FBGRatings Seeks to Recalibrate Madden Ratings, Change Game

[Chris Hitchens]

probably just a publicity stunt.

[DCEBB2001]

Quote: Originally Posted by Chris Hitchens

probably just a publicity stunt.

How do you figure that?

[Hooe]

I've taken up the task of building a roster file with the ratings from the website on PS4 (charter04 is doing the same for XBOX One). I've played a couple games, and I've been writing up some detailed impressions to DCEBB as I play the games to give him some feedback.

Going forward, I'm going to cross-post those impressions as to my experiences with the FBGratings player re-rates on an individual game basis here. All my games played will be Play Now games against the CPU, 10 minute quarters with accelerated clock run-off to 15 seconds; slider changes will only be made as deemed necessary. I'll be adding to this thread shortly impressions of my first two games with these ratings.

Given that I'm on a PS4, I'll also see about broadcasting to Twitch going forward to let others watch the games I'm playing to get a glimpse for themselves as to what I'm seeing. Unfortunately, I did not do this for the first two games I played.

EDIT: just to preface my impressions - the game I'm playing is still Madden, and player interactions in the game still are bound by the rules defined by Tiburon; i.e. OL/DL pass blocking interactions are still very much win/loss affairs, the re-rates won't add LBs defending assigned gaps if that logic doesn't already exist in the game, etc. Rather, what the re-rates are attempting to accomplish is get player behavior within Tiburon's system producing more realistic results, under FBGratings' premise that consistently assigning all the players' ratings based on empirical data will yield a better-playing game that the media-influenced ratings Tiburon provides.

[Hooe]

Game 1 - Cowboys (me) v Eagles (CPU)
All-Pro difficulty, default sliders across the board
10-minute quarters, accelerated clock to 15 seconds

Cowboys 24 - Eagles 17

Tony Romo - 20/26, 244 yds, 2 TD, long of 44
Mark Sanchez - 17/19, 157 yards, TD, long of 44
Nick Foles - 1/1, 12 yards

Demarco Murray - 17 attempts, 64 yards, 11 after first hit, long of 12
Darren Sproles - 7 attempts, 41 yards, TD, 4 after first hit, long of 26
LeSean McCoy - 3 attempts, 6 yards
Lance Dunbar - 2 attempts, 10 yards, fumble
Nick Foles - 1 attempt, 9 yards

Dez Bryant - 8 catches, 158 yards, 31 YAC, long of 44
Riley Cooper - 5 catches, 79 yards, TD, 20 YAC, long of 44
Demarco Murray - 5 catches, 29 yards, 36 YAC, long of 17
Jeremy Maclin - 4 catches, 46 yards, 6 YAC, long of 15
Darren Sproles - 3 catches, 2 yards, 9 YAC, long of 4
Jason Witten - 3 catches, 38 yards, 2 TDs, 4 YAC, long of 25
Terrance Williams - 3 catches, 7 yards, TD, 1 YAC, long of 4
Jordan Matthews - 2 catches, 18 yards, 6 YAC, long of 14
LeSean McCoy - 2 catches, 23 yards, 12 YAC, long of 12
Josh Huff - catch, 2 yards, 2 drops
Cole Beasley - catch, 12 yards, 2 drops
Gavin Escobar - drop

George Selvie - sack, 2 TFLs, fumble forced
Connor Barwin - sack, 3 TFLs
Fletcher Cox - sack, TFL
Sean Lee - sack, 2 TFLs
Malcolm Jenkins - pass deflection

IMPRESSIONS - QBs still felt too accurate on default settings - typical of Madden - though Sanchez was content to throw dippy short passes most of the game once Foles and McCoy left due to injury (both left in the first quarter). Sanchez did go down field in the 4th quarter with success, however. Coverage never felt very tight for either side, most of the incompletions were dropped passes. Sanchez's 44-yard pass was the result of a missed jam by Brandon Carr (PRS 87) against Riley Cooper (RLS 97) who then couldn't win a footrace down the sideline. Clearly FBGratings has graded Mark Sanchez as not a very accurate passer (TAS 76, TAM 70, TAD 68), yet he was pretty solid in this game. This will likely require a slider adjustment to correct, but I anticipated this.

The players with high ratings looked and felt very good; Bryant, Free, Witten, Sproles, Lee, Ryans, Barwin all stood out easily. Kickoff length seemed very short compared to today's touchback-heavy NFL, but that's a simple slider adjustment. Dez Bryant in particular was really fun to use; he kept getting open and was able to break some tackles after the catch to get further downfield.

The one thing that jumped out at me about player movement was that it felt harder to break away for a game-breaking play. I think this is by design with the ratings, though; Bryant isn't graded so much as a burner (80 SPD, 86 ACC) as much as he is a physically dominating receiver who can be dangerous after the catch (79 STR, 74 SFA, 76 TRK, 82 ELU, 93 SPM, 95 JKM, 94 BCV) and everyone knows Jason Witten isn't running away from anyone (78 SPD, 76 ACC). Dallas in general doesn't really have a "home-run hitter" sorta player. Putting drives together to move the ball was important.

The other phases of the game to me felt really good. I could tell that Ron Leary was a weak link in the Cowboys' line in all phases, while Doug Free was a bit of a monster to run behind and Tyron Smith was trustworthy in blindside protection. I'm not sure I've ever noticed the differentiation in OL skill in Madden so obviously before. At least some of that is the new advertised OL logic, admittedly; I admittedly don't have a large breadth of experience playing Gen-8 Madden yet. Melton wasn't consistently a force, but with his high athletic ratings (SPD 76, STR 86, AGI 76, ACC 80), he was able to really blow up a play a time or two.

Other than the pass accuracy being too high - which again, is typical Madden - this is one of the best experiences I've had playing a Madden game out of the box with no settings adjustments. Encouraging so far, but I obviously want to play more games against other teams to get more data to this end. Probably the first thing I will do is reduce QB accuracy to 25 for both User and CPU before my next game. Everything in general felt a little more measured than a typical Madden game, for lack of a better phrase for it. I'm a fan.

[Hooe]

Game 2: Redskins (me) @ New York Giants (CPU)
All-Pro difficulty, default sliders across the board baseline
HUM and CPU QB Accuracy slider at 25
10 minute quarters, accelerated clock to 15 seconds

Redskins 26 - Giants 20

Robert Griffin III - 17/25, 179 yards, 2 INT, long of 27
Curtis Painter - 9/22, 111 yards, TD, 3 INT, long of 22
Ryan Nassib - 10/15, 91 yards, 2 INT, long of 15

Alfred Morris - 16 attempts, 54 yards, 14 after first hit, long of 11
Roy Helu Jr. - 9 attempts, 42 yards, 6 after first hit, long of 15
Rashad Jennings - 6 attempts, 18 yards, TD, 2 after first hit, long of 10
Robert Griffin III - 4 attempts, 15 yards, TD, 2 after first hit, long of 11

Andre Roberts - 6 catches, 71 yards, 10 YAC, long of 27, drop
Desean Jackson - 5 catches, 64 yards, 7 YAC, long of 21, drop
Mario Manningham - 4 catches, 41 yards, 26 YAC, long of 15
Reuben Randle - 3 catches, 39 yards, 9 YAC, long of 19, 2 drops
Victor Cruz - 3 catches, 38 yards, TD, -2 YAC, long of 22

Jon Beason - 7 tackles, TFL
Ryan Kerrigan - 2 tackles, 2 TFLs, sack
Stephen Bowen - tackle, TFL, sack
Cullen Jenkins - 4 tackles
Prince Amukamara - 6 tackles
Stevie Brown - 5 tackles, INT
DeAngelo Hall - 4 tackles, INT
Nat Berhe - 4 tackles, TFL, INT
Jose Gumbs - INT
David Amerson - 5 tackles, 3 INTs

IMPRESSIONS - This was a sloppy game to play overall. The Giants dug themselves a hole early with some rather bonehead throws into obvious coverage by Ryan Nassib (he has the "Trigger Happy" player trait), which led to easy interceptions. He left the game partway through with an injury but came back. In the meantime, Curtis Painter's accuracy issues (TAS 71, TAM 62, TAD 52) were readily apparent, though inconsistent; sometimes he would throw darts and sometimes his passes would sail. His inaccuracy was most obvious when he was under pressure (he has the "Paranoid" DPP trait, and TOR of 54). An improvement over the previous game I played for sure, however, where Mark Sanchez pretty much couldn't miss. RGIII's low mid-range accuracy rating (66 TAM) appeared to manifest itself in receivers having to come back to passes more often than not, which led to longer down-and-distance situations and a few more failed 3rd down conversions than the previous game. The INT I threw to Stevie Brown was user error where I picked the wrong receiver to throw to and he obviously wasn't open.

The opening quarter of the game also featured a fumble returned for a touchdown after Brandon Meriweather destroyed an effectively-stationary and dead-to-rights Henry Hynoski (CAR 68) in the flat.

The Giants were content to let me keep putting the ball into Morris' hands on the read option, they stayed at home on RGIII far more often than not and I never attempted to force the issue. Griffin's TD run came off of a playaction bootleg where he just beat the defense to the corner without much of a challenge after a long drive.

I don't normally sub heavily in Madden, really only doing so as need as the formation sub interface is clunky, but by the end of this game the fatigue of my players on the offensive side of the ball was obvious. I struggled in the later stages of the game to gain much on the ground with Morris and had to yank him for Roy Helu at times. Going forward I want to try subbing players earlier and more often to see if I can have a more effective team in the 4th quarter, as this game was a bit of a struggle to close out after I got off to a hot 13-0 start. The fatigue was also obvious on the defensive front; when I used the DL sub package in the 4th quarter (having not really subbed at all in the game before that point), the fresh players did have a burst on the ball my starters simply didn't have at that point. Nothing which struck me as unrealistic - in fact, I think the toning down of player ACC ratings on both the OL and DL helps out the play there a lot, I don't see any unusual bursts - but a difference was evident.

One comment I will make about running plays is that, I'm guessing on account of the general rise in defenders' AWR and PRC ratings, draw plays are consistently tough to succeed with, and screen plays can also be hit or miss; I struggle to even get back to the line with draws, and screen passes certainly don't feel like the easy yards they were at times on XBOX 360 Madden 25 on account of a more active and smarter defense. When there's a particularly athletic DL in the mix - such as Henry Melton for the Cowboys (not in this game, I was just running plays in practice mode afterward) - the issue is exacerbated a bit.

It was readily apparent who on my defense was really bad at coverage, particularly man coverage where guys with low MCV ratings flat-out would wait sometimes as much as (eye-balling) about a third of a second before reacting to a receiver's cut. The more balanced / less extreme SPD / ACC ratings FBGratings provides also seem to either mitigate or eliminate weird boosts on the ball downfield players sometimes get towards a ball when tracking a pass; I didn't feel gipped out of anything on the two INTs I threw and I didn't ever feel like the AI for my team or the CPU's was cheating.

I remain happy with what I'm seeing from the game with these rosters, and in particular I'm impressed by how few slider adjustments I've had to make to get competitive games. Again, the only slider I've touched is pass accuracy (which I think I may have overdone, but I need to actually play against a good QB for a whole game, they keep getting hurt!) and I don't really see the need to touch anything else right now. That said, I probably will mess with the auto-sub sliders because my curiosity about fatigue is piqued.

[Hooe]

Game 3 - Eagles (me) vs Cowboys
All-Pro difficulty, default sliders across the board baseline
HUM and CPU QB Accuracy slider at 35
10 minute quarters, accelerated clock to 15 seconds
Auto-Sub sliders: 90 out, 95 in for all defense

Tony Romo - 22/34, 176 yards, TD, INT, long of 29
Nick Foles - 15/22, 140 yards, 2 TDs, long of 28

LeSean McCoy - 14 carries, 104 yards, 23 after first hit, long of 56
Demarco Murray - 8 carries, 37 yards, TD, 5 after first hit, long of 25
Nick Foles - 8 carries, 40 yards, 8 after first hit, long of 11
Darren Sproles - 6 carries, 42 yards, TD, 9 after first hit, long of 14

Cole Beasley - 6 catches, 58 yards, 27 YAC, long of 29
Jason Witten - 5 catches, 43 yards, TD, 12 YAC, long of 16
Dez Bryant - 4 catches, 36 yards, 4 YAC, drop, long of 20
Jeremy Maclin - 4 catches, 49 yards, 10 YAC, long of 15
Terrance Williams - 3 catches, 27 yards, 9 YAC, 2 drops, long of 13
Riley Cooper - 3 catches, 55 yards, 18 YAC, long of 28

Sean Lee - 4 tackles
Rolando McClain - 4 tackles, TFL
J.J. Wilcox - 4 tackles
George Selvie - 3 tackles, 2 TFLs, sack
Bruce Carter - 3 tackles, TFL
Brandon Boykin - 4 tackles, fumble forced
Mychael Kendricks - 4 tackles
Demeco Ryans - 4 tackles
Nate Allen - 4 tackles
Casey Matthews - INT
Phillip Hunt - tackle, TFL, sack

IMPRESSIONS - this was a tight game up through the third quarter, with Dallas up 14-13 at half, but their defense just appeared to wear out by the end of the contest.

McCoy is a ton of fun to use, even with his speed reduced. I had to keep giving him the rock, but I eventually was rewarded with a 56-yard run in the fourth quarter on the game-sealing drive. I'm pretty sure on the default ratings this run would have been a touchdown, but again McCoy's top-end speed isn't that great with FBGratings (81 SPD, 87 ACC) so I wasn't really upset. He was obviously more quick than fast and was very effective evading tacklers.

Romo played reasonably well but never really challenged downfield. I was playing a lot of zone coverage on account of the Eagles' cornerbacks not having great ratings, which probably had something to do with the matter; would have been nice to see the Cowboys try to stretch the field more, though. He was obviously capable though; well, minus the one INT he literally threw right at me when I was playing on Casey Matthews in coverage.

The Eagles' OL played very well in pass blocking, with the exception of the Selvie sack I'm not sure that Foles was touched. Between that and a strong running game, I was able to have success throwing the ball without much issue.

This is mostly not a ratings concern but I'll say something anyway - I could be wrong, but it was almost like the adjustments to auto-subs affected the CPU team more than it did mine, which I did not expect at all. By the end of the game I was seeing plenty of Cowboys' reserves on the field for extended periods of time; Rolando McClain, Orie Lemon, Jeremy Mincey, Tyrone Crawford, Ben Bass, Ken Bishop, and Ahmad Dixon all logged significant minutes in the fourth quarter. There were some subs on the Eagles' side of the ball - I saw a lot of Philip Hunt in place of Trent Cole in the 4th quarter and Jason Phillips was in there for a few plays in place of Mychael Kendricks - but not nearly to the extent as what I saw from Dallas. I did win time of possession 23:00 to 17:00, which contributed certainly, but bottom line is that the changes I made to auto-sub sliders appeared to have a profound impact which I don't yet understand (I'm not a slider guru) and that likely swung the outcome a bit. I also assume that the reduced stamina ratings across the board contributed, as most of the defensive linemen's STA ratings are now in the low 70s.

I had 98 total offensive plays. Might be necessary to up the quarter length to get a realistic total amount of offensive plays (which I believe is supposed to be closer to 120 or 130?). These ratings so far appear to really reduce the number of big plays in the game; the 50+ yard run by McCoy is easily the longest play I've seen with these ratings in the three games I've logged (and one additional game which I did not record). This in turn results in more longer and sustained drives. Worth noting that I did not attempt to play at the uptempo no-huddle pace which the Eagles typically play at in real life.

[HC0023]

The Subs are seriously broken i would just leave them at default . Well atleast a minimum of 120 plays last year the most was 142 i believe but minimum 120

[Argooos]

Quick question: Does anyone know how much on average the ratings on Madden are inflated compared to the FBGRatings?

I'm creating a bunch of players based on their FBGRatings for my next franchise, but I plan to use them in a regular Madden roster. I plan to raise my created players ratings to match those of players on other teams (to make them competitive). I just wanted to see if anyone has done any work on comparing these ratings to the Madden ratings.