RFF's M21 32-man CFM USER v USER Sliders

[Aestis]

You should use these sliders IF:

• You are in a CFM with other humans, preferably 32 but a few leagues of 4-10 do use these as well.
• You want gameplay to be as realistic as Madden can get--this means coverage is tougher, pass rush comes faster, and your INTs & sacks may be high if you are more aggressive than an NFL QB would be.
• You want sliders driven by actual data from hundreds of games (strictly User v User games from After Work Football League), and based on per-play efficiency metrics built by an analytics director IRL. This set is not based on how I personally feel after playing a handful of games.
• You are comfortable with the fact that these may not lead to a great User v CPU experience--these are entirely optimized for our 32-man league, not CPU games.

The Methodology:

Spoiler

Dating back to M15, I've been measuring our league's User v User game stats & benchmarking against the NFL using per-play advanced metrics (e.g. Pts per Play, Net Pass Yds per Drop-back, Air & YAC Yds per Pass Comp, Sack & INT rates, 20+ yd run %, etc.) as well as NFL weekly medians (to understand volatility) to balance our sliders.

This means hard data from 32 humans across hundreds of games are what dictate my slider adjustments, NOT the subjective judgment by the slider creator on how they think they should be performing.

So if I'm personally having a hard time running the ball, but the league as a whole is tracking great vs NFL run success...the run game goes untouched. Without this visibility into league-wide stats in USER games only, there is no honest & objective way to balance a 32-man league.

Again, this is all benchmarked against NFL median per-play success, not just game-end point & yardage totals. Sim leagues who run 6-minute quarters where each team snaps 35-45 plays/game might end up with realistic looking 34-24 type scores, 280 yds passing by QBs, 90 yds rushing from RBs...but that's off 21 pass attempts and 13 rushes. Despite calling themselves "sim leagues," their actual gameplay is incredibly arcadey. Scoring is MUCH easier in these leagues. Scoring in the NFL is tough, teams have to punt often, and our sliders are designed to reflect this.

If you run this set, to succeed you will have to:

Spoiler

- Stop throwing it into the middle of sagging coverage. Hardest thing for most Madden players to learn. NFL QBs don't force it, and they don't try to thread tight needles 15 yards downfield at a huge risk of INT. Madden players tend to force it downfield into coverage no matter what then complain about INTs. NFL QBs check down & live to fight another day--even if it means punting.
- Get the ball out quick. Pass rush can come quickly. You better actually read Ds & have a progression. The "drop back 10+ yds and wait for someone to come open" age-old video game approach won't work.
- ID & adjust to possible blitzes. All the pass block sliders in the world won't make your 5 blockers protect against 6 rushers. In the NFL, 5 man pressure vs 5 blockers reaches the QB in 2.5 seconds or less about half the time. 4 man rushes vs 5 blockers reach the QB in 3 seconds or less about half the time. Better have your "hot" read ready when the D dials up a blitz.
- Get in 3rd & manageable. 3rd & long is just really really tough on these sliders, as it is in the NFL. Don't anticipate converting a whole lot of 3rd & 8+ on these sliders.
- Stop assuming a route will always get open because it always got open in previous Maddens. E.g. corner routes & outs are no longer automatic wins on these sliders. (But they can still be highly effective vs the right coverages with the right matchups!)

tldr; Good defense + limiting turnovers + field position = winning, with these sliders.

ver 1.17 - 2020 Season Wk 17
Updated: Oct 29
latest changes highlighted in yellow



*Important notes on XP sliders:

Quarterback
Due to QB regression being too harsh for age 30+ players, our league is 'refunding' some attributes for QBs in their early-to-mid 30s, and the 70 slider reflects this. If you are not doing this, my advice is to run QB slider at ~80. 100 is still too high.

Offensive line
Candidly, no matter what you read in other XP slider threads, there is no good XP slider for OL. What I mean is, at default or near it, you will see the mid-tier starters inflated massively in several years, which will really throw off talent and even gameplay balance, as it will be FAR too easy to get any scrub to high 70s or even 80+ OVR. But if you put it lower, you won't see ANY OL hit 90+. The reason is dev trait progression. The game doesn't generate nearly enough Superstar/XF players to match what you start with. To solve this problem, our league will be manually promoting 1 or 2 OL per year from Star to SS, and 1 or 2 OL per year from Normal to Star. In this way, you can keep XP sliders lower for better overall talent distribution, while still getting some guys to hit 85+ and 90+ OVR.

[umbe88]

Re: RFF's M21 32-man CFM USER v USER Sliders

Thank you for the best simulation slide every year!

[AlessandroTaballione]

Re: RFF's M21 32-man CFM USER v USER Sliders

waiting for..when they will come out?

[Aestis]

Re: RFF's M21 32-man CFM USER v USER Sliders

waiting for..when they will come out?

I will post the sliders we're going to start the file with in the next few days, still doing some eyeball-based play.

But again the selling point here is the data from dozens of user games, and we probably won't kick off our file until roster update ~Sep 8, meaning probably mid-Sep before I get a couple of game-weeks of data to start really making data-driven adjustments. Until then it's based on how I feel.

But the good news is I kind of know by now what the gameplay should generally look & feel like from years of seeing how that translates to #s, so the initial set should at least be in the ballpark. I'm usually wrong about a couple of things, though.

[bucknut7]

Re: RFF's M21 32-man CFM USER v USER Sliders

I will post the sliders we're going to start the file with in the next few days, still doing some eyeball-based play.

But again the selling point here is the data from dozens of user games, and we probably won't kick off our file until roster update ~Sep 8, meaning probably mid-Sep before I get a couple of game-weeks of data to start really making data-driven adjustments. Until then it's based on how I feel.

But the good news is I kind of know by now what the gameplay should generally look & feel like from years of seeing how that translates to #s, so the initial set should at least be in the ballpark. I'm usually wrong about a couple of things, though.

Do you anticipate you'll be bumping up coverage to 100, and threshold to 80, as in last year's file?

[AlessandroTaballione]

Re: RFF's M21 32-man CFM USER v USER Sliders

I will post the sliders we're going to start the file with in the next few days, still doing some eyeball-based play.

But again the selling point here is the data from dozens of user games, and we probably won't kick off our file until roster update ~Sep 8, meaning probably mid-Sep before I get a couple of game-weeks of data to start really making data-driven adjustments. Until then it's based on how I feel.

But the good news is I kind of know by now what the gameplay should generally look & feel like from years of seeing how that translates to #s, so the initial set should at least be in the ballpark. I'm usually wrong about a couple of things, though.

thank you, really curious to test it. We run a 32 Human league from 2013 and always open to improve the slider set for our Franchise

[Aestis]

Re: RFF's M21 32-man CFM USER v USER Sliders

Do you anticipate you'll be bumping up coverage to 100, and threshold to 80, as in last year's file?

Already done threshold testing this year and we'll be running at 70. Last year in order to get the diff between fast and slow correct, the game itself played a bit too fast. This year the game speed is a bit slower, so that's great. 70 is pretty close to the ideal sweet spot.

Coverage/PDRT will def be starting high, 75+, and I'll see how it goes.

Unless it's tuned in a future update, PBK will be at 0. The slider makes a difference but not a huge one, so 0 PBK is still too slow of a pass rush unfortunately.

QB accuracy will be lower. Previous years if you dropped too low, you'd get too many WTFFFFF throws 15 yds in the wrong direction, often straight to an uninvolved defender in what was clearly preprogrammed INT. It's much more organic now, lower slider means more off target throws but they're all in the vicinity of the target. I don't know what the sweet spot will be until I see some data, but this will feel much better even as we drop accuracy.

Everything else close enough to stock to start that I can't tell much without seeing some data. I suspect 50 RBK will be a tad high, may start around 47 and see how the #s look.

[Aestis]

Re: RFF's M21 32-man CFM USER v USER Sliders

Our initial set when we launch will look pretty close to this... keeping it out of OP for the time being until I get some data to validate:

[bucknut7]

Re: RFF's M21 32-man CFM USER v USER Sliders

Already done threshold testing this year and we'll be running at 70. Last year in order to get the diff between fast and slow correct, the game itself played a bit too fast. This year the game speed is a bit slower, so that's great. 70 is pretty close to the ideal sweet spot.

Coverage/PDRT will def be starting high, 75+, and I'll see how it goes.

Unless it's tuned in a future update, PBK will be at 0. The slider makes a difference but not a huge one, so 0 PBK is still too slow of a pass rush unfortunately.

QB accuracy will be lower. Previous years if you dropped too low, you'd get too many WTFFFFF throws 15 yds in the wrong direction, often straight to an uninvolved defender in what was clearly preprogrammed INT. It's much more organic now, lower slider means more off target throws but they're all in the vicinity of the target. I don't know what the sweet spot will be until I see some data, but this will feel much better even as we drop accuracy.

Everything else close enough to stock to start that I can't tell much without seeing some data. I suspect 50 RBK will be a tad high, may start around 47 and see how the #s look.

Thx. Curious to know the thoughts behind the cpu sliders given your leagues focus on only user games

[AlessandroTaballione]

Re: RFF's M21 32-man CFM USER v USER Sliders

We are testing it. One thing i can say, from the feedback of our members is that running is really really hard against Human.
But the set is really appreciated between the 32 members. We Will continue to test and use it as the base for the franchise.
Thanks for the hard work, the IML ( Italian Madden League) running from 2008 and seems have found their base sliders for years to come.

[Aestis]

Re: RFF's M21 32-man CFM USER v USER Sliders

We are testing it. One thing i can say, from the feedback of our members is that running is really really hard against Human.
But the set is really appreciated between the 32 members. We Will continue to test and use it as the base for the franchise.
Thanks for the hard work, the IML ( Italian Madden League) running from 2008 and seems have found their base sliders for years to come.

Ty sir! If running is low, it'll definitely come out in the data pretty quickly. I love the feel of the run game this year, so I feel the inside/outside balance and the balance between short 2-3 yd positive runs vs long runs is solid too. There have been years where too many huge gains made it tough to balance YPC to NFL standards, for instance, without making it so too many runs went for negative or no gain.

I will say, there is going to be an adoption/enjoyment curve either way. Madden has for years ingrained bad habits in all of us.

For years it was, to buy time in the pocket, keep dropping way way back. We all intellectually knew this was bad, in the NFL you have to step UP to buy time, back = death, but it's what the game drove home as muscle-memory. And now when guys do it they often get sacked.

The game taught us we can definitely throw downfield WAY more often than a real NFL QB. Now that coverage mechanics have caught up, we throw a lot more INTs. We are used to having more time in the pocket than NFL QBs, so we get frustrated when we sit in the pocket and get sacked "before the route even finished, WTF!" But go watch an NFL game and see how often QBs are throwing within literally 1-1.5 seconds.

How realistically people think they play compared to the NFL & how they actually play are two different things. And that's a very sensitive subject, because "unrealistic" is associated with "cheesy" or "dumb" in a lot of our minds.

So, just be prepared. I bring this up because in a given week, the median team in the NFL averages roughly 3.7-4.1 ypc. My balance target is usually just a slight hair above that, but the goal is to keep it very close. That means games with bad run games will be below that mark. Nobody identifies as being a 3.3 YPC team. You WILL receive feedback that running is too tough if your league is used to higher YPC rushing. It may actually be too tough, in which case the median YPC is lower than target range, and I bump RBK. Or it may be just right and a few guys won't love where they fall in that 1-32 range of rushing success.

[Aestis]

Re: RFF's M21 32-man CFM USER v USER Sliders

Thx. Curious to know the thoughts behind the cpu sliders given your leagues focus on only user games

Short answer, if I'm being truly candid, it's just a SWAG at keeping CPU games halfway fun. We don't do any CPU games in our league, ever. CPU gets AM bonuses that User v User games do not, so I keep their sliders a bit lower than HUM levels. It helps with practice mode a bit and for leagues that do have CPU games.

So, I don't think the CPU games will be TOO much different than User games, but by no means is it some exact science. Feel free to modify at will. Presumably none of them should affect User games, but I can't guarantee they never do. It's possible CPU sliders act as a baseline and HUM sliders modify from that baseline. Not saying I believe that, I *doubt* they affect User games at all, just saying it could be true, in which case changing CPU sliders to be different than ours may affect things.

[Mike Lowe]

Re: RFF's M21 32-man CFM USER v USER Sliders

I love and respect the detail and effort put into these each year. I'm actually thinking of trying this as a base for solo, but I'd guess two sliders I would need to adjust right away would be the user PDRT and Pass Coverage as those are geared for user v user.

Will the AI be pathetic with the settings that high? In thinking maybe 60-75?

[bucknut7]

Re: RFF's M21 32-man CFM USER v USER Sliders

Short answer, if I'm being truly candid, it's just a SWAG at keeping CPU games halfway fun. We don't do any CPU games in our league, ever. CPU gets AM bonuses that User v User games do not, so I keep their sliders a bit lower than HUM levels. It helps with practice mode a bit and for leagues that do have CPU games.

So, I don't think the CPU games will be TOO much different than User games, but by no means is it some exact science. Feel free to modify at will. Presumably none of them should affect User games, but I can't guarantee they never do. It's possible CPU sliders act as a baseline and HUM sliders modify from that baseline. Not saying I believe that, I *doubt* they affect User games at all, just saying it could be true, in which case changing CPU sliders to be different than ours may affect things.

Thx. I run a 10 team league so a good amount of cpu games. About half the schedule with the way we have divisions.

Incorporated some of your set and it’s playing really, really well

[AlessandroTaballione]

Re: RFF's M21 32-man CFM USER v USER Sliders

Ty sir! If running is low, it'll definitely come out in the data pretty quickly. I love the feel of the run game this year, so I feel the inside/outside balance and the balance between short 2-3 yd positive runs vs long runs is solid too. There have been years where too many huge gains made it tough to balance YPC to NFL standards, for instance, without making it so too many runs went for negative or no gain.

I will say, there is going to be an adoption/enjoyment curve either way. Madden has for years ingrained bad habits in all of us.

For years it was, to buy time in the pocket, keep dropping way way back. We all intellectually knew this was bad, in the NFL you have to step UP to buy time, back = death, but it's what the game drove home as muscle-memory. And now when guys do it they often get sacked.

The game taught us we can definitely throw downfield WAY more often than a real NFL QB. Now that coverage mechanics have caught up, we throw a lot more INTs. We are used to having more time in the pocket than NFL QBs, so we get frustrated when we sit in the pocket and get sacked "before the route even finished, WTF!" But go watch an NFL game and see how often QBs are throwing within literally 1-1.5 seconds.

How realistically people think they play compared to the NFL & how they actually play are two different things. And that's a very sensitive subject, because "unrealistic" is associated with "cheesy" or "dumb" in a lot of our minds.

So, just be prepared. I bring this up because in a given week, the median team in the NFL averages roughly 3.7-4.1 ypc. My balance target is usually just a slight hair above that, but the goal is to keep it very close. That means games with bad run games will be below that mark. Nobody identifies as being a 3.3 YPC team. You WILL receive feedback that running is too tough if your league is used to higher YPC rushing. It may actually be too tough, in which case the median YPC is lower than target range, and I bump RBK. Or it may be just right and a few guys won't love where they fall in that 1-32 range of rushing success.

You can't imagine how happy I am to hear these words.
When we talk about being Sim in our league, we exactly talk about this.
Time in the pocket, YPC, Sticky coverage.
And oblviously we are all aware that withe learning curve starting now, what now seems impossibile tomorrow maybe will be easy.

We always search the perfect balance between a true football experience, as high as Madden can offer, and the arcade soul of the game.
I agree at the same time that the cpu need to be slight easier. We are a full league, 32 humans, cpu games are very rare and we can't risk to alter the outcome of the season for a game played against the CPU.

Thanks a lot Aestis, waiting for the first release of sliders. Will be really fun to contribute to test them and make them even better