Wednesday, August 26, 2009

Madden 10: Realism Sliders v3

Quarter length: 12
Play clock: On
Accel clock runoff: 15 seconds
Injuries: 50 (still checking)
Fatigue: 50 (still checking)
Game speed: Fast* (game balance will not be the same at other speed settings)
Player min speed thresh: 70
Camera angle: Wide
Fight For the Fumble: Off

Offside 100
False Start 100
Holding 52 (very touchy--don't go higher than this)
Facemask 62 [1]
Def PI 100
Off PI 100
KR/PR Int 70
Clipping 75
Int Grounding 70
Rough Passer 75
Rough Kicker 75

Note: at first, it looks like there are only four sliders. Press "A" on each to access the sub-settings (thanks to Bill Abner for the tip).
QB Accuracy 20
Pass Blocking 75
WR Catching 50
Broken Tackles 51 [2]
Run Blocking 100
Fumbles 25
Pass Defense:
Reaction Time 40 [3]
Interceptions 20 [4]
Pass Rushing 75 [5]
Rush Defense:
Reaction Time 20 [6]
Block Shedding 0
Tackling 45

QB Accuracy 25 [7]
Pass Blocking 80 [8]
WR Catching 50
Broken Tackles 53 [9]
Run Blocking 100
Fumbles 28 [10]
Pass Defense:
Reaction Time 100 [11]
Interceptions 25
Pass Rushing 95 [12]
Rush Defense:
Reaction Time 60 [13]
Block Shedding 0
Tackling 45

FG Power 60 [14]
FG Accuracy 40
Punt Power 60
Punt Accuracy 80
Kickoff Power 45 [15]

What's being addressed in this version (in order):
[1] I was using Facemask penalties to compensate for the lack of defense of pass interference penalties (even with the setting at 100). It does, but it's annoying and unrealistic, so I lowered the setting.
[2] I love broken tackles, because they're dramatic, but there were way too many, so this has been lowered (and tested extensively). If you're using the truck stick and special moves extensively, you'll want to reduce this setting even further. This was balanced for players who usually let the CPU run the ball for them, or just sprint with the RB without using special moves.
[3] This improves reaction time, but the CPU offense is still plenty effective. It makes pass coverage look more realistic (although still far from ideal).
[4] The Human player was generating too many turnovers in the previous version, which enabled players to beat much stronger CPU teams. So in this version, there's a slight bias toward the CPU in both the INT and the Fumble settings.
[5] the biggest change in this version is to the pass rush. Based on stopwatch timing, neither the CPU nor the human player had enough time to throw the ball against a standard pass rush in comparison to real NFL numbers (both were averaging below three seconds against a four-man pass rush). Now, time and the pocket should be much more realistic, and the CPU was given a bonus of about three-tenths of a second to give them more time to make decisions.

Oh, and if you have a crappy defensive line and can't get a good push, don't blame me. It's supposed to be that way.
[6] The CPU was generating unreasonable rushing totals. This will put them in a more realistic range.
[7] I tested this setting with both the most accurate and least accurate quarterback in the league, along with a few others. It works fine with starters and backups, but third stringers with accuracy below 60 are going to struggle (as they do in real life). The CPU has a 5-point advantage over the Human player.
[8] Slightly improved to reduce the Human pass rush. There needs to be a difference to compensate for the CPU's less effective decision making compared to the human player.
[9] Slightly reduced to lower the number of broken tackles by the CPU. Again, a slightly better setting for the CPU than for the Human player.
[10] There were slightly too many fumbles for the CPU with the old setting.
[11] This setting is really cranked up from version one. CPU pass defense was just too weak, and this significantly improves the CPUs ability to cover receivers. Do not think that CPU pass coverage becomes psychic or something, because it doesn't.
[12] Slightly lowered to increase the time that the human quarterback has to pass. This should also allow play-action to work, although not always.
[13] This should make the Human rushing game more difficult. In v1, it was just too easy to run. I did extensive testing to improve the yards per carry average compared to real NFL statistics for several different backs as test cases, involving watching over 1,000 plays (in total) in practice mode.
[14] Slightly increased because too many lower-tier kickers had less kick power than in real life. The correspondence is very close now. If you don't think you can kick far enough when you control the kicker, don't forget to lower the trajectory for long kicks.
[15] There were an unrealistic number of touchbacks with the old setting. This reduces kickoff power for the best kickers by about 3 yards, which produces more kicks right at the goal line.

Now I will bore you with details. Testing this version took 30+ hours in total, and it was a fairly mind-numbing process of adjusting settings, watching results in practice mode, and adjusting again.

I did do a few things different this time, and I think they were more effective from a conceptual standpoint. To balance the pass rush, I broke out the stop watch again and actually timed how long it took to get to the quarterback, using different teams and watching hundreds of plays. That's how I established that the pass rush was just too effective in comparison to the real NFL.

For rushing, I used a selection of six running plays (each to both the left and the right, for a total of 12), and ran each play 15 times in practice mode. So each test consisted of 180 play results in total, and I calculated average gain for pitch, outside, and inside running plays.

I used a selection of three teams as defense: best, average, and worst in terms of yards given up per rush last season. What I wanted was for the running back to be above his real average yards per carry last season against the bad rushing defense, below against the best, and about the same against the average defense.

I also established profiles for teams, watching 60 random plays in practice mode (about the number a team runs in a real NFL game) multiple times. It helped me understand how to adjust the settings to make it more difficult for the Human player without cheating.

What makes Madden different from all the other football games I've ever played is not that ratings matter. Ratings have always mattered.

Usually, though, only a few ratings matter.

For instance, in almost every football game I've ever played, I could just find and draft the fastest halfback. Maybe he had terrible hands, so CPU teams didn't rate him highly, but he'd never fumble enough in the game for it to matter.

Actually, speed in general was always a killer. Just draft the fastest guys and you'd win, both on offense and defense.

With this version of Madden, though, if your very fast halfback has a crappy carry rating, he will be a fumble machine. If your linebacker can't tackle, he will get run through like paper. If your quarterback isn't accurate, it won't matter how strong his arm is, because he will be spraying balls all over the field.

What all this does, besides making the game much more fun to play, is make your GM decisions in Franchise mode meaningful. There are no decisions without some kind of adverse consequence. That will make Franchise mode an absolute blast.

One last note. If you want to send me feedback on this version, please don't do it on the basis of one game. It takes several games to get a feel for how these play, and each game will not play out in identical fashion.

Site Meter