Significant post from Joe Stallings...

[Ben E Lou]

He posted some screenies today. I know that many of you will consider most of 'em fluff. However, this post shouldn't get lost in the shuffle:

One of the debug features (to help measure the long-term stability of the game engine) is that at the end of every season, we output a text file that shows you the league's per-game averages for major stat categories. Not sure if we will keep it in the final release, but it's interesting stuff. Here is the 1993 output:

All Stats are per game

Total League Stats
Total Points, 43.203125
Total Yards, 676.85546875
Total Plays, 129.98828125
Total Yards Per Play, 5.20704991435526
First Downs (Pass), 21.14453125
First Downs (Run), 13.31640625
First Downs (Pen), 2.52734375
Third Down Conv, 10.859375
Third Down Att, 28.73046875
Third Down %, 0.377974167233175
Fourth Down Conv, 0.56640625
Fourth Down Att, 0.8828125
Fourth Down %, 0.641592920353982
Penalties, 10.046875
Penalty Yards, 77.02734375

League Passing Stats
Attempts, 65.0546875
Completions, 34.24609375
%, 0.526420079260238
YPA, 6.52642007926024
Yards, 424.57421875
TDs, 2.65625
INTs, 1.9765625
Sacks, 4.85546875

League Rushing Stats
Rushes, 60.078125
Yards, 252.28125
YPA, 4.19921976592978
TDs, 2.03515625
Fumbles, 3.734375
Fumbles Lost, 0.93359375

League Return Stats
Kick Returns, 7.375
Kick Return Yardage, 160.73828125
Kick Return Ave, 21.7950211864407
Kick Return TDs, 0.01171875
Punt Returns, 7.56640625
Punt Return Yardage, 71.23046875
Punt Return Ave, 9.41404233350542
Punt Return TDs, 0.02734375

FG Stats
FG Converted, 2.953125
FG Attempts, 3.796875
FG %, 0.777777777777778
XP converted, 4.75
XP attempted, 4.87890625
XP %, 0.973578863090472

Punting Stats
Punts, 11.21875
Punt Yards, 487.90625
Punt Ave, 43.4902506963788
Net Punting Yards, 388.2734375
Net Ave, 34.6093314763231

Defense Stats
Defense TDs, 30

[Ben E Lou]

Oh....and the screenies can be found here

[Dutch]

Wow that game is looking gorgeous! And hopefully those numbers all equate to NFL styled numbers. They look pretty on spot at first glance.

[RawIsDan]

They do look good.

[The_herd]

They look pretty good. The only that that jumps out at me is the completion %, looks too low for today's NFL.

[maximus]

SkyDog, I posted in that thread and I voiced my opinion. I say Joe and Arlie keep those stats in the final version. This could work to our advantage on several fronts. De-bugging but also utilities and record keeping, ect.

[sabotai]

The only problem I see is that kick return TDs and punt returns TDs are extremely low.

0.02 punt returns per game would equate to 1 every 50 games, and there's a kick return for a TD once every 100 games.

Now..that's just my speculation. It may be that the NFL numbers are that low too. It's just doesn't seem like that (seems like there's at least one PR or KR taken for the distance each week). I'm too tired right now to check it out.

[Eaglesfan27]

I must admit, the games looks gorgeous, and my plan to wait until various people on here review it before buying it is flying out the window. Hopefully, it will have a three day demo which I will get as soon as available and likely buy the game if it plays at least as half as good as it looks.

[Buccaneer]

He said 1993. Is it accurate for that year? I would suspect that 2003 would be different - or is the test year arbitrary. Again, I would bring up the same issue when we discussed this with FOF5. Do you want a model that gives the results within a narrow margin year after year after year with no potential for historical or future differences?

[Ben E Lou]

Originally posted by Buccaneer
Do you want a model that gives the results within a narrow margin year after year after year with no potential for historical or future differences?

Yes. One qualifier--fairly narrow.

The problem (and Arlie and I have discussed this at length and agree) with a changing model is the very real concern that it could veer off tremendously in one direction over many seasons, potentially rendering one portion of the game (say, running or passing) useless. This happened with one version of Baseball Mogul. After a period of time, it would inevitably go into an dead-ball era that it couldn't extricate itself from--so much of a deadball era that ERA's were pretty much all under 1.00, and a rare player would hit over .100.

[maximus]

SkyDog, what do you think? Should Joe and Arlie keep this feature?

EDIT: ....Beyond the beta phase that is? I would like for them to keep it. There are several uses for it. IMO

[Ben E Lou]

Originally posted by maximus
SkyDog, what do you think? Should Joe and Arlie keep this feature?

EDIT: ....Beyond the beta phase that is? I would like for them to keep it. There are several uses for it. IMO

I have no clue why they'd drop it. It is a text file output in the main game directory that takes up virtually no HD space.

[The_herd]

I didn't notice that it was 1993 until you mentioned it. If the game can realistically handle an evolution of offenses over several years that would without a doubt be a huge step for football text sims. Regardless, I was just pointing out that it was low for the way football is played now. In 1970 the league completion % was 51.1, it jumped up to 56.1 in 1980 and stayed there until the 90's when the West Coast offense became prevalent. In 2000 the completion % was 58.1 because 75% of the teams now run a version of the West Coast offense.

As I said, if the game can handle an evolution like that, its great for us, otherwise, I was just pointing out its a bit on the low side.

[maximus]

Originally posted by SkyDog
I have no clue why they'd drop it. It is a text file output in the main game directory that takes up virtually no HD space.

But you did see where Joe said he wasn't sure if they were going to keep it, correct? Well, if you can... try talking them out of it. I don't want to see it gone.

[Buccaneer]

Originally posted by SkyDog
Yes. One qualifier--fairly narrow.

The problem (and Arlie and I have discussed this at length and agree) with a changing model is the very real concern that it could veer off tremendously in one direction over many seasons, potentially rendering one portion of the game (say, running or passing) useless. This happened with one version of Baseball Mogul. After a period of time, it would inevitably go into an dead-ball era that it couldn't extricate itself from--so much of a deadball era that ERA's were pretty much all under 1.00, and a rare player would hit over .100.

I would suspect that my level of comfort would be a slightly wider margin. It sort of aggravated me when folks complained of having 6 4000+ yds passers in FOF5 instead of the expected 2 or 3. But I certainly agree that what you said about the one version of BM in being totally unacceptable. I'm just really stuck on the different default era settings in OOTP but knowing also that it might be much harder to do in a football game. However, I would be willing to predict that sometime in the future, 6 4000+ yds passers would be the norm in real life. So why not include that in the margins instead of thinking that a football model can only mean 2001-2003 stats? But part of this is also looking to the past. Couldn't some of the more defensive minded years be included in the margin as well?

[Buccaneer]

dola

the_herd: Good post.

[Philliesfan980]

I have no idea if this question has been answered or not, so please forgive me for the lazyness, but is this game going to have a database import function similar to the OOTP series? And if so, is anyone going to make a leihman type database for this game?

[Buccaneer]

Phillies: That's a great question. I recall that Draft Dodger was working on something like this. To me, if not only can we import a full 1970 roster but in a 1970 league structure (with perhaps the bonus of 1970-era stats) would be my ultimate wish.

[Comey]

I've heard that we can look for something similar to the Lahman Database in 2004. Don't know an exact ETA, but it's something to look forward to!

[Ben E Lou]

Originally posted by Buccaneer
However, I would be willing to predict that sometime in the future, 6 4000+ yds passers would be the norm in real life. So why not include that in the margins instead of thinking that a football model can only mean 2001-2003 stats? But part of this is also looking to the past. Couldn't some of the more defensive minded years be included in the margin as well?

I agree with this part in theory completely, but the problem is making it work in practice. If the engine could have a nice ebb and flow over time, moving from a defensive era to an offensive era to a running-oriented era to a passing-oriented era, then by all means go for it. OOTP is the only text sim that currently even makes any sort of attempt to model different eras, and as you know, it is a manual setting, not something the engine can do on its own.

Considering it has never been done (although surely considered), I think it would be too tall an order to expect a "dynamic" stats engine such as this in a first-generation sim. I do, however, think it would be fair to push Jim, Markus or Shaun to go for a dynamic stats engine in their next releases. Of course, the other problem is the time-spent vs. what-is-gained issue. I'd imagine it would take quite a bit of tweaking to get it right (and avoid the dreaded point-of-no-returns of BM), but that it wouldn't generate that much more in terms of excitement or sales.

Just my $0.02.

[Ben E Lou]

Originally posted by Buccaneer
Phillies: That's a great question. I recall that Draft Dodger was working on something like this. To me, if not only can we import a full 1970 roster but in a 1970 league structure (with perhaps the bonus of 1970-era stats) would be my ultimate wish.

They've stated publicly that league structure (teams/divisions) is going to be static for the release, but that they're working in flexible league structures for a patch. I haven't heard anything different, so I assume that is still the case.

[Buccaneer]

Ben: I agree with you about the dynamic stats engine and the complexity it involves. It was more hypothetically speaking.

Also, I hope that a flexible league structure would come sometime but I know full well that it would add to the headaches of what they got now.

[kcchief19]

When it comes to statistics simulation, all I want is a "realistic model." If that means that the game has a static statistic model based on real-life 2003, I'm fine with that. What I want to be able to do is play the game out of the box like it is 2003 without having to do anything.

Now, if a game wants to be ambitious and give me a chance to play in different eras, I'm fine with that too -- if it works and is easily understood by someone without a degree in mathematics. The OOTP era engine is an example of a poor engine. If I click that I want to play in the deadball era, all I have really done is click a button that doesn't actually do anything. If I want to recreate the dead-ball era, I have to break out a slide rule, an etch-a-sketch and a pot of coffee every single sim year.

If I'm playing a sim that purports to recreate historical eras, this is what I want: At the start of the game, I type in the year I wish to use to start. The game automatically uses a profile of that year to sim the games. The game then updates the profile automatically without me having to calculate anything. It just does it. Once you get past 2003, it would be nice if the game evolved as well, creating new profiles within a small deviation. Maybe in 2010 everybody is running the Rams/Chiefs offense, and by 2025 maybe its reversed back to an era dominated by defense. That's fine. A static future is fine.

If you want do something like OOTP and give the user flexibility for modeling their own eras, that's fine. But don't sell it as something that it is not. In the feature list, I want it to say, "Statistic majors will enjoy calcuating ratios create their own football eras. This feature will be useles for most of you, so please do not buy this game for this feature if you do not have a statistics degree."

[kcchief19]

By the way, this is the kinder, gentler kcchief. How did I do?

[Ben E Lou]

Originally posted by kcchief19
If you want do something like OOTP and give the user flexibility for modeling their own eras, that's fine. But don't sell it as something that it is not. In the feature list, I want it to say, "Statistic majors will enjoy calcuating ratios create their own football eras. This feature will be useles for most of you, so please do not buy this game for this feature if you do not have a statistics degree."

LOL. You're right about that. Changing the statistics manually in OOTP is the most bizarre process I've ever seen.

[Ben E Lou]

Originally posted by kcchief19
By the way, this is the kinder, gentler kcchief. How did I do?

Pretty boring until the last paragraph.

[Buccaneer]

The OOTP era engine is an example of a poor engine. If I click that I want to play in the deadball era, all I have really done is click a button that doesn't actually do anything. If I want to recreate the dead-ball era, I have to break out a slide rule, an etch-a-sketch and a pot of coffee every single sim year.

kc, that's the second time you have said that recently and you are pissing me off again (in a politely respectful way, of course ). You know that I am not a fan of stats and numbers but I find playing in any OOTP era to be easy and fun. What I do is start in a year and just make sure that the right era is set (e.g., 1901-1920 = Dead Ball Era). What's so hard about that? Maybe you are talking about wanting to start in 2003 and simulating the dead-ball era. Why would anyone want to do that or expect it to be realistic? The era's stats are tied directly to the players that are imported or generated. You can't change the 1934 Canton Bulldogs into the 1999 Rams. This probably goes along the same lines as expansion which I find one of the most fun things to do in OOTP. In my recent career, I went through all of the 1960s expansions and changes and it was easy and great to do! But then again, folks like me love that kind of stuff.

[Ben E Lou]

Originally posted by Buccaneer
He said 1993. Is it accurate for that year? I would suspect that 2003 would be different - or is the test year arbitrary. Again, I would bring up the same issue when we discussed this with FOF5. Do you want a model that gives the results within a narrow margin year after year after year with no potential for historical or future differences?

I thought they'd made this public, but couldn't find it. I wasn't looking in the right place. This is from the main feature list of TPF, published on the front page of the .400 main site:

Fast, Accurate Pro Football Game Engine

Developed from the ground up, based on 2000-2002 National Football League* game results. Simulate an entire season in minutes.

[kcchief19]

Bucc, what I'm saying is that if I want to start in 1910, OOTP will do a nice job in adjusting some things, such as bunting, pitching staffs and the like. But recalculating the stats to recreate the dead ball era accurately is extremely difficult and time consuming. The recalculate stats buttons in my experience do not work as designed.

If you want to play in the dead ball era, you can have a three-man staff, but you'll also end up with out-of-whack statistics for the era. The margin of error is too great to make me feel like I'm actually playing in that era.

Now, it can apparently be done, but it involves putting in all the numbers, and you also have to account for league sizes and other factors. What I want from a game promising historical replay is actual true-to-life historic replay without an hour or work. In fairness, I think FOF is the only game to get that right -- no other sim I've played produces statistics as dead-on as FOF. Not OOTP, not Baseball Mogul and certainly not a Madden or such. Sierra Football Pro was up there, though.

I still play OOTP because it's a fun game, but the historic features of the game are non-functioning for me because the time it requires to do expansion and other such stuff is too boring and time consuming for my tastes.

In other words, it's kind of like how you feel about TCY.

[Buccaneer]

kc, fair enough.