tackles is diminished?
Scout: Nah, we didn't touch that. Actually he grades as a 62 but we dropped him.
GM: Because he is injured? And you are expecting a decrease in ability?
Scout: I don't know I am not a medical doctor or anything; we just dropped him to convey to you that...
GM: Get out of my office.
*end scene*

Of course I know that GMs do not care about overall rating like certain Madden players. It makes the whole damn thing pointless though. It is arbitrary and I have a hard time believing that professional scouts would manipulate data in this way. Going back to the draft, sure that may be different, but it is my opinion that abilities, injury concerns, and character should all be separate. If some front offices combine those categories to arrive at a final grade that is understandable but they do not arbitrarily drop the football grade.

I agree 100% with this

[ggsimmonds]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

One thing I should add, this is not about your Madden ratings per se and whether or not I would use them. You are a scientist so you should understand this better than most. You probably do not plan on going for a swim in lava flow but you still want to know how hot it is and how it works. I am not attacking or arguing your methods for rating guys. It is about the scouts. If you took it on yourself to change the rating to illustrate a player's circumstance that is different and understandable.

[DCEBB2001]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

One thing I should add, this is not about your Madden ratings per se and whether or not I would use them. You are a scientist so you should understand this better than most. You probably do not plan on going for a swim in lava flow but you still want to know how hot it is and how it works. I am not attacking or arguing your methods for rating guys. It is about the scouts. If you took it on yourself to change the rating to illustrate a player's circumstance that is different and understandable.

Hey man, I just interpolate the data based on what I was told about it when I started the project. If I could give you the contact info for the front office of the team or the person who supplies it, I would, and you could take this up with them. In your hypothetical situation, what is more likely to happen is that the GM would simply take the player off of their board. What would happen is the GM could see how the player graded individually in each skill then look at the data as to why his overall grade has been lowered. These guys are pretty smart; they know what to look for.

The real problem here is that Madden ONLY allows us to manage the OVR rating by the attributes. We can't "tag" a player as injured or suspended, leave his attributes alone, and then see his OVR drop as it is supposed to (like in the scouting data).

Now, I am quite tired of the back and forth on this and it is getting quite off topic. So, let's move on. The banter back and forth is not being productive to the topic. This looped discussion will not change how the scouts rate their players. It will not change how I use their data. So, logically, all of this is for not. Let's get the thread back on topic please.

[DCEBB2001]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

Hey Dan, have you ever considered also including traits in your system? I would love to know based on real data what those traits should be. Throws tight spiral, how often they take off running, pocket presence, big hitter.

I have a TON of trait data. The problem is it is all written out in text and not assigned a value. Multiply each report by about 70,000 players and you have a TON of work to do, which requires a ton of time...time I do not have.

[A_Quiet_Pro]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

I could be wrong here, but my understanding is that there's like an algorithm that more accurately portrays the existing Madden ratings for players and teams. Would it be possible to just take a spreadsheet of the Madden ratings (spreadsheets are available once the ratings are released I believe), but take the spreadsheet, and feed that info into the algorithms and allow the computer to change everything over? Maybe it's not as easy as I think it is. (OR Maybe EA can hire Dan and have him work with Donny lol).

[DCEBB2001]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

I could be wrong here, but my understanding is that there's like an algorithm that more accurately portrays the existing Madden ratings for players and teams. Would it be possible to just take a spreadsheet of the Madden ratings (spreadsheets are available once the ratings are released I believe), but take the spreadsheet, and feed that info into the algorithms and allow the computer to change everything over? Maybe it's not as easy as I think it is. (OR Maybe EA can hire Dan and have him work with Donny lol).

The problem is that the scouting data references different criteria (attributes) and formulas (the OVR calculation formula) to reach an OVR rating versus that in Madden. In fact, there are some attributes that either don't exist in Madden or Madden has some attributes that don't exist in the scouting data. For instance, there is no "balance" rating in Madden (which is a key attribute in the scouting data).

So the problem arises when converting one set of data to the other. Neither is mutually exclusive and exhaustive. I can pick out criteria that matches Madden and find attributes in Madden that match the scouting criteria, but there is not a 100% match amongst all criteria for all positions (like the "Clutch" grade for K's and the "Release In Time" grade for QBs). So, we have to ignore some and include others.

In addition, the scouts value certain criteria very differently than Madden does. Madden places a multiplier of 0.329 on THP, but the scouts only have a similar multiplier (or significance, if you will) of 0.16. That means that Madden values THP twice as much more than what the scouts value "Arm Strength" for the QB position. This automatically throws the equations off.

What EA needs to do is get a real scouting department involved that uses real data and real formulas. Then, they need to overhaul the ratings system based on the scouting variables and OVR scores. Only then will you see something close to 100% realism. FBG Ratings can use real data to interpolate Madden ratings based upon it, but as I said earlier, it won't match the scouting data 100%. If someone like me was brought in, we would have to, first and foremost, revamp the entire way that players are graded and rated. We can still use attributes and OVR ratings as forms of measure to differentiate between players, but we would have to make it mimic the data source.

So one must ask themselves, what data sources are EA-Tiburon/Donny Moore using and how well do they actually correlate to the Madden ratings we see in the game? From what I have heard they do a lot of youtube scouting, look at simple stats, and even use some PFF data (which I have pointed out in several topics are nearly useless for determining attributes, but are great at determining traits).

[DCEBB2001]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

Whats up with the website right now? I noticed a bunch of players OVR calculations went way down compared to where they were a week ago, except a few players stayed the same. I'm not expecting 99 OVR guys, but seeing the best player at a position rated at 80 confuses me. I see Calvin Johnson at 97, Fitz at 92, and then AJ Green at 81, whats up with that? Did the OVR calculations change or something? I don't think there is that big of a gap between WR's. Same with LOLB, Tamba Hali is rated 97, the next best is Suggs with a 83. I don't think Hali is even the best LOLB in the league, let alone being 14 OVR higher than the next best guy.

I have a few people testing the Equal-Interval ratings method right now so they are producing test rosters. Based on PLAYMAKERS' findings from last year, these may give even better results. They also mimic the scouting data at a perfect 1:1 ratio as the mean-rating has been removed from every rating category. This method allows you to see how the scouts rated the players on a scale that is unbiased toward the mean. PLAYMAKERS posted some results on this last year in this thread with some vids as well. This is as close to the real distribution as I can get within Madden's OVR rating formulas. You will see way fewer 90+ ratings for attributes as originally intended. This is the system we ran last year to great success, albeit much debate.

[michiganfan8620]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

I have a few people testing the Equal-Interval ratings method right now so they are producing test rosters. Based on PLAYMAKERS' findings from last year, these may give even better results. They also mimic the scouting data at a perfect 1:1 ratio as the mean-rating has been removed from every rating category. This method allows you to see how the scouts rated the players on a scale that is unbiased toward the mean. PLAYMAKERS posted some results on this last year in this thread with some vids as well. This is as close to the real distribution as I can get within Madden's OVR rating formulas. You will see way fewer 90+ ratings for attributes as originally intended. This is the system we ran last year to great success, albeit much debate.

Gotcha, I understand that, just kind of confused about the few outliers in the data, such as Hali being rated so much higher than everybody else. Also, just wondering, why do so many WR's have 99 juke and 99 spin moves? Not sure if that's an error or done intentionally.

[DCEBB2001]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

Gotcha, I understand that, just kind of confused about the few outliers in the data, such as Hali being rated so much higher than everybody else. Also, just wondering, why do so many WR's have 99 juke and 99 spin moves? Not sure if that's an error or done intentionally.

All of the players are rated on the same scale. An outlier would be three standard deviations about the mean. All players, up to an OVR of 99, are within three standard deviations of that mean.

Once again, you must consider these ratings HOLISTICALLY. What are the SPD, AGI, and ACC ratings for those players with 99 JKM and SPM? Trust me, a player with an AGI of 70 will do a juke move a bit differently than a player with an AGI of 90, even if both have a JKM of 99. Think about the ratings as they work SYMBIOTICALLY for a player.

[DCEBB2001]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

Well then, I'm thrown off big time, I don't understand how somebody like Tamba Hali can be rated 14 points above the next best LB, when some people don't think he is even the best LB on his team. Of his 11 sacks last year, almost half were in two games. Same with Fitzgerald, he is definitely on the decline. He is still a great player, but he isn't 11 points above Green, Marshall, Jones, Jeffery, etc.

And about the 99 juke and spin move thing, LeSean McCoy, one of the shiftiest backs in the league, has a 81 agility and 74 juke move. Larry Fitzgerald, not known for making moves after the catch, has 78 agility and 99 juke move. That isn't right, and there is no way around that.

The great thing about Madden is that you can personalize your rosters to how you see fit. If you disagree with something, just change it in the game. Works for you, works for me. I feel like I am on the hamster wheel of hell with you on a lot of this stuff, so from this post forward I will cease continuing to address it. Everything you need to know, as far as an answer, lies on our FAQ page. Please refer to our FAQ page.

[DCEBB2001]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

Another thing, you have Clowney as a 75 in man coverage and a 85 in zone coverage? Based on what? He has never been used in coverage other than an occasional zone blitz used maybe twice in a game. In comparison to that, Joe Haden, one of the best CB's in the league gets a 74 man coverage and 77 zone? How is that even remotely plausible?

Please refer to our FAQ page.

[DCEBB2001]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

I'm just trying to make your ratings site better. I don't see anywhere in the FAQ page where it explains why a DE/OLB who has never played in coverage before should be rated higher in coverage ratings than one of the top cover corners in the league, nor do I see where it explains why the shiftiest back in the league is rated worse at juking, agility, spinning, etc than most WR's, including some older ones who have lost a step over the years. I'm being reasonable here, but you don't want to take the time to explain these things. I am looking at the big picture, and I see glaring errors that I am trying to help you fix. I just don't see how these ratings make sense.

However, if you don't want my advice, so be it. I was just trying to help you get the best ratings for realistic gameplay.

The FAQ page states quite clearly that we use data to drive the ratings. If the data doesn't warrant a particular rating, then we don't give it out. You must think big picture as no rating works statically. No attribute in player analysis works independently of other traits. McCoy has a significant edge in both AGI and ACC (both of which are needed for employing sharp SPMs and JKMs). You must consider how ratings affect eachother on they playing field. Search for some of PLAYMAKERS videos to see this. He has a youtube channel dedicated to it.

As for Clowney v. Haden, you must, once again, consider all of the other ratings. There is a direct positive correlation between AGI and MCV ability in the data, so much so, that the R-Squared value is 98%. 98%!!! Look at their agility ratings. Now look at who they have to cover in such positions. It is highly unlikely that Clowney would be used to cover the deep 3rd or be asked to man up against a WR. However, the data suggests that when he WAS asked to drop into coverage (and he will be doing some of that in Houston), he was very effective against the type of players he was asked to cover and the types of zones he was responsible for. If I didn't have data for Clowney's Zone and Man coverage abilities, it would be a 0.0 on the scale, but it isn't. Also keep in mind that much of this, for rookies, is based upon PROJECTION. If he struggles in pass coverage (which he will be asked to do at least a few times this year), then his ratings in those categories will drop.

So in essence, each attribute is relative; relative to other attributes and relative to the competition one is facing on the field. After all, we don't expect Danny Amendola to hold up as well against a NT in run-blocking as he would a CB, would we?

[michiganfan8620]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

The FAQ page states quite clearly that we use data to drive the ratings. If the data doesn't warrant a particular rating, then we don't give it out. You must think big picture as no rating works statically. No attribute in player analysis works independently of other traits. McCoy has a significant edge in both AGI and ACC (both of which are needed for employing sharp SPMs and JKMs). You must consider how ratings affect eachother on they playing field. Search for some of PLAYMAKERS videos to see this. He has a youtube channel dedicated to it.

As for Clowney v. Haden, you must, once again, consider all of the other ratings. There is a direct positive correlation between AGI and MCV ability in the data, so much so, that the R-Squared value is 98%. 98%!!! Look at their agility ratings. Now look at who they have to cover in such positions. It is highly unlikely that Clowney would be used to cover the deep 3rd or be asked to man up against a WR. However, the data suggests that when he WAS asked to drop into coverage (and he will be doing some of that in Houston), he was very effective against the type of players he was asked to cover and the types of zones he was responsible for. If I didn't have data for Clowney's Zone and Man coverage abilities, it would be a 0.0 on the scale, but it isn't. Also keep in mind that much of this, for rookies, is based upon PROJECTION. If he struggles in pass coverage (which he will be asked to do at least a few times this year), then his ratings in those categories will drop.

So in essence, each attribute is relative; relative to other attributes and relative to the competition one is facing on the field. After all, we don't expect Danny Amendola to hold up as well against a NT in run-blocking as he would a CB, would we?

Sorry I asked then. I'm going to just delete my posts from this thread tonight, and we can both move on.

[garret2]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

dcebb2001 have you ever heard of anyone doing an in-depth ratings project like this for the NBA? you may have inspired some people to try!

[WingWest36]

Re: FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

I recently adjusted the Bears roster ratings and did a quick mirror match. I enjoy the rating scale FBG has provided. I can clearly tell the difference between Brandon Marshall and Alshon Jeffery. Marshall actually has more yards after contact and Jeffery's speed is noticeable as well. Only downside I see is the time involved adjusting ratings for the teams (obviously lol). I cannot wait to see how M15's gameplay with the FBG rating scale will look and perform!