Madden NFL 11

I know, I know. This doesn’t come out for awhile, but the development team is starting to put up blogs detailing what they are working on. Right now they are focusing on pretty low-level gameplay stuff, and there are some pretty interesting and welcome changes.

Madden NFL 10 was a pretty dramatic departure from Madden NFL 09’s emphasis on twitch control. We really focused on trying to make the game look and feel more like what you see on Sundays. In the end we did a decent job, but we found out pretty quickly that the tech we were using wasn’t giving us the control that we needed to match our vision. Things like acceleration, direction changes, and momentum were very difficult to tune and while the end result was a fun user experience, we knew we had to make a pretty dramatic change in our tool set to allow us to achieve our goals.

At the start of the Madden NFL 11 cycle our goals were to integrate acceleration, momentum, and direction changes into the game in a way that felt responsive; was easy to tune; and looked more like what you see in an actual NFL game. With those targets in mind, we have completely ripped out our old locomotion engine and replaced it with a brand new system. We collaborated closely with the FIFA team and are combining the best parts of their locomotion system (called RailTracks) with a brand new tech we created specifically for Madden NFL 11. This new system has allowed us to meet our goals, and I’ll discuss how the tech affects the way the game looks and feels - focusing on momentum, acceleration, and the look on the field.

Momentum: In Madden NFL 10, players couldn’t turn as fast when they were sprinting as they could when they were not. One of the mechanics around this was forcing users to let go of the sprint button when they needed to make sharp turns (like trying to choose a run hole for example). This wasn’t very intuitive for a lot of our users. We knew that we had to make a change that would allow people to move where they wanted to on the field, while still making skills like Agility matter. We tried a few different things, but the mechanic that felt the best was to allow the player to go where you are pressing on the left stick, but slow him down to make the turn you are requesting (based on his ratings of course). It feels pretty natural to gradually round off a turn if you want to keep your speed up, or move the stick sharply to sacrifice speed for making a necessary cut. As soon as we put this in, the game felt way more responsive. You can feel the momentum of trying to turn a big lumbering lineman, but at no point do you feel like he isn’t going where you are trying to make him go.
If you try to make an extreme turn, your player will play a direction change that will dynamically update to match changes in desired movement direction. For instance, if you start a 180 degree direction change, but then change your mind in the middle, you can branch out. If you decide halfway through that you need to adjust where you want to run just a little bit, the animation will compensate and make sure you exit in the direction you are pressing on the stick. Direction changes have always been a tough balance between look and responsiveness, but our newtech has made it easier for us to improve both the look and feel of direction changes.

Acceleration: In years past, a lot of Madden NFL fans cared more about speed than any other rating. Our game was tuned around the fact that our old tech couldn’t really handle acceleration gracefully, so acceleration didn’t have much of an effect at all. This didn’t become a problem until the end of Madden NFL 10 when we really started to re-introduce momentum and acceleration into the game. From the start of Madden NFL 11, acceleration has been a huge focus. We set up tuning controls and a locomotion debugging mode I like to call the Track Meet to enable us to do a lot of fast iteration on how speed and acceleration interact. This debugging mode lets me set up 40 yard and 100 meter races between players so I can tune how acceleration and speed affect 40 and 100 times. Our speed curves are modeled after real life 100 meter race data from Olympic athletes. If you plot the speed of an Olympic sprinter at 10 meter splits you get a curve that represents his acceleration. In our animation tool, we can recreate those graphs of speed over time to make the acceleration of our football players match the shape of the curves from the Olympic sprinters. I’ve included the curves we use to represent a max acceleration, max speed player and a min acceleration, max speed player. Time is in units of 1/60 of a second and speed is in yards per 1/60th of a second.

Obviously, players in pads can’t run the same speed as a world class sprinter coming out of blocks but their acceleration curves have the same shape. They get up to around 80% of their top speed relatively quickly, but the last 20% is stretched out further (it takes the Olympic guys around 60 meters to hit 100% of their top speed). Once I had the curves created, I tuned them around combine 40 times and 100 meter times of players who ran track in college. I then passed that info on to Ratings Czar Donny Moore so he could set speed and acceleration accurately for our rosters. In total, we have 24 different curves that we can tune to control everything from how a player’s speed increases over time (based on acceleration and speed ratings), to how much he can turn based on his current speed (driven by the agility rating). That level of control really lets us differentiate players on the field.

It is now possible for our ratings guys to make players that are quick over 10 to 20 yards, but get smoked at longer distances. A 90 speed player can now run the same 40 time as a 99 speed player, but lose over longer distances. I have a video here to highlight this. I created two players. The first is 99 speed, 90 acceleration (#99, on the top in the video). The second is 90 speed ,99 acceleration (#90 on the bottom in the video). The players race for 90 yards in the video. As you can see, the player with higher acceleration jumps out to an early lead, but gets passed at around 40 yards by the higher speed player. The animations aren’t final (particularly the initial move that plays), and I still have some fine tuning to do on how we map the acceleration and speed ratings into our set of curves, but you can see how powerful acceleration is going to be in Madden NFL 11.

Visuals: A big issue with the look of our previous locomotion was that it looked robotic over long distances. Our run cycles were only a few steps long (another limitation of our old tech), and we didn’t have a good way to introduce subtle variation dynamically. We now have the ability to fix both of those problems. At mocap, we set up a 50 yard long runway and recorded our mocap talent over the whole distance. Our forward run cycles for each body type now don’t loop until the player has traveled for 40 yards. The differences in each stride are subtle, but that small difference is enough to break up the robotic feel of our old locomotion. We layer in dynamic lean based on how much a player is accelerating or turning to get rid of the upright turning from Madden NFL 10 and contribute to the feeling of momentum. Our layering technology also allows us to do things like add ballcarrier arms, cover ball arms, trucking, or stiffarming to any of our locomotions, enabling you to have full control of the player (including direction changes) during these states. In the past, control felt inconsistent across these types of locomotions because each was an entirely new set of animations. Sharing the base run cycles across multiple movement types allows for a more consistent experience and gives the user more control than in the past.
http://maddennfl.easports.com/blog.action?blogId=locomotion

I can’t wait for new feature info and what they are doing with online franchise.

From the previous blog post:

Finally, we have gotten sideline and back of the endzone catches to trigger in the appropriate situations. In previous years we have seen some situations where a receiver tries to make a catch near the sideline without making any attempt to get both feet inbounds. This was most commonly seen on flat and out breaking type routes.

For Madden NFL 11, with the help of the Catch Debugging Tool, we have corrected all of our sideline / back of the end zone catches. You no longer have to be afraid of trying to squeeze that Deep Out or Goalline fade inbounds, as now the receivers have much more advanced AI to get both feet in.

And ditto what Cubit said re: new features and online franchises.

I hope there’s a way to migrate our existing QT3 franchises into Madden 11; on the other hand, as a Lions guy, I’d much rather have the real Lions 2010 roster given all the improvements they’ve been making during the offseason :) Realistically, given how they’re changing the underlying models / stat engines / etc, I don’t see how it’d really be possible anyway.

“We heard reports from players that in some cases during Franchise Mode, the AI-controlled Detroit Lions were winning games. In Madden 11, we will fix this bug.”

Ha :)

I forgot about short fade routes. Those are pretty much broken in Madden 10. I’ve never successfully completed one. Good to hear that is being fixed.

I LOVE THE FADE ROUTE. I’ve wanted to use it in Madden forever.

I’ll be watching this closely, and I imagine picking it up for online franchise with you guys.

That momentum thing is something that I think Visual Concepts had right in NFL2K ten years ago. Sheesh.

Visual Concepts isn’t exactly making football games anymore, so if Madden finally implements it well, you won’t hear me complain.

Thursday is gonna be a big day for Madden 11. They are releasing the first trailer, along with demo and cover announcements.

Drew Brees is gonna be the cover athlete:

He’ll challenge the cover curse, then – Brees has been scary consistent for the last six-seven years.

Larry Fitzgerald already debunked the curse for Madden 10, but the same can’t be said for Troy Polamalu. Does that mean the curse is considered a wash for Madden 10? :-)

Nah, just that it focused on one guy. The curse isn’t broken until someone proves they have a successful season/career after it.

It is official:

Why is Brees wearing a towel like a Hooters girl trying to hide her camel toe?

Well, now we know that New Orleans will have a down year in 2011. :)

Debut trailer is out:

Wow, that was terrible.

What I want to know is what this “350 plays down into 1!” nonsense is.

When the Madden 11 team was looking at stats from Madden 10, they found out that a pretty significant amount of a player’s time (like 40%) is spent navigating the play call menus rather than actually playing the game. They are not cutting down on the amount of plays, but they want to streamline things so that you spend less time in the play menus which then shortens the games. I don’t know how they plan on doing it though.