Statistics in sliders-Madden18

[Oraeon1224]

I use All-Pro because prior testing has shown higher settings give statistical boost to CPU players and penalize Human players (Set up a 50 yard dash with a running back and a stopwatch and see the different results at different settings if you want to confirm). This was confirmed by a developer in around 2013 (they don't make the AI smarter like chess they just give stat boost across the board). The most important factor for getting realistic results is true plays per game. Your goal should be about 130 plays per game (both sides total), which is done by adjusting the play clock and game time.

All-Pro (Sim)
Speed: Normal
Clock: 15min
Minimum clock: 19sec
Injuries: 43 (seems popular this year)
Fatigue: 62
Threshold: 50 (The speed gaps are realistic so don't change this generally)
--Threshold has been tested by me in the past at low settings WR and RB get too much separation from LB's and there are too many big plays. It also gives conflicting results based on the players involved. It may pork your fast CB covering less fast WR if set high, but if set low may pork your skilled CB covering a slightly faster WR. In summary don't muck around with this slider).

Player/CPU:
QB ACC: 25/28(goal Comp% 60-70%)
Pass BLK: 50/100 (CPU gets sacked too much)
WR catching: 43/45 (goal 3-5 drops per game total)
Run blocking: 46/52
Fumbles: 33/27(goal 1 per side per game)
Pass Coverage: 0/50 (lower for player reduces AI sacks)
Pass Reaction: 85/50 (goal of 4-5 Defl per game which is real and reduces AI QB COMP%)
Interception: 43/43 (goal 1-3 per team per game)
Tackling: 42/44 (goal 5-6 BTK per team per game)

Special teams:
Kickoff 50 (seems pretty accurate so far)
FG Power: 50
FG Acc: 43 (50yd 50-60%)

Penalties (This gives realistic penalties with 70 yd per game not fun penalties)

Offside - 81
False Start - 54
Offensive Holding - 90(barely works at 80)
Defensive Holding - 0
Facemask - 61
Defensive Pass Interference - 65
Illegal Block In The Back - 53
Roughing The Passer - 51 (may put back at 1 to mimic PI holding calls)

[Oraeon1224]

Principle thread:
The main purpose of my sliders is to create a starting point that I find helpful. From NCAA to Madden I have found that recreating realistic statistics in these games (using NFL data) results in more realistic looking games. I generally start off testing how the sliders work then start with a preliminary set and then adjust the sliders to get the statistics to match real world NFL statistics.

When you are using any sliders your individual skill level will always impact results. That is why it is important to get a starting set and then adjust to make them work for you, by understanding what is not realistic and then moving the correct slider to match your game play strengths and weaknesses.

TIPS:
Play realistic football. NFL teams blitz 5 or more 30-40% of the time but only blitz 6 or more players 10-15% of the time. The playbooks make the AI wacky due to too many 6 man blitzes and too many vert routes. If you blitz all the time QB Comp% will be high but so will sacks.

For now:
Tackling slider effects broken tackles (NFL 5-6 per game): Starting 46
--Every 2-3 points of adjustment may result in 1 more or less broken tackle

Pass coverage: Sticking with 36 from last year reduces AI QB getting sacked, and leaving at 50 for more difficulty for me making passes (may even up higher if too easy)

Pass Reaction: Sticking with 91 from last year. The AI is open with lower coverage setting but the reaction time being higher means more deflections and hits that jar the ball loose keeping % down. Also in Madden17 this looked like real football to me.

[Oraeon1224]

Tackling slider:
This was most obviously broken with my sliders from last year. NFL avg 20% of carries result in a broken tackle, or 5-6 per game. I ran matched 26 plays at different sliders.


Slider = BTK
50 = 4
47 = 5
45 = 5
43 = 8
40 = 9
37 = 12 (stop point


It appears Tackling set to between 45 and 50 is realistic. I will try 46 initially.

[Oraeon1224]

By running the same routes and defense over and over and adjust sliders from extremes you can see the effects of sliders. I tested against main coverage and zone with sliders at 0,50,100 while the other slider was set at 50. The CPU is generally very reliable on same target and distances play to play with only 20% varying from a script (those I disregarded) with 20 passes thrown.

Man coverage results: Pass coverage Pass Reaction:
0 /50 /100 0 /50 /100
Space on catch: 4yd/3yd/3yd <1yd/ 3yd/5yd
Space on cut (out route): 7yd/5yd/5yd 5yd/5yd/5yd
Deflections: 1 /1 / 0 6 / 1/ 0
Sacks: 3 / 2 / 2 1 / 2 / 1

Subjective: Space on catch (i.e. screwing up the route was noticeably different when Pass coverage was adjusted (more of a buffer or lag). Speed of closure to the receiver with the ball in the air (not after catch) was noticeably different with adjustment of pass reaction.

Zone testing:
Same for all variables (less predictable passing) without any trends or even subjective findings. There were more sacks with higher pass coverage but settings by 2 sacks at 100 vs 0 but zone in general had an insane number of sacks (about 1/4 plays resulted in a sack) so I am not sure this is relevant.

Results:

To increase deflection and thus decrease completion percentages (but not necessarily interceptions) increase Pass reaction slider (NFL average 3-5 per game). To create more open receivers and potentially decrease sack chances a lower pass coverage slider might be better (opinion more than numbers) and to decrease completions and increase deflections you should increase Pass reaction (didn't cause any interceptions).

Zone appears to be responsible for most of the coverage sacks but I can't tell either subjectively or objectively how these two sliders effect Zone defense. It is entirely possible Zone is mandated by player statistics and the sliders only impact the coverage during man defense.

Since this seems relatively similar to prior years I am going to stick with my strategy of low pass coverage and high pass reaction to get realistic deflections and hopefully less coverage sacks. I will try in the future to test how pass coverage sliders effect sacks per game but that is very time consuming since it can't be labbed easily (i.e. you have to play an entire game so getting sample size is hard).

[jersam515]

I just want to let you know that this sounds like an awesome project and to wish you luck.

[Aestis]

I can tell you in M17 with statistical basis the coverage slider affected zones. Run a 32-man league and I track stats for every single User v User game across a number of seasons. Hundreds upon hundreds of games. When the passing game started to get a little high for my taste, I'd bump coverage slider and saw numbers come down. And virtually nobody ran Man in M17 vs users, so the vast majority of that effect was coming from zones.

My understanding is zones are a bit busted currently in M18 awaiting a tuner update ~mid September. Wonder if that's affecting the coverage slider.

[Oraeon1224]

It may be that it effects Zone in some way. However I can't control variables and measure it. It definitely is visible on coverage during cuts when in man.

[Oraeon1224]

Finished testing Fumble slider. Very disappointing. I think the new tackling and running changes porked this. After 25 runs with this set at 0 and another 25 at 100 there were 0 fumbles. I then ran it at 0 with holding the ball for a a sack. 1 out of every 5 times the QB fumbled on 0 and none on 100.

I am leaving this the same as last year since I can't interpret how it works now. It seems to have some effect on QB fumbling on sacks. Since I suspect fatigue and hit type impact this maybe it will work in actual games, but I can't predict how.