Classic Roster Ratings

[Iteachpercussion]

Alright folks, most of you on here know me and the classic rosters I have created over the last few years. Before I get to my question, let me give you some background on what I generally try to accomplish in my rosters.

1.I want to create a roster for a particular season which showcases each team and player as close to their real life counterparts as possible. I use season stats, film, understanding the Madden system, etc to create the rosters.

2. I want the roster to look right as well. I try to match ht and wt or body type as best as I can. I also try to get equipment as correct as possible. Though the limitations of Madden can be frustrating on that end.

3. I want to be able to simulate a season and have SIMILAR results to the real life season. Exact is not possible, and honestly no fun. I want good teams to play well, bad teams to play poorly, but as in real football anything should be able to happen. Also, sometimes in real life a bad team outperforms itself and vice versa.

4. I have tried to make rosters which allow users to play multiple seasons, but that is where the issue is.

Ok, so that is some background. Now I have been making rosters for Madden since before Madden Share was a thing. I made the rosters to start with just for myself. Once Madden Share was introduced I began putting the rosters out for everyone else, and they have been very popular. Over the years I have learned a great deal about the ratings, the sim engine, playbooks in Madden, and overall what works in game and what does not work. So on to the dilemma...

Ratings... There are a couple of things going on here. Madden's rating system is what it is and I do not want to debate the system or scale here, instead I just want to deal with the reality of what it is.

Quarterbacks: Madden deals with Qb ratings based on the modern NFL. As a result, high ratings produce high completion %s. This is especially true when simming games. This also works in played games. This makes perfect sense. However, for classic rosters this is a problem. If I am making Dan Marino, I want him to be good. Marino was an amazing QB, but his completion percentage was not in line with the modern NFL. If I make him with an OVR in the 90s, he will perform like a modern QB with an OVR in the 90s. People looking at the roster see Marino in the 90s OVR and are happy because that is what they expect. However, if I want the game to play out as a game from 1985, Marino should NOT complete 75% of his passes. He should be around 58% maybe low 60s. In order to accomplish that, I need to lower his ratings. However, then he ends up with an OVR in the 70s or low 80s. This concerns some people.

However, here is what I have found...
If I place a QBs accuracy ratings in the 70s and 80s for SAC and then down from there, the results are MUCH more true to life. This is for both sim and played games. Also, dropping AWR causes the CPU to sim more interceptions, which is a good thing.

What is the downside?
Well, if you are planning on playing multiple seasons in CFM, the CPU will draft QBs which have higher ratings than some of the good QBs in the game. So, before long, Marino is replaced by a rookie or 1 year veteran.
However, we have the ability to edit players, especially rookies, in CFM. So, I say, after each season you need to edit the rookies to be in line with the roster.

Other positions:
WR: in order to help the passing game and prevent too many guys from having really high numbers, most WRs need to be rated lower. There should be very few WRs in the 90s OVR.
TE: Except for a very few guys, most of them in the 80s and mid-90s should not be great receivers and therefore their OVR is lower than you might expect.
RB: This is a big one for simmed games. The Ball carrier vision must be lowered. This does a slight drop in the OVR, but does not affect gameplay much, but DRASTICALLY helps simmed games.
DB: These guys have the opposite problem of the offense. The coverage ratings need to be higher and the Press rating also needs to be higher. This helps in reducing simmed passing yards, but also helps played games by making passing windows smaller. Also, the game "feels" more right due to the lack of emphasis on flags being thrown in the 1980s for pass interference and such.

So, I know that was longwinded and I really only scratched the surface there. But my big question to this community is:
Would you rather have ratings that make the game play better and sim results better? OR Would you rather have ratings which mimic the current Madden system?

Please let me know your thoughts. I am willing to answer any questions I can about the ratings, results, etc... Like I said, I have been doing this for a while and I have learned a lot about how the game functions. I have discussed and used the findings of others in this community as well.

[rplatypus]

I prefer the player ratings to be in line with what EA puts out. The only reason is due to the incoming draft classes. It's not a real option for people to rerate every player in a draft class (Not possible anyway) to conform to a different ratings system. A different ratings system limits someone to only playing one season because the incoming rookies usually have better ratings then some pro bowlers if you use even a slightly lower rating scale than the official ratings. Sim stats are great and all but most people that sim do so over multiple seasons.

When I have created classic teams in the past, I used pro football reference AV value to guide me on player overall ratings. I then make 3-8 90+ overall players per team and work my way down to players around a 60 overall. I made a best of teams roster (89Sf, 95Dal, etc.) instead of a specific year which would have less 90+ overall players. Conforming to their ratings is the only way to currently make a classic roster that can be played for multiple seasons, at least until they decide to let us edit and share draft classes.

[Iteachpercussion]

Quote: Originally Posted by rplatypus

I prefer the player ratings to be in line with what EA puts out. The only reason is due to the incoming draft classes. It's not a real option for people to rerate every player in a draft class (Not possible anyway) to conform to a different ratings system. A different ratings system limits someone to only playing one season because the incoming rookies usually have better ratings then some pro bowlers if you use even a slightly lower rating scale than the official ratings. Sim stats are great and all but most people that sim do so over multiple seasons. When I have created classic teams in the past, I used pro football reference AV value to guide me on player overall ratings. I then make 3-8 90+ overall players per team and work my way down to players around a 60 overall. I made a best of teams roster (89Sf, 95Dal, etc.) instead of a specific year which would have less 90+ overall players. Conforming to their ratings is the only way to currently make a classic roster that can be played for multiple seasons, at least until they decide to let us edit and share draft classes.

I hear you and understand. I will say that though we cannot change the draft class ahead of time, we can change it after the fact, and it is not really that hard. It does take some time, but not too much.

Beyond the sim aspects, I think the whole game plays better with some slightly different ratings. The passing game especially works better with lower rated QBs and higher rated DBs.

My ratings have always been slightly off from Madden's ratings, because I do not like the number of guys in the 90s that Madden has. Also, I struggle between making a player the way he should play with his other ratings and conforming to the OVR formula. In other words, in order to Barry Sanders rated with a high OVR, I have to significantly inflate his Trucking, Stiff Arm, and Blocking skills. It is even worse for a power running back.

Christian Okoye is impossible to make a correct OVR for his 89 season because of the OVR formula. I can put his Trucking at 99, ball carrier vision at 99, but his OVR will be low until his Elusiveness comes up into the at least the mid-80s. Christian Okoye was not elusive. lol

I appreciate the feedback!

[capa]

It has always been my experience that you cannot create realistic classic rosters that reflect how players played and also have a realistic OVR. Like you said, the OVR formulas are ridiculous in my opinion. Most FBs wind up being 99 or so since you have to inflate their AWR, blocking and passing ability, to ensure they pick up the right guys.

I prefer skewed OVRs with players that look and perform how I remember them.

C

[Iteachpercussion]

Quote: Originally Posted by capa

It has always been my experience that you cannot create realistic classic rosters that reflect how players played and also have a realistic OVR. Like you said, the OVR formulas are ridiculous in my opinion. Most FBs wind up being 99 or so since you have to inflate their AWR, blocking and passing ability, to ensure they pick up the right guys. I prefer skewed OVRs with players that look and perform how I remember them. C

You are correct, the OVRs are not a really good indication of how a player plays. On my rosters I have tried to balance the two concepts. I know a lot of Madden players look solely at the OVR and if they are "off" by Madden standards they think the roster is not accurate or what not. So, I have tried to keep that in mind. Plus, if you allow the CPU to control teams in CFM, the OVR is the most important thing in terms of money, trades, draft, free agency and starters. None of that is important if you do a 32 team CFM and control those aspects.

I am leaning more towards ignoring the OVR for many positions in order to create the most realistic roster.

Thanks for the feedback! ANy additional info or words of wisdom are appreciated.

[rplatypus]

I didn't mean to say it's not possible to edit the draft class, I meant that it's not practical for the average user to edit an entire draft class after each season. Editing names, ages, appearance, college, ratings, etc. is pretty time consuming. Especially if you are trying to recreate a specific year. Adjusting contracts also takes forever. You can easily have multiple hours into editing a draft class that is only available in that particular franchise save. If a person is okay with a random class and just wants to keep the ratings in line with a specific roster, it obviously takes much less time.

I checked 3 random draft classes for an idea of the incoming draft class overalls on M17. It looked to me based on this small sample that there are around thirty to forty 75-79 overalls, thirty to forty 70-74 overalls and over a hundred 66-69 players coming into the league on average. If these results are typical for most users then I would set a ratings scale of:

90-99 All Pro/HOF'er
85-89 Pro Bowler
79-84 Above average starter
73-78 Average starter
68-72 Below average starter

That's without editing any player in the draft class and is in line with how the Official ratings are set. I think you may be able to lower overalls 5 points more on a classic roster without the CPU replacing them if a user adjusts the ratings downwards of those 75 overall or higher players. Any overalls for starters lower than that and each team will be drafting 4-5 starters every year without heavy editing to the draft class by the user.

You can also set overalls based on a curve for each position instead for the entire roster. Let's say you make a roster with multiple 90+ rated fullbacks, you can still have your top rated quarterback as an 87 overall. As long as each players overall rating is scaled to their position, then the overall ratings for the roster as a whole don't matter as much.

* I just tried creating an in his prime Christian Okoye at Hb and he had a 78 overall. I thought I even embellished some of his ratings. He would likely be a priority replacement for the CPU if the player isn't adjusting the incoming draft class ratings. I didn't try to move him but I'm sure he would have a 99 rating if I changed him to Fb. I haven't edited much in M17 and that was a bit of an eye opener.

[Culture Rot]

i wouldnt ruin your ratings system just to have it in line with maddens draft classes. There's numerous ways around that as you mentioned, as well as actually creating the rookies for the "next" season yourself, and assigning them after the season.

[Iteachpercussion]

Quote: Originally Posted by rplatypus

I didn't mean to say it's not possible to edit the draft class, I meant that it's not practical for the average user to edit an entire draft class after each season. Editing names, ages, appearance, college, ratings, etc. is pretty time consuming. Especially if you are trying to recreate a specific year. Adjusting contracts also takes forever. You can easily have multiple hours into editing a draft class that is only available in that particular franchise save. If a person is okay with a random class and just wants to keep the ratings in line with a specific roster, it obviously takes much less time. I checked 3 random draft classes for an idea of the incoming draft class overalls on M17. It looked to me based on this small sample that there are around thirty to forty 75-79 overalls, thirty to forty 70-74 overalls and over a hundred 66-69 players coming into the league on average. If these results are typical for most users then I would set a ratings scale of: 90-99 All Pro/HOF'er 85-89 Pro Bowler 79-84 Above average starter 73-78 Average starter 68-72 Below average starter That's without editing any player in the draft class and is in line with how the Official ratings are set. I think you may be able to lower overalls 5 points more on a classic roster without the CPU replacing them if a user adjusts the ratings downwards of those 75 overall or higher players. Any overalls for starters lower than that and each team will be drafting 4-5 starters every year without heavy editing to the draft class by the user. You can also set overalls based on a curve for each position instead for the entire roster. Let's say you make a roster with multiple 90+ rated fullbacks, you can still have your top rated quarterback as an 87 overall. As long as each players overall rating is scaled to their position, then the overall ratings for the roster as a whole don't matter as much. * I just tried creating an in his prime Christian Okoye at Hb and he had a 78 overall. I thought I even embellished some of his ratings. He would likely be a priority replacement for the CPU if the player isn't adjusting the incoming draft class ratings. I didn't try to move him but I'm sure he would have a 99 rating if I changed him to Fb. I haven't edited much in M17 and that was a bit of an eye opener.

You are right, if you try to redo an entire draft class each year to make it true to life, it is a VERY time consuming process and one that most users have no interest in. That is not necessarily my goal.

Your findings on draft class OVRs is in line with what I have noticed as well. This allows, with the ratings I normally use, for my rosters to be used without too much trouble, at least for a couple of seasons.
If I used the ratings that I feel create the best possible game play as well as sim stats, it requires editing rookies mostly with the QBs. Some RBs and WRs may need need editing as well, but to a lesser extent.

Making Okoye is a tough and you are right, if you make him true to life, he ends up low OVR and the CPU will draft a replacement. I am okay with embellishing some, but I think the players begin to lose their identity with too much.

Madden makes most WRs too good as well, and it is difficult to make them "fit" correctly. In order to get a guy into the upper 80s OVR you have to have really high numbers in all of the major categories.

For most of the rosters I have made, I do something like you mentioned, curving the OVRs by position. Not only the OVRs, but other ratings as well. I create a kind of min/max for each position and the major ratings for that position. Occasionally a superstar goes above the max. For example, I tend to make TEs slower than they really are because of how Madden works. In the modern NFL there are more TEs who contribute. Linebackers are not good enough in coverage on Madden to keep up with most average TEs. So, I slow down the TEs and drop their receiving ratings. That is, except for the big name guys who really stood out.