Super Bowl XLV: Madden NFL 11 Predicts A Steelers’ Victory

Added on Feb 03, 2011 by Jack Thurman in

Super Bowl

Pro football video games have come a long way since the arcade classic Atari Football. Today’s sports video games are more serious simulation than anything else with online roster updates and an insane level of detail. When it comes to NFL football, the leader of the pack is the EA Sports ‘Madden NFL’ series and the video game manufacturer has hit the publicity goldmine with their yearly ‘simulations’ of the Super Bowl game. The fact that their yearly simulation has picked the outright winner in 6 of the last 7 years has only enhanced the media attention they’ve received. Based on their Super Bowl XLV simulation, the Pittsburgh Steelers are going to win the game outright which will no doubt come as good news to adult video enthusiasts.

We’ve got a video recap of the Madden simulation at the bottom of this post, but here’s the highlights:

The Pittsburgh Steelers will snag their seventh Super Bowl title this weekend in Dallas — at least according to Madden NFL 11.

Electronic Arts has unveiled its annual prediction for the outcome of Super Bowl XLV, based on a simulation of the game run through Madden.

The results? The Steelers hang on to defeat the Green Bay Packers 24-20, thanks to a late interception by Aaron Rodgers in the closing moments.

Steelers wide receiver Mike Wallace finished as the Super Bowl MVP with five catches for 111 yards and the game-winning touchdown.

In addition to a 6-1 SU record in the Super Bowl, the simulation has produced good results against the pointspread. Betting the outright winner in the Madden simulation has also produced a 5-3 ATS mark. If you compare the final score of the yearly Madden simulation to the Super Bowl line you did even better–using the Madden score result to determine your pointspread play has produced a 6-1 ATS record. The Madden sim hasn’t fared so well on totals–betting the Super Bowl total based on the final score of the simulation has produced a 3-4 record.

So how seriously should you take the Madden result in placing your own Super Bowl bets? Not very. The casual fan/sports bettor is always looking for some mythical ‘system’ to circumvent the hours of work, experience, statistical analysis and research necessary to make money consistently. Whether it’s Paul the Octopus picking World Cup winners or the performance of the Dow Jones Industrial Average in the last quarter predicting the Super Bowl winner the public gravitates toward their mythology. It doesn’t take a genius to realize that an octopus has no soccer handicapping ability or that there’s no causal relationship between the stock market and NFL football teams.

In theory, we respect the work that goes into making the Madden game an accurate ‘simulation’ of NFL play. In fact, many serious handicappers–myself included–use computerized game simulations as part of the handicapping process. They don’t have the intricate detail and fancy graphics–they just crunch data and come up with a score projection–but the ‘message’ is the same. There is considerable difference in how they’re used–typically, the system I use will simulate a game over 10,000 times and come up with an aggregate score. Even so, the information they generate is just the first step of the handicapping process–blindly betting on games based on these numbers is ineffective. Giving the Madden game the benefit of the doubt as to its simulation methodology, the fact remains that its only a single simulation. From a statistical standpoint, it’s meaningless.

Furthermore, computer sports simulations of any type have a fundamental flaw–the fact that sports are played by actual human beings who are subject to all sorts of emotional/subjective influences that simply can’t be replicated in any statistical model. Examples of this are countless–there’s no way a computer can account for potential player injuries, injured players playing at less than 100%, players disgruntled with coaches or teammates, exceptionally good or bad play due to motivational factors, so on and so forth. Another example–when you’re dealing with high priced favorites a statistical simulation will invariably overestimate the superior team’s margin of victory over their opponent. The reason is that the computer assumes a constant linear effort throughout the game–meaning to the computer a team will be exerting the same amount of effort to score and increase their lead when they’re up by 40 as they did when the game started. Anyone who watches sports knows that’s simply not the case. While there’s no shortage of touts that try to hawk ‘computer systems’ they are at best only a part of the handicapping process. That’s why that despite an obscene amount of technology available to the general public bookmaking remains a profitable business and they keep building those big hotels in the desert.