FanGM

Looks like the developers at Madden are tweaking their ratings a bit - http://insideblog.easports.com/archive/ ... w-era.aspx

(if you don't want to read the entire article there, the important information is in the third paragraph)

Of course, we'll have to wait to see what the new grading system looks like, but we may need to do some tweaking of our own with the grade system here.

Lovely.. now all I gotta do is get Manning, Fitz, and Johnson. That should be pretty cheap!

Thanks for sharing that. Yeah, we eventually are going to get more away from that and hopefully into more independent rankings based upon more factors than just one, but not ready for that yet.

I was going to look at adjusting our conversion scale a little bit this year anyway as it was a very simple and hurried manual range conversion before that could stand to be more sophisticated with more time put into it.

Grades for players now were the same from the end of last year, and likely will be somewhat different when the new game version came out with updated ratings that we'd also incorporate. Sounds like they have a different approach to that and could have different range of grades this year. To compensate, we'd adjust our conversion scale a bit also to fit the original intentions:

Grade Scale:
A+ : Hall of Famer, peak years
A : All-Pro, Pro-Bowl, etc.
B+ : Above Average Starter
B : Legit Starter
C+ : Below Average Starter
C : Average Backup
D+ : Below Average Backup
D : Fringe Backup
F : Worst, liability on the field

We'd tweak out scale that says a madden score of X = 6.0 = B accordingly to where solid starters and still solid starters with B's no matter what madden score says, and make our conversion scale try to best group those types of players together. Hopefully there will be more info about a score of X next year was more like a Y last year and we can also use that to help in conversion scale adjustments.

Thanks!

So I've been thinking about how we could develop more of a system of our own for quite some time now, and I think I may have come up with an objective way of grading players. I've got a google spreadsheet published where I'm entering in different players so you can see how it compares to what their current grades are: http://spreadsheets.google.com/pub?key= ... 9dAs2BDrVw

If you have time and want to help compile the list, send me a note with your email and I'll send you a link to edit the document.

Troy has wanted for a while to make sure we stick to something like what he posted just before me on here. And the problem I've had with that is that you can't predict when a future hall of famer is having his peak years - so you end up guessing who should be an A+ versus an A. To top that off, the Pro-Bowl these days has become somewhat of a popularity contest with the fan voting involved. In my opinion, the Associated Press's All-Pro team is much more distinguished.

Therefore, I propose the following requirements of each grade (all of which are explained in the spreadsheet):
A+ - Current All-Pro who has been selected to multiple All-Pro teams
A - Current All-Pro OR current Pro-Bowler who has been selected to multiple Pro-Bowls
B+ - One year removed from All-Pro selection OR current Pro-Bowler
B - Started 9 or more games last year OR current 1st round pick
C+ - Has started at least 1 game in career OR former 1st round pick, 2nd (current or former) or current 3rd round pick
C - Never started before or 3rd (former), 4th or 5th round pick (current or former)
D+ - 6th or 7th round pick (current or former)
F - undrafted player

Within these ranges, we could use Madden grades (or some other method that hasn't been thought of yet) to differentiate between the specific numerical grades.

Furthermore, we could update periodically on our own within the season whenever a player reaches a new grade level (if a C ranked player starts a game, they become a C+). This would make weekly grade appeals possible, along with significant changes based on pro-bowl rosters (give some excitement to late season events and have playoff implications).

Some good suggestions and detail on the level descriptions there. That's somethat the general idea of the grades initially for different groupings but expanded upon nicely.

There's a lot to think about from that front about independent grading so perhaps not ready to do that quite yet (as also effects changes to the simulations if grades scales and values altered) but I do like where you are going there. There are two objectives there generally about an independent grading system:

- Avoiding problems being tied so directly to another game's grades not only for potential legalities but mostly just because it's entirely out of our control and that's problematic long-term. Plus there are lots of opinions on players out there, not just one and lots of people have different opinions than them that are just as valid and perhaps more about on-field play directly than video game skill application.

- In the broader scope of things, it is hoped to expand in to more leagues and more players and that's good but also could lower the floor on player experience. For the best experience for everyone overall, it's ideal usually if mostly informed decisions are made by all the teams. Ripping inexperienced players off may be good temporarily for the advantaged team, but bad overall for everyone and fair competition. I think it's a natural for extremely detailed simulation leagues like this that rely upon making decisions on 3rd string offensive tackles or 7th round draft picks from North Dakota State also have a content component that aids in player research.

There are sim leagues and there are analysis sites/blogs out there. To me having a merger of the two would be great for both the players getting more info to make fairer decisions, as well as the two feeding off each other and bringing more fans to both.

There will be some of that already with the draft tool updates this weekend, and I think I mentioned that before would be a heavy part of the baseball sim league whenever that gets going with a prospect site integrated into it to assist with basic minor league player opinions.

When I think about independent grading, I like the additional levels but it would be partly driven by an accompanying analysis site to where people just seaching for opinions on the top NFL lineman would find that content but also see a strong link to a sim league where they can play something beyond fantasty football that matches their more intense interest.

As opinions (or game start data or awards or madden updates or whatever) are collected on the analysis sites, those variables would shift the baseline grades either up or down both for the rankings within those analysis sites as well as the grade within the tied sim leagues all automatically across the entire network.

So that's kind of what I have been thinking and much of it at least partially in development still in hopes of enhancing the experience and giving people more opportunities to be involved as well as help influence player values through gathering opinions on their players.

First, some preliminary findings now that I've got an unscientific sample of about 170 players - taken from All-pros from 2008, 2007, and some from 2006, some 2008 pro-bowlers, and my AFFL Arizona roster:

• There are more A+ grades under this system (unless there are some players that are rated A+ and aren't included in my sample yet).There are less A grades under this system. In fact, there were 65 A or A+ grades last year and 64 so far this year. So it seems this just spreads it out a little better between A and A+. I'm not sure if this is a good idea or not. Perhaps that could be fixed by changing to A+ being a 3x or more All-Pro who is currently an All-Pro, but that would make the only A+ players for 2009 Peyton Manning, Tony Gonzalez, Walter Jones, Steve Hutchinson, Alan Faneca, Kevin Mawae, Julius Peppers, Kevin Williams, Kris Jenkins, DeMarcus Ware, Joey Porter, Ray Lewis, Ed Reed, and Troy Polamalu. I guess that could be okay.
• It is very difficult for a player to maintain of A or more. Even more so, B+ becomes a bit more exclusive also and there is a possibility of jumping back and forth between B and A. That could be a problem as we have a lot of people who have been to multiple pro bowls but would drop from A to B if they don't make the game from one year to the next. Perhaps throw in a tag where a player can be B+ if they were in the Pro Bowl within the last 2 years?

Now on to respond to Troy...

I agree with you on the first objective and want to add that I'm sure there were a number of players who were improved in the madden ratings last year but weren't included in the posted update. As a result, we have players that could probably be rated a bit higher (or lower) but aren't.

Second objective I somewhat agree with you. I think we need to have a system that changes the grade before the off-season begins to make sure that the inexperienced players aren't taken advantage of with players that have significant changes involved. I hope that this system kind of takes care of that by allowing us to update and change the grade immediately.

However, I don't really like the idea of incorporating a lot of different opinions. A lot of the opinion and analysis blogs out there are specific to teams (run by fans or sports writers of the city) and hence can be a bit biased. We need something that is as objective as possible.

I'd like to add a third objective to try and pursue with any grading system and that is a bit more parity within the league. Right now I see the average rating as 6.0 (B) and we need to make it difficult for a team to go much higher than that. I think that means lowering the maximum grade a single player can contribute.

I can agree with some of that but advantages/disadvantages to everything. It's stuff I want to capture and consider amongst everything but too much could be made of a flat number of "starts" or draft round outside of context also.

There might be a stud young player who gets little playing time with a super bowl champ behind an aging all-pro but a worse youngster who plays like crap 16 times a year being forced into action on a horrible team with no alternatives but racking up starts. Numbers can often flat out lie. Just as there can be tremendous bad reaches and busts in high rounds and great steals and players who slide into lower rounds that produce much better prospects.

I know some baseball stat guys and they both intrigue me with knowledge and drive me up the wall in frustration with some bonehead conclusions that nobody watching the game would come to. Going just by the numbers sometimes can be misleading in terms of valuing a player if there isn't some common sense context given to the situations.

I seem to agree to most of what that system would entail, however, here's a flaw that I can see:

Running-back tandems

Before last year, according to that system, Michael Turner and Marion Barber would have been C's or C+'s, as they had not started many games. However, I think that we can all agree that neither of them would have deserved to have been rated that low, as they showed last year. And now with the Dallas RB situation, Jerious Norwood in Atl, and the two-headed attack in Ten, I think those players (and others in the league) deserve consideration beyond having started a game.

Perhaps if we were to use this system, performance-based statistics could also be worked into the grading system (i.e. # of touchdowns, yards, etc. - or sacks, tackles, etc for defensive rotation players) as a way to compensate for starting-caliber players who, for some reason, haven't been given a start.

And this doesn't just apply to RB's - look at Anthony Gonzalez in IND or Justin Tuck in NYG prior to last season.

We can also probably use similar numbers to adjust the grades of players who are starting, but shouldn't be, sort of like our own football-version of VORP.

I think the real worry is when it comes to offensive linemen. They don't accumulate stats, and would be very tough to evaluate without an outside source.

OL's do accumulate stats - you just need to know where to look.

http://hosted.stats.com/fb/findplayer.a ... osition=95

That site has OL stats - not sortable, which makes things a real pain in the arse, but it has them (such as false starts, penalties against, sacks against, etc.). The only other thing, mostly for inside linemen that it doesn't have is how well an OL does against the run, but that can usually be figured out by a team's YPC.