Now you might be asking yourself what the hell video games have to do with stats analysis, but the connections are not as convoluted as you might think. NHL video games, particularly the EA versions, have had a profound effect on a generation of young hockey fans, beginning with the blockbuster NHL '94 and making major advancements in technology/format in the '99, '01, '04, and '08. While boxcar stats had always spoken to previous generations of young hockey followers, attribute ratings have brought a new element of understanding the game to recent generations, and have likely fostered a desire for concise explanations and expressions of player values and talent. At the same time, they might also be perpetuating some of the same things we've been resisting in the stats community.
This is really only the tip of the iceberg...
Possibly the biggest change for me was the introduction of a pseudo-franchise mode in '99, where at the end of a season you had the option of moving to the next season. This was replicated from that year onward, most memorably for me in NHL 2000 for PC (though I'd been playing since NHL '96), a game which took up countless hours of my youth, vitality, and promise. Little did you know, but I could've cured cancer and water-borne illnesses had I not made that choice. We'll get 'em next time.
The excellence of NHL 2000 was that a number of players emerged from nowhere to become NHL greats, players you had pegged for greatness were simply mediocre or regressed, players 25 and over only rarely got any better than their beginning rating and production, and there was very little difference between performance among goaltenders. In other words, it was actually incredibly reflective of the real NHL. There was even an option to "fantasy draft" (essentially, a clean slate to build your team and league), or "expansion draft", where each team protected all but a handful of players whom you could select for the brand-new Atlanta Thrashers (and like a real expansion draft, there weren't elite players available). In other words, it was a playground for those video game players that enjoyed franchise modes where you simply try your hand at building a team in a variety of ways.
The game wasn't without its eccentricities. There were a few players who inexplicably dominated in any scenario, including Christian Dube, Ville Peltonen, Fredrik Lindquist, Patrice Lefebvre, and Rico Fata (maybe Fata not so inexplicably, as he was 20 at the time). Yeah, I know. Christian Dube. Players retired at odd times, and at the peak of their careers, with frustrating regularity; and unlike Brett Favre, you couldn't bring them back by sending the coach's favourite players to coo and whisper sweet-nothings into their ears. Future entry drafts included players that actually had the names and pictures of some of the creators of NHL 2000, with their exact age and birthplace, and inflated attributes that mucked with the results of simulated seasons. Trading was also way too easy, as you could essentially get a trade to pass as long as the person you were trading had a close overall rating to the player you were trading for, regardless of age. There was no salary cap, but there also wasn't a salary cap at the time. In general, you could try to exploit these loopholes, and have fun doing it anyway, but there was also the opportunity to get a good idea of player development and decline, and experience some of the hits and misses of player acquisitions and drafting. It was as enjoyable for me simulating all the games as it was playing them because of this.
The player development/decline was what really pulled me in, as from one year to the next a player could jump from a pretty average 70 overall rating to an 80 in a year, and to an 89 in two seasons. Yet it didn't necessarily always go that smoothly; you also saw players simply peak at 80, or dip back down to 75, only to rise up to an 82 the next season. More importantly, all of the players began to see some decline after the age 30, which occasionally involved as drastic a drop as you saw in the increases. At times, what brought about the change was random, but half the time it was because of performance. There were also a large number that simply didn't make it, and since there was no AHL in the game, they simply stewed on the free agent wire at overall ratings of 50 until they retired. In other words, and once again, very reflective of the NHL.
This style of game brought me to examine what made the players good in simulation, and likewise made me think about what made players good on the ice. Speed was definitely a virtue in NHL 2000, but what really made the difference was shots. If your team could generate a large volume of shots, you had a greater chance at winning the game (that unusual list of players above was actually a list of guys who generated a lot of shots). So when I first began reading Gabe's work at BTN, I already had that understanding of hockey. Right or wrong, video games placed in me (and places in others) a sort of "elementary" understanding of the game, something that I would bring with me to the new things I was learning at BTN.
Since then, the kinds of drastic and dynamic attribute changes in video games have altered, to the point that I've pretty much quit on video games like the NHL series because it has essentially snuffed out what made GM'ing so much fun. Nowadays (hold on, I forgot to take my Centrum Silver...okay...here we go)...nowadays, player attributes move nowhere; even the most promising players can have two seasons of incredible hockey and only bump from a 78 to an 80 overall rating. I'd feel pretty bad for the person who plays all those games for that kind of outcome. No player in the game makes a 10 point or 15 point jump, nor does any player make that level of a decline, either. On the other hand, and ironically, should a player have an excellent season in real life, the change in that player's attribute level from one year's version of the video game to the following might very well feature a 10 to 15 point jump. Furthermore, you have newer features like "attribute boosts" when players win fights, and a particular attribute called "poise." A quick explanation: while players have an overall rating (0-100), within that rating are ratings (0-100) of things like "speed", "acceleration", and "checking", among other attributes, that are supposed to add some individuality to player talent and combine to make up the overall rating. "Poise" was one of the more recent additions, and it's as arbitrary and ridiculous as it sounds. Yet if you have a player cranked up in all the other attributes but low on poise, their overall rating suffers badly (we're talking going from elite-NHL to third-line-AHL). "Toughness" is in there, too, and though initially I thought it referred to the ability to avoid injury, it turns out the highest-rated players in toughness are players who fight (which made me wonder why there's also a rating for "aggressiveness"). To be fair, toughness and aggressiveness were in NHL 2000, but in all the games from 2000 to present those ratings meant very little, certainly not on the level that poise means to a player's overall rating.
The point is that what was there before, for me, was a game that genuinely made me think differently about hockey, and it unconsciously made me aware of the occasional fluctuations of player (and team) performances. There are fewer such fluctuations in today's games; "good" teams attribute-wise play well, "bad" teams attribute-wise play poorly, and that's that. And they won't get substantially better or worse through drafting or trading for prospect-like players; you just have to trade for established talent, which is simply unrealistic and unreflective of today's NHL. I've seen this happen in some of the games for NFL and FIFA as well, and have even heard that attribute boosts are being sold separately, so your team can get better if you pay above and beyond what you've already thrown down for the game. So not only has the development curve for athletes moved away from the game, but it seems like the games themselves are becoming almost exploitative.
You can ignore it, and say this doesn't matter, but in a time where new generations are growing up more "plugged in" than we ever dreamed, you have to admit that video mediums (including television, NHL.com, YouTube, movies, and video games) are shaping the way the younger NHL fans understand the game. In the meantime, if I do play a video game I play college football, because they have to get better in 4 years otherwise they aren't going to get drafted (or I give their spot to Rudy Ruettiger). In that case, the video game gods decided this one time the players could improve the way they had in the past. I've sunken low, people.
Oh, and by the way, Lee Corso and Kirk Herbstreit need to cool the F down when I call for the onside kick. I know what I'm doing.