Attributes - VPNFL logic and research

[Mitch-Oilers]

As I work on my 1983 Project, I came across the original 'README" file for the VPNFL. It describes in great detail some of the VPNFL team's insights to attributes. I admit, I did snicker where they suggest having a "narrow" rate range of "about 15 points for SP and about 20 points for AG, ST and HA" is what helps create more realism. I think our range is about 5 to 7 points. Here are some of their findings:

Attributes.

The two issues with attributes are ranges by position and ratings of individual players. Sierra has never done a systematic, good job of this. For instance, in the NFLPI97r league that ships with Sierra’s 1.08 patch, WRs Sean Dawkins (Indy) and Yancey Thigpen (Pittsburgh) have the blazing speed of guards. Seattle backup RBs Lamar Smith and Steve Broussard are rated so low that they could never have the success their real-life counterparts had last year when starter Chris Warren was running poorly. An FBPro coach using NFLPI97r would never even think of giving Smith or Broussard a try.

When it comes to ranking players – estimating where a particular player ranks with others of his position – nobody has more experience or expertise than the Kang Brothers of IKNFL fame (http://www.dimensional.com/~kang). So we started with their 1997 season and 1998 draft files.

But we didn’t leave them alone. Excellent as their work is, we were aiming at something different, at what made Tom Flores and Bob O’Connor entitle a book, Football: The Violent Chess Match. We wanted positions to work as closely as possible to the way they do on the field. For instance, we wanted it to make sense to put in extra wide receivers on passing downs and extra tight ends in short yardage. For various legitimate reasons, the IKNFL doesn’t require this the way it’s set up. For instance, WR Jerry Rice has a ST in the 80s, which means he is a far better blocker than any TE2 is likely to be.

Excellent work has been done by Rolando Jimenez and others timing sprints on the practice field and translating SP-AC ratings into 40-yard dash times. (And by the way, SP is SP and AC is AC for all positions no matter what the manual says. And anyone who has watched a linebacker bounce off a high-ST QB knows that ST is ST too. It’s our opinion that there are no “relative” attributes no matter what anybody tells you.) Basically, the results are as follows:

SP-AC 95-95 Time in 40-yard dash 4.18
Every 10 points less of SP is worth about .3 seconds
Every 10 points less of AC is worth about .1 seconds

So the fastest player in the NFL would check in at about 95-95. The fastest DE around 72-75 or so and the slowest OL in the 50-point range. The relations are linear up and down the scale. A lot of “realism-oriented” remote leagues scale their positions on precisely this scale.

Our players are considerably slower. They may look ridiculously slow to experienced FBPro users at first. But they work. Here’s why:

1) Sure there are a very few players in the NFL who run the 40 in less than 4.3 seconds. But they don't do it in pads and they don’t do it on grass. When I asked experienced football players and coaches to estimate the speed penalty for playing in such game conditions, they said it averaged .3 seconds and ranged from .1 to .5 or even more depending on a number of factors. You’ve heard of players who “play faster than they’re timed.” That’s what they mean.
2) To be able to run reasonably authentic plays you need to be able to get the snap off. Everyone is familiar with the woeful caliber of OL as opposed to DL play in FBPro. To be able to operate under center the DL must be slowed way down. That means that other positions must be slowed down accordingly.
3) “Wider is better.” We’ve found that the key to an authentic-looking passing game is the ratio of receiver (and defender!) SP to QB ST. To actually hit a slant pattern or split a zone, to catch a receiver in space and in stride means overcoming FBPro defenders’ tendency to collapse on the ball the instant it’s in the air. Slower receivers mean QBs are less likely to be throwing just as receivers come to the end of their patterns and sit a minute while the defender catches up.
4) Experiment proved that the Kangs were right: unless you want to rely on gimmicky offenses, defenders have to be scaled somewhat below their offensive counterparts. It’s the only way to produce reasonably realistic stats and scores with authentic plays.

Our other guiding principle comes from Carroll, Palmer and Thorn’s Hidden Game of Football: “The difference between a great quarterback and a merely average one is three or four plays a game.” This holds for other positions too. Most elite running backs average less than a yard per carry more than a mediocre back does (5.0 to 4.0). That means the great back is worth but a long stride more than the also-ran every time he gets the ball. So for each physical attribute at each position our range from top to bottom is considerably narrower than in any of the Sierra leagues, about 15 points for SP and about 20 points for AG, ST and HA. Yes, we know about fatigue. In fact, EN is our great fudge factor. The range from top to bottom for EN, IN and DI is more like 30 points, and EN especially helps make players heroes or goats late in the game. You can count on Barry Sanders and Terrell Davis to get yards in the fourth quarter in a way lesser backs might not. Kordell Stewart does not have impressive IN or DI, but his high EN means that late in halves and games his throws still have that zip on them, so he really is a proportionately greater threat as everyone else on the field wears down.

Which is no guarantee that Sanders, Davis or Stewart win. With the narrow attribute ranges the VPNFL really does have an “on any given Sunday” effect. Over the course of an entire season though, the advantage the best players and teams possess tends to tell.