NFL Predictions
I'm not a fan of Barnwell over at Grantland. He way overvalues statistics as a predictor. Last year, he predicted a major 49er regression because most teams who jumped in performance from 7-9 to 13-3 had regressions. I thought this was an entirely stupid argument because the reason the 49ers jumped was because Harbaugh was a good coach and there was no reason to think they would regress to under-performing. His logic was akin to flipping a coin 9 times and 7 of them hitting tails and figuring the next flip therefore has a higher probability to hit heads. Wrong. The probability is still just 50/50.
This year, in his analysis of the Falcons, he points out how they perform well in close games and are due to regress because over time close game results tend to even out. Really? Wouldn't you conclude the reason they perform well in close games is Matt Ryan and Mike White know how to manage the clock and have good offensive weapons? Their performance, in other words, isn't random. I watched the 49ers all growing up during the Montana and Steve Young era. Montana always managed the clock at the end of games well. This is why he is known as the greatest quarterback of all time. Young did not. They used the same scheme and had similarly good offensive players. There was a difference between players and that's the reason Montana won 4 Super Bowls and Young won 1. It wasn't random.
In any case, here are my predictions:
1. At least one team that looks shitty on paper will be decent - Oakland, San Diego, Tennessee, Jacksonville, Arizona, Buffalo. I'm picking Tennessee, although Arizona and Buffalo have a shot.
2. At least one team that looks like a contender will suck - Denver, New England, New Orleans, Atlanta, Seattle, 49ers, Green Bay, Houston. My prediction is Denver or the 49ers, sadly. I am picking them based solely on if their quarterbacks get hurt, I think the team will really suffer and they both have a decent likelihood for different reasons.
3. Kansas City, Philadelphia, Baltimore, Cincinnati, Chicago, and Tampa Bay will all be over .500 and vying for the playoffs.
4. The most competitive divisions will be the AFC North, NFC East, NFC South. In some ways, I think any of the four teams in all these divisions could steal it.
5. I am terrified of the Seahawks.
2 comments:
Yes, there is something amiss in these statheads' writings. I don't find it an enjoyable way to experience sports. Was just reading Barnwell's latest piece, I thought this sentence summed him up pretty well - "My guess is that if you played the season one million times, one of these eight teams would win about 65-70 percent of the Super Bowls played." So, he took months and months of research to get at a beyond vague probability. I know all the professional teams are putting stats to use to analyze player performance etc, I just think there is too much an element of the fan feeling like they are an equal or superior in some way to the athlete when they can offer a snarky data driven proof that a choice in the middle of the game was not the optimal efficient one and thus the player or coach is an idiot.
it also gets at this other thing that really annoys me -- a sort of 20/20 in hindsight approach to things -- where the statistics tell an obvious story that correspond to exactly what happened during the game and offer no predictive value.
i mean, for godsake, he has this statistic of "win-total" prediction and then gauges whether a team under performed or over-performed. what the hell is the win total prediction? how many total points scored? so you're telling me a team that scores more points than the opponents has a high win total prediction? no shit. you consider this advanced statistics? and all these totals are like 9.4 wins...and the team has 9 wins. that's considered underperforming? last time i checked, you couldn't win 9.4 games. this is a problem with the statistic, and bears no relevance to the team. furthermore, teams change over seasons. last season has some (but not much) predictive value over the season before.
the whole thing is a dynamic process. injuries, strength of other teams, game-planning, fitness, execution, skill, lucky breaks, inspiration, the weather, all this stuff will always factor. the statistics are reflections of those elements which decide a game -- just like the outcome of the game itself. they don't, in other words, decide the game. but these statheads think they do. totally moronic. i feel like these dudes need to play more sports and watch less.
it reminds me of these screenwriting books which break down why movies work into a formula. the obvious question to these geniuses writing the books -- if it were so easy as following the formula -- why don't you write successful screenplays as opposed to writing the how-to manuals? the answer is obvious. it is easy to look backwards at a great movie and break down why it works -- it is easy to look backwards and break down a shitty movie and why it doesn't work. much harder to take a germ of an idea and execute into something good. an entirely different skill set.
the stats also do not take into account game plan
Post a Comment