FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

[DCEBB2001]

Quote: Originally Posted by shagrugg

You've probably already answered this somewhere in this thread... but from your previous post ^^ If injuries affect the OVR, why wouldn't statistical achievements or stat trends? I get the scouting data as a baseline of physical attributes coming out of college... but over time that should become less a contributor as there is statistical data seasonal achievement and career trends up and down that should contribute to OVR.

I do not believe in using stats in figuring out the OVR of a player. If I did, I would set the person with the highest statistical data to 99 every year. In this system, that is reserved for players who are playing at an all-time level. See Brady of 2007, Rodgers of 2011, Rice of 1995, etc. Also, stats lie in my opinion. Statistics do not always tell the entire story. Sometimes great players have bad days. Sometimes JAGs (Just a Guy) have great games...one hit wonders if you will. Scouting data is always being updated and compiled. Why do you think pro scouting departments for NFL franchises are always bringing in players for Tryout Tuesdays? They don't just scout college talent you know.

I prefer real NFL scouting data because it shows the ability of a player. Stats are the product of TEAM ability...the end result of a joint effort. That RB with a 20yd TD may have broken 6 tackles to get there with terrible blocking. That is great individual effort by the RB. Then again, a RB may go for the same 20yd TD without even getting touched. Perhaps the OL and play calling needs more credit for a run where all the RB had to do was run straight at 75% of full speed for 20 yards. If you looked at the stats line, you wouldn't know the difference as it would just show up as a 20 yard run. There is no reasoning for how the run occurred and no in-depth analysis. Scouting data for individual players does that.

[DCEBB2001]

Quote: Originally Posted by luisffsilva

Pure opinion ??! I don't think so ! you see nfl games or just play videogames and do scouting !?? I think alot of users agree with me. Adrian Peterson is behind Jones Drew and Steve Jackson ?!? Fullback best rating is 86 ?!?? See a NFL game, than talk... Scouting is better than stats, OK, i agree but they are REAL numbers. Like Todd Heap, he is third TE in Cardinals depth chart and the rating is 91 ?!? He never was a PLAYMAKER and lost alot of games with injuries. Now numbers: Carrer Todd Heap RATING 91 years pro 11 REC 480 YDS 5,642 TD 41 Jason Witten RATING 90 years pro 9 REC 673 YDS 7,680 TD 41

I do watch NFL games, but I rely on scouting data for this project. It is unbiased. Peterson is behind those other RBs because he is not a complete player. He doesn't even play on 3rd down. If you would watch games, you would know THIS. He is a poor blocker. Jackson and Jones-Drew are willing blockers and good receivers. They are complete backs. Peterson is not a complete back, hence his OVR being limited.

You don't have to be a play-maker to be a great player. Heap has great AWR and is a solid route runner. Is his athleticism declining, yes. But he still has good hands, is a strong blocker, and runs good routes. Just so happens that those are important ratings for evaluating TEs so why wouldn't he be rated well? Depth charts lie as well. I have been in coaching since 2005 as both a HS assistant and a GA, and I can tell your right now that coaches tailor the depth chart according to what they do best. I have seen several players who may have been more complete, better OVR players get buried on a depth chart because they didn't fit as good into a system. This also happens when grooming young talent - staring an inferior player with less experience over a veteran in the hopes of GETTING that young player more experience. Happens all the time. Do you really think that Ponder up in MN is a more complete player than McNabb? Do you think that he can read a defense better or identify a rush better at this point in his career?

The point is that stats and numbers offer less-complete data IMO. They never tell the whole story. I consider scouting data to be more complete for several reasons. If you disagree, then don't use these ratings. Use Donny Moore's. After all, if a no-name player has a good game, Donny will just increase his SPD, AGI, and ACC by 10 and call it a day without actually doing any heavy analysis. Perhaps that would fit your style a bit better.

Oh and if you are going to argue a point with someone who does not believe in statistical numbers, please bring something that he is willing to believe in. I don't see any blocking data in those numbers; a large part of the contributing OVR value.

[Culture Rot]

So is this project geared toward producing more accurate sim stats, or more realistic on the field gameplay?

Both is not an acceptable answer btw

[DCEBB2001]

Quote: Originally Posted by xlatinoheatx

So is this project geared toward producing more accurate sim stats, or more realistic on the field gameplay? Both is not an acceptable answer btw

Why is both not an acceptable answer?

This is a gameplay adjustment. There are some testimonies scattered in the thread regarding the gameplay should you want to know how it plays out in-game.

However, the primary goal of the project is to incorporate real data into the Madden ratings process with less opinion and conjecture, but more objective data. This data is then used to diffuse rating inflation.

[shagrugg]

Quote: Originally Posted by DCEBB2001

Why is both not an acceptable answer? This is a gameplay adjustment. There are some testimonies scattered in the thread regarding the gameplay should you want to know how it plays out in-game. However, the primary goal of the project is to incorporate real data into the Madden ratings process with less opinion and conjecture, but more objective data. This data is then used to diffuse rating inflation.

Incorporating real data, except all NFL statistics... seems to be a ratings process geared more towards opinion and conjecture, as opposed to actual data... excluding data which is more telling of NFL gameplay.

Whether or not the gameplay is improved is purely opinion i.e. testimonies... is there any statistical evidence to prove that game play is better?

From what I understand of your methods... your site is a good place to get ratings for guys who have little to no NFL experience, for guys EA does not have ratings for.

However, for anyone who has established careers or statistical significance at the NFL level, ratings for those players should be taken with a grain of salt, as the ratings are based on physical attribute data that is in many cases several seasons removed from college scouting data.

[DCEBB2001]

Quote: Originally Posted by shagrugg

Incorporating real data, except all NFL statistics... seems to be a ratings process geared more towards opinion and conjecture, as opposed to actual data... excluding data which is more telling of NFL gameplay. Whether or not the gameplay is improved is purely opinion i.e. testimonies... is there any statistical evidence to prove that game play is better? From what I understand of your methods... your site is a good place to get ratings for guys who have little to no NFL experience, for guys EA does not have ratings for. However, for anyone who has established careers or statistical significance at the NFL level, ratings for those players should be taken with a grain of salt, as the ratings are based on physical attribute data that is in many cases several seasons removed from college scouting data.

1. NFL statistics never tell the full story. They never fully explain how good a player is at a particular skillset. Do the typical NFL statistics measure how good a WR can block downfield? Do they measure how good he is at catching a pass under duress by a defender? Scouting data based on watching the behavioral patterns of a player in practice and game situations can give any scout an idea as to how good a player performs in a given situation. From that, some correlations can be made.

2. Statistical evidence is elusive as gameplay seems to be pure opinion. I only hear about the gameplay from those who use the ratings, so you would have to read what they say to get any viable impression or play the game yourself. Neither case is as objective as one may desire, but it is likely all we have at this point.

3. Statistics are the product of a team game. You can't have a completed downfield pass without a line to protect your QB. You can't have a run for a large gain without some level of blocking. It is nearly impossible for a LB to make a tackle up against 5 OL-men without them occupying some blocks. The point is that in a team game where the outcome of a play is determined upon a near infinite number of factors coming together in any more infinite variety of outcomes, we have to be able to single out the parts that total to the sum. How well did the OL pass protect? How well did the QB read the coverage? Did the WR utilize pure athletic ability to get open or rely on reading the coverage and altering his route to open a window in the defense? Did the defenders get to the ball before the receiver did? Did the receiver get his hands on the ball while under duress? Did he catch it? Did he break a tackle to get more YAC? Was the defender bearing down on him taking a poor angle? Is he a poor open field tackler?

As you can see, all of these factors may lead up to a gain of 40 yards downfield on any given play, but it may have also resulted in a sack 1.5s after the center snaps the ball. The point is that in a team game, statistics do not fully measure everything. They do not tell the whole story. Scouting data gets us a bit closer to that.

4. The physical attribute data I use is derived from a combination of measurables from combines and pro days, but it is also derived from objective scouting data that can be used to measure things not considered at a combine. The "Raw" data can be universal to correlate (as opposed to determine) the SPD, ACC, AGI, JMP, and STR of a player. However, the real meat of the data is in the scouting information which is updated weekly. The other 40+ attributes not covered by the Raw attribute data is covered by this scouting information for NFL players. So in essence, only 5 of the 50+ attributes are determined by "physical attribute data that is in many cases several seasons removed from college scouting data" as you put it.

Even then, that RAW information is altered by the scouting data and changed as more information is gathered, so to say that the site is only good for new players that EA does not currently have in their rosters is incorrect. The scouting data I use is not only from the college scouting information, but also from the professional scouting information I have acquired. This data, like I said, changes after every NFL week completed and every month in the offseason, so the wealth of information is quite substantial.

[Big FN Deal]

Of course this general topic is opinionated and open for discussion but I agree with DCEBB. I was watching the Steelers v Chiefs and the commentators mentioned how when people say to Hines Ward his speed has declined, he says "I was never fast to begin with", lol. The point being, the notion of great statistical performance equaling superior physical traits is sometimes popular perception, not fact.

This is always a difficult topic partly because of the mystery surrounding the exact meaning of Madden player ratings. I personally believe there should be far more individual rating categories but they all should be hidden and working "under the hood". Showing gamers the inner workings of how the program processes on field performance, is like watching a puppet show with neon strings and exposed puppet masters but I digress.

We all know player ratings is not an exact science but it makes sense to use a quantifiable formula with the least amount of variables.

[luisffsilva]

How this work in Madden 12 ?!?
Because players like DE and LB don`t have a option to put ratings like:

Throw accuracy, throw on run and others.