FanGM

Madden grades increasingly get more difficult to use with more complaints. Thank goodness for brhlbryce and some of his tech friends finding a way to pull data efficiently, or manual grade updates would have been too much for me this season with some of their public list site formatting changes and the difficulty quickly getting the information.

We've had grade discussions before. Most probably think Pro Football Focus does a better job in player performance grading, or have other ideas about us grading players ourselves. PFF just wasn't an option because of it being a paid service, so making that premium info freely available here being wrong and problematic.

Maybe there's a way to make updating grades easier for site administrators and arguably better for most. Here's a suggestion to get everyone's feedback:

- The league updates grades every 4 weeks (instead of weekly) during the season. Arguably a player's value shouldn't change too much from one week to the next, but evaluated over more time for better judgement. Also it's less administrative work leading to more efficient weekly sims if not updating grades every single week.

- We change our grades from relying upon one source (Madden's free access publicly posted grade list) to a more proprietary grade system.

- Instead of one source, we come up with our own formula that mixes in a couple different grading sources together to come up with a wider consensus value across multiple top grading source opinions. Maybe it starts with a 50/50 kind of mix between popular grade sources.

- We aren't reliant upon just one source and whatever they change from year to year. We aren't publishing anyone else's grades at all (whether public access or premium info), as we have our own calculation being published that's not exactly any other grade based upon a variety of different grades out there.

- We keep things independent and objective just like now with no control over the grades ourselves, but just use a mix of various resources instead of one directly.

Some administrative/legal advantages long-term after some short-term figuring out how to calculate a mixed grade. Gives us some flexibility to further customize our own unique grades ahead as we move past one source. Perhaps gives worthy players their rightful values incorporating more grading opinions and not just one that some often complain about. But maybe also makes it a little more difficult for average players here to understand the grades and how they're changing if less of it is public (compared to just look at Madden's grades posted online now) and more of it is based upon information most others can't see and recalculated in more complex ways.

One note on PFF, the current player grades are now publicly available. That's a change from the past. I don't think it solves the administrative issue though.

I'm almost always in the boat of more data being better, but as Knight said the PFF overall grade is now publicly available. I know that's all we see on the page, but do you pull more than that from Madden, Commish?

JonC wrote:do you pull more than that from Madden, Commish?

No it's just the overall grade now.

I didn't know that about PFF. Good to know. Looking around there now, I see that in some places, but doesn't seem easy to export that information on first glance but I'm not too familiar yet.

It's possible, perhaps, that we just switch from one public grade to another and still rely just on one resource, but would have to look into that more. Feels a little riskier to rely upon a new resource that just started being public (less certainty that would for sure continue ahead) and still has premium access interests.

If most feel PFF is much better than Madden grades, but we end up wanting our own custom calculation published, it's also possible that we gear that formula much more to one than the other (75/25 or whatever) but still get in a mix of grade opinions for our player value.

What about a monthly change unless a player has an increase higher than a certain amount any given week.

For me personally, Wyatt Teller improved very quickly. Not sure what the difference in my grades and game outcomes wouldve been had I had to wait a month to get them into the mix.

And at the same time, DJ Humphries is a 88.3 on PFF but only 77 in Madden. Joel Bitonio 85.5 PFF but 88 in Madden.

I like the idea of more data from more sources but finding the right balance would be crucial

Goodell wrote:
Looking around there now, I see that in some places, but doesn't seem easy to export that information on first glance but I'm not too familiar yet.

I think it may be easier to export data if you have a premium subscription. That would give you access to the data tables.

I'm a fan of a Madden/PFF composite score if that is feasible.

PFF is great!!! I have a subscription and it is awesome. I think it could be a slippery slope here though. PFF is just the player's grade on every play, not taking potential or responsibility on the play into consideration. Secondary is particularly tricky. Some corners are given safety help and some are put on an island. Both Tredavious White and Stephon Gilmore had PFF grades below 70 for 2020, mostly because they are asked to cover the other teams #1 receiver and are rarely given safety help to their side of the field.

Offensive Line is similar in that some Tackles are given TE help, or the RB will chip on their side.

One idea:
Use the Madden grades as a baseline, a player's grade can never go lower than that. Then every four weeks, check Madden grade vs PFF grade (with snap minimums) and if Madden grade is higher, add half the difference to the FanGM grade.

Example:
Madden grade 70
PFF grade (with at least an average of 15 snaps per game played) 82
FanGM grade 76

Example:
Madden grade 70
PFF grade (with al least an average of 15 snaps per game played) 62
FanGM grade 70

We all play by the same rules so I don't want to overload what you already do. I have PFF Elite and I like it. Like someone said (Jon C maybe) more data is better. Whatever is best for you.

Larry

I like the idea of using multiple sources but PFF's grade would obviously have to be adjust to a scale of 99. Like the top ranked guard is a 99 while 2nd is 98... 10-15 is 85... 40-60 is 70... I didn't put a huge amount of thought into that but I think ya'll understand the jist. Then do like 50/50 with Madden grade.

robroach wrote:PFF is great!!! I have a subscription and it is awesome. I think it could be a slippery slope here though. PFF is just the player's grade on every play, not taking potential or responsibility on the play into consideration. Secondary is particularly tricky. Some corners are given safety help and some are put on an island. Both Tredavious White and Stephon Gilmore had PFF grades below 70 for 2020, mostly because they are asked to cover the other teams #1 receiver and are rarely given safety help to their side of the field.

Offensive Line is similar in that some Tackles are given TE help, or the RB will chip on their side.

One idea:
Use the Madden grades as a baseline, a player's grade can never go lower than that. Then every four weeks, check Madden grade vs PFF grade (with snap minimums) and if Madden grade is higher, add half the difference to the FanGM grade.

Example:
Madden grade 70
PFF grade (with at least an average of 15 snaps per game played) 82
FanGM grade 76

Example:
Madden grade 70
PFF grade (with al least an average of 15 snaps per game played) 62
FanGM grade 70

PFF grades are brand new to me, but does sound like some complexity to them to figure out if we end up wanting to incorporate.

May need more suggestions like this above from those more familiar with PFF for us to decide if there's an approach or formula to exactly implement.

What I like about this suggestion is that it keeps things grounded in grades will be at least the familiar Madden grades we've been using, but that if Madden is drastically undervaluing someone that other graders like PFF think is performing much better, that we'd boost up that base Madden grade some.