Comparing the final Football Outsiders F/+ projections to the AP poll and conventional wisdom.
09 Jan 2007
by Bill Moore
Let me start by saying, I love Madden Football. I've played Front Page Sports, ESPN's game, and all the other games over time, and they are all great. However, I suck. Some of it has to do with the fact that I play the games on the PC rather than a game console, but mostly it's because I suck. I spend too long staring down my receivers and don't "feel" the pocket collapsing all around me. By the time I recognize the read, my QB is inhaling turf. I can never remember what button is juke and what is jump, and I end up spinning my running back in circles rather the gaining that extra yard.
When Madden added a franchise feature, I gave up taking control of the players and focused on calling plays and managing teams. Unfortunately the AI in Madden has equivalent knowledge of my mother. That's no knock against my mom, but she's no Matt Millen.
I grew up playing lots and lots of Strat-o-matic, so ultimate franchise management in a PC game has always been the holy grail to me. Years before I stumbled across Football Outsiders, I discovered a downloadable version of a new game called Front Office Football in 1999. It had recently been licensed to Electronic Arts for distribution, and it consisted of the managerial features of Madden -- but way more. The game involved intricate play that provided a realistic approach to managing a football team without the responsibilities of being the player -- a perfect game for me. I led my New England Patriots over 25+ seasons and eight FOF Championships (back at a time when a Patriots dynasty was hardly realistic). I retired from that job and built up last-place Buffalo into a contender and finally led them to two Championships.
It was an intellectually stimulating game, but in the end, FOF was missing the graphics that added life to the game. I was not emotionally attached to my players, a process which comes from watching your quarterback engineer game-winning drives or your running back run for 2,000 yards. My hope was the EA would buy the game rights and integrate the AI into Madden. However, after two or three seasons, EA dropped the game and I thought the days of FOF were done.
When Aaron Schatz asked the staff of Football Outsiders to review a new updated version of Front Office Football 2007, I leapt at the chance. My original version of FOF is on a PC that is no longer operable, but I remembered enough about the game play to get started right away. Although the look is marginally different, I knew what to expect. However, it should be noted that a few other FO Staffers who tried FOF had difficulty figuring out how to begin. There are no tutorials and the interface isn't particularly user-friendly.
|Click picture to expand|
The game is laid out in traditional Windows format with dialog boxes that the user needs to open individually. Considering they are each important to managing the game, I kept them all open simultaneously, and I wonder why it doesn't do that automatically. My first reaction was the game didn't fit. The game requires a screen with resolution of 1024x768 or higher, but I use a Vaio laptop which has a reduced-size screen. The boxes are fixed size and do not have scroll down bars. Consequently, some boxes didn't fit on my screen. That was an annoying start. Nevertheless, I managed.
Zipping through the opening choices. I actually don't remember what level of play I picked, other than I didn't want to start with a free-for-all draft. I wanted teams to have opening day rosters, and it was pretty close. With most non-EA games, the players names are the same, but the teams need to be "forged." The Boston Colonials, The Green Bay Green Wave, the Seattle Cyberhawks all need to be manually changed to reflect the non-licensed names. I did enjoy that the AFC East team was referred to as the New Jersey team. True. True.
Getting your team ready for the season ahead offers a remarkable level of complexity at the planning level that's exciting but also remarkably overwhelming. Overall, there are no spectacular graphics, images, or start up screens, but rather just wads and wads of stats, formations, coverage options, and game planning decisions. You can leave everything with default values, but even for someone who has played before, all this information is a little intense. For someone new to the game, it must be quite daunting.
For instance, on the Depth Chart and Game Plan drop down alone, there are 37 -- yes 37 -- option and execution screens. Things such as Depth Charts under Nickel Defenses, Depth Chart under Dime Defenses, Defensive Game Plan Adjustments, Formation Use in Extreme Running Situations, and Pass Coverage Slight Running Situations. Each of those screens has tens if not hundreds of choices and decisions. For instance, under Formation Usage -- Extreme Passing Situations, you allocate the percent of plays that your team uses certain formations in situations that are heavily weighted to the probability of a pass. You must allocate from up to 18 formations -- I-formation, Normal; I-formation, TE-Pairs; I-Formation, WR moves to slot; Weak formation, Normal; Single back formation, TE-pairs, and so on. Did I mention that your quarterback has only so many formations that he knows how to run?
|Click picture to expand|
The Offensive Game Plan Adjustments is truly intimidating. On this screen, you can adjust your play calling by situation of lead and field position. There are 40 different situations from which to evaluate four different play calling actions per situation. That's 160 different adjustments for just this screen. Yikes.
Under Basic Defense Play Selection Chart, the user allocates whether to expect 'aggressive runs', run, pass, or 'aggressive passes' based on each of down and distance. It would be useful to have an option that adjusts all situations automatically with some super-sliders that would determine your level of your running or passing. Each of the Depth Chart and Game Plan options does have a "Recommend" button, but not a universal recommend button that covers all options.
Above and beyond getting updates on injuries, locker room cohesiveness, and depth strength, your scouting staff provides you a Scouting Report of your upcoming opponent, but I didn't find much value in it. I wasn't sure how to utilize the report's information to maximize my highly complex game plan. For instance, the report tells me that Atlanta's free safety is excellent at intercepting passes. What do I do with that? Run more? Yet, it also tells me that the middle linebacker is excellent at defending the run. So with all the intricacies and nuances of the game plan, I'm not sure exactly what to plan for.
A player with patience and diligence can take days (or even weeks) to plan a game or a season. However, an ADD-player like me can run an entire season in minutes. The choice is yours. You can have the AI simulate a game, a week, all games up till the playoffs, or the entire season. However, if the user is making all the decisions, the simulation will stop if depth chart decisions need to be made. Often, that can mean stopping the simulation each week.
There is a more simple approach. You can have the coach reset your depth chart and make all game planning decisions. Although this feature saves you the hassle of aligning your depth chart each week, losing control of that feature eliminates your ability to mold the team. If you want Laurence Maroney to start over Corey Dillon, you need to have control over that depth chart. Frankly, however, my teams played better (and more realistically) when I let the coach make most of the decisions. Under those same global options, you can allow the scouts to fill needed roster spots. The biggest problem with the latter is that your scouting department can end up releasing your favorite third-string running back, because the scouts believe you need a backup tackle.
As for game play, there's good news and bad news. The good news is that the AI behind the simulation is very good. It provides very reliable and reasonable results. FOF creator Jim Gindin and Football Outsiders are of like mind. Jim collected and analyzed a wide variety of statistical data and utilizes it in a similar fashion to the creation of DVOA and DPAR. Look at Jim's own words in the Help file:
"I quickly realized that basing player ratings on the basic set of common statistics was not going to provide a realistic simulation. I took the play-by-play information for all of last season's games, generating a profile of all the rushing plays by distance and direction. The net result was a grid of every running back and team matched against each opposing defense. I could then normalize each running back by modifying his averages and standard deviations against each defense. In a similar manner, I rated offensive linemen based on the average running play in their direction. Since linemen often have set plays that involve them in rushes in other locations, a portion of their rating is based on all running plays. These, too, were modified by the average runs against each defense. For defensive ratings against the run, I compiled a grid of responsibility for each defensive position. For instance, the left defensive end had 25 percent of his run rating based on his team's performance against runs around right end and only 4 percent of his rating based on runs around left end. For offensive and defensive linemen, the ratings were based on whether or not the ball carrier reached the median distance for a carry in each direction."
The rest of the Help file reads like a FO-contributed article. It discussed items such as incorporating an unknown "X-Factor" in individual game play to mimic real-life variability in performance. That X-factor determines the random element that changes the core ratings each time you start a new game. When the X-factor is 0 (a new player), the core ratings can vary by up to 50 percent from the assigned rating. When the X-factor is 100, which it is for about one-third of the players in the league, the core ratings only vary by up to 10 percent. Front Office Football generates a new universe every time you start a new game. Most of the time, a quarterback with a 50 percent rating will be a decent quarterback. A good percentage of the time, he'll be outstanding. And some of the time, he'll be fairly mediocre. Like real life, sometimes the good guys have bad games. The established players will perform pretty much as they did in the past, but there will still be some variety.
|Click picture to expand|
Although not my forte, the ability to call a team's offensive and defensive plays is a strength of the game. It too is a very intricate process, and requires a more than peripheral understanding of football strategy. You call the formation (18 choices), the type of play (run, short pass, long pass), and the play itself (one of 20 choices per setup). You call for the formation to have strength on the left or the right, and who is to pick up the blitz. The interface does an excellent job of describing and diagramming each choice and identifying the key personnel. Additionally, it allows you to toggle the personnel. Unlike Madden and many other football games, you are not held to the play clock, which I had always found an unnerving process of play selection.
The bad news is that unfortunately, "watching" the game is undoubtedly low tech. When I say low tech, I mean it. If Madden's visuals are a Cray Supercomputer, then FOF is a typewriter. You watch the play-by-play scroll past. You are essentially Owner, GM and Coach rolled into one, but you are forced to monitor the game by the equivalent of the NFL.com Gamecast. In one particular game, I called a Single-back formation, "shotgun TE cross 25, FL fly 25," in which the tight end runs a crossing route from 19 to 26 yards as the primary receiver, and the flanker runs a fly route from 19 to 26 yards as the secondary receiver. The resulting interface shows this:
New England is lined up in a Single-Back formation, aligned to the right.
The defense is in a 43 and 4-deep zone coverage.
They are keying on the pass. Hill is blitzing.
Tom Brady receives the snap from center.
Roderick Coleman simply ran through Matt Light to pressure the quarterback.
Brady was sacked by Defensive Tackle Roderick Coleman for a loss of 5 yards.
I called the play again:
New England comes out in the Single-Back formation, the right side is strong.
The defense comes out in a 43 alignment and 2-deep bump-and-run coverage.
They are aggressively expecting a pass.
Tom Brady fields the snap and drops back to pass.
Mike Patterson just blew past Logan Mankins and is all over the quarterback.
Brady heaves a long pass to Wide Receiver Steve Smith.
Patterson is credited with forcing that incompletion.
Can't you just feel the excitement! I must admit that the post-game log provides endless detail that any FO reader is sure to love. This lasts a long time if you are calling every play. Eventually, I abandoned "watching" the games altogether. I became the Billy Beane of FOF. Sure, this game could be so much better if I could really watch my new rookie running back break tackles and give stiff arms, but this game is not designed for that. It is not attempting to compete against Madden, but rather be smarter and more intricate for the football strategist looking to build a dynasty.
|Click picture to expand|
And if you love information about your team, this is the game for you. The game keeps 135 different individual statistics and 182 different team statistics. That includes FO's very own Adjusted Line Yards, on a page which also features a link to our site.
(Ed. note: Some of our readers who also play Front Office Football asked Gindin to use FO stats, so he offered us a branded page in the game if he could put one of our metrics on it. There wasn't time to work the complex DVOA equations into the game, so we went with Adjusted Line Yards. Full disclosure: Part of the deal was for us to review the game, although we probably would have done that anyway; we like reviewing games.)
Although the game play feature can be frustrating at times, the crown jewel of FOF is the AI behind the real aspect of dynasty building -- drafts and free agency. I really enjoyed the draft and the free agent process. It actually is probably my favorite part of the game. In the draft, not only do you get every draftee's combine results, but you also have the ability to interview up to 60 players. Interviewing a prospective player will yield interesting results such as "very overrated" or "as scouted." You have to interview the players prior to the draft during the free agent process. That step like many others is not intuitive and requires that you pay attention to each step, or forfeit the opportunity. Players are drafted in a realistic manner, and draftees perform in a manner similar to their draft position.
FOF tackles a traditionally tricky aspect of drafting in a fairly interesting way. The program generates players for the draft in a random fashion. Your perception of those players is solely dependent upon your scouting reports on that potential draftee. Although I did not get an opportunity to try this feature, Solecismic Software has a college simulation game in which graduating seniors and some of the most talented underclassmen will leave college and can be packaged into a draft file that one can import into FOF in place of the randomly generated amateur draft pool created each season. As a result, your draft has life that a random generator can not replicate.
Drafts may build the team for the future, but the free agent process is the way to build up your team now, and the AI for this step is among its best and unique features. Players choose the team they are willing to sign with based on money (of course, in this game like life, it's always about the money), but it also factors things such as loyalty to existing team, proximity to home, and other factors. I wish there could be some kind of dialog with agents. Having a player burn you for another team without notice is a pain, but it does happen in real life. Ask the Jets who thought they signed Antoine Winfield until Minnesota whisked him away by private plane.
|Click picture to expand|
You not only need to manage a realistic salary cap (players contracts have signing bonuses, salaries, and even voidable years), but also manage personalities. Certain players have affinity with some others, while some have conflict. Understanding a free agent's leadership qualities and personality will help your team building process as much as the understanding a player's physical attributes. When acquiring or retaining free agents, the attitude of the player toward the team (or your general reputation among your players) will factor in that player's willingness to accept your offer. The responses you receive from players when conducting transactions will help you understand your standing.
As a side note, the hiring of the scouts and coaches is a much more frustrating and vague process. The strengths and weaknesses of the staff members are laid out well. But here's the thing: 1. you can't fire your coach and hire one of your own coordinators to do the job; and 2. you can't hire someone else's coordinator as your head coach. The basis behind this seems lacking given the detail involved in player negotiation.
Similarly to a good free agent system, a realistic simulation needs a good trading protocol. If you can cheat the computer franchises out of their top players, you'll build a great team, but quickly lose interest. The computer GMs will evaluate the importance of their players on the rest of the team and factor such intangibles as fan favorite and cap impact. As a bonus, you can trade future draft picks as well as trade during the draft to move up spots. The other GMs will interact with your trade offers (and their future responses will depend upon your trading integrity), but it remains mostly a one-way dialog. Although the opposing GM will offer you feedback, the ability to seek a counter-offer would make the process less grueling.
Overall, the "game" is in building a team and leading a franchise. When I got into it, I found myself not caring about the game-to-game action, but only the finality of the season. I was most curious about how the team that I built had performed. You can run a whole season in a couple of hours, or stretch it out as long as you want to work on it. Certainly there are features that I would like to see improved, but in totality, I think the AI of this game makes for a realistic and interesting front office simulation. I enjoyed the basics of the game, but some of the intricate features were too complicated or time consuming for me. Nevertheless, I can see others getting into the minutiae of game planning and roster administration.
FOF ends up with a strange combination: it falls short in some of the basics of game design, but it is far ahead when it comes to the subtle back-end elements and realistic football details. While a game this complex is not for everyone, the great thing about FOF is that you will never feel that your intelligence is being insulted -- and there aren't many sports games where that's the case.
51 comments, Last at 26 Apr 2011, 10:38pm by xiaoz