AUTHOR'S NOTE: This article is only half-serious. All jokes about the inaccuracy of anyone's picks are purely tongue-in-cheek. We all pretty much sucked. And in the context of my Pulp Fiction reference in the article, that's an apt description.
With Fooch off fulfilling his role as SBN's Carnac of the Courtroom, and the NFL's player and ownership factions currently filming Mark Burnett's next reality show, Survivor: Millionaires vs. Billionaires, now seems as fine a time as any to dig up NN's staff predictions prior to the 2010 NFL season. To refresh your memory, 7 of us ventured last September into the shark-infested Sea of Expert Prognostication knowing full well that few, if any, of us would return with our reputations intact. Depending on your perspective, the choice was borne of either confidence or masochism. As you'll find out in this post, it looks an awful lot like the latter in retrospect.
Today, I'll be reviewing the accuracy of our 2010 predictions, which included the following:
- Win totals for all 32 teams
- Division champions
- Playoff teams in each conference
- Conference Championship game teams
- Super Bowl teams
- Super Bowl champion
To measure accuracy for win total predictions, I'll be using the same stat I used to evaluate NFL pundit predictions for 2009, mean absolute error (MAE). MAE is a very easy stat to understand, as it just takes the average number of wins we were off - in either direction - across the 32 teams. So, underestimating a team by 2 wins is just as inaccurate as overestimating a team by 2 wins, and being off by 64 total wins across the 32 teams means an average inaccuracy (i.e., MAE) of 2.00 wins. Simple enough, I think.
For descriptive purposes, I'll also list how many of the following each of us had:
- Hit - Win total prediction for a team was exactly right
- Near Miss - Win total prediction for a team was off by 2 wins or fewer
- Barn Miss - Win total prediction for a team was off by 4 wins or more
For the other predictions, I'll just look at accuracy as a percentage, such that correctly picking 8 of the 12 playoff teams is 75% accurate, correctly picking both of the Super Bowl teams is 100% accurate, and so on.
Finally, I'll do my best Casey Kasem impersonation, presenting all this stuff as a countdown going from our least accurate NN writer to our most accurate (more precisely, our least inaccurate). For each, I'll give his (in)accuracy stats, and detail some of his best and worst picks; the latter of which, I have to admit, is primarily for comedic purposes.
After the jump, the "Why We Haven't Retired on our Vegas Incomes yet" countdown...
Just as a point of reference before the hilarity ensues, it's useful for you to know the following "mindless prediction" accuracies:
- If you had just predicted 8 wins for every team in 2010, your MAE would have been 2.56.
- If you had just predicted every team in 2010 to match their 2009 win total, your MAE would have been 3.31.
#7 - Ninjames
MAE: 3.31
Hits: 2
Near Misses: 8
Barn Misses: 11
Division champs: 12.5% (Colts)
Playoff teams (Correct picks): 50.0% (Colts, Ravens, Jets, Falcons, Saints, Packers)
Conference Championship game teams: 0.0%
Super Bowl teams: 0.0%
Super Bowl champ: 0.0%
Best picks (Prediction/Actual): Packers (10/10), Dolphins (7/7), Cardinals (6/5)
Worst picks: Bears (2/11), Bengals (12/4), 49ers (14/6), Buccaneers (2/10)
From the looks of it, Ninjames was either drinking heavily the night he worked on his predictions or he was employing an office-pool-esque strategy of trying to not pick what all the other guys are picking. Game theory says that can be a pretty successful strategy (the office pool one, not the drinking one). In this case, unfortunately, it was an abject failure. I'll give him the benefit of the doubt though, and assume that there actually was a strategy at work here.
In comparison with our "mindless prediction" accuracies, Ninjames would have been nearly a full game more accurate per team if he had simply drank 8 shots and succumbed to the coincidence of that number. As it is, he somehow achieved a much more impressive mathematical coincidence. Namely, despite not actually using the 2009 win totals to make his 2010 predictions, he somehow matched the inaccuracy of that strategy exactly. It's as if 32 numbers were on his computer screen, then the drinking started, then the double vision set in, and then the predictions were made. So maybe I should take back that whole benefit of the doubt thing I said a minute ago.
Unfortunately, there really isn't much in the way of positive results for me to pat Ninjames on the back for here. Although hitting their win totals on the number was certainly accurate, it didn't exactly take Nostradamus to know that GB was going to be an above average team or that MIA was going to be an average one. In my mind, his best pick was accurately predicting the decline of ARI without Kurt Warner. If only it were the case that inaccurately predicting the stratospheric rise of SF with Alex Smith didn't more than offset that one. But hey, it's the Niners, so I guess I'll have to give him a pass on that one.
#6 - Andrew Davidson
MAE: 2.94
Hits: 3
Near Misses: 9
Barn Misses: 11
Division champs: 25.0% (Patriots, Colts)
Playoff teams: 58.3% (Patriots, Colts, Ravens, Jets, Falcons, Saints, Packers)
Conference Championship game teams: 25.0% (Packers)
Super Bowl teams: 50.0% (Packers)
Super Bowl champ: 100.0% (Packers)
Best picks (Prediction/Actual): Lions (6/6), Dolphins (7/7), Jets (11/11), Raiders (7/8)
Worst picks (Prediction/Actual): Bengals (11/4), Buccaneers (3/10), Bears (5/11), Cowboys (12/6), Chiefs (4/10)
First off, two things: (a) kudos to Andrew for correctly picking GB to win it all, and (b) notice that Ninjames was nearly a half-a-game worse than the next-least accurate NN writer. With that said, let's not lose sight of the fact that, like Ninjames, Andrew would have been better off just predicting 8 wins for every team.
Also like Ninjames, Andrew nailed MIA. Well, not the whole city. That'd be way more impressive what with the model contingent down here. He also did pretty well for himself picking an unusually not-horrible season from OAK. In terms of his bad picks, Andrew was again channeling Ninjames for his CIN, TB, and CHI win totals, and you can't really fault anyone for missing the barn on DAL and KC, seeing as how few people saw their 2010 seasons coming.
Aside from correctly predicting GB to win it all, the other highlight of Andrew's predictions was correctly picking 7 of the 12 playoff teams. That 58.3% accuracy rate tied him with 2 others who ended up much higher in the countdown. The same thing goes for having correctly predicted 4 of the 6 AFC playoff teams, and 3 of the 6 NFC playoff teams.
#5 - Drew K
MAE: 2.78
Hits: 4
Near Misses: 14
Barn Misses: 11
Division champs: 12.5% (Colts)
Playoff teams: 50.0% (Colts, Ravens, Jets, Eagles, Saints, Packers)
Conference Championship game teams: 50.0% (Jets, Packers)
Super Bowl teams: 50.0% (Packers)
Super Bowl champ: 100.0% (Packers)
Best picks (Prediction/Actual): Browns (5/5), Dolphins (7/7), Seahawks (7/7), Redskins (6/6), Eagles (9/10)
Worst picks (Prediction/Actual): Bengals (11/4), Buccaneers (3/10), Patriots (8/14), Steelers (6/12)
Hopefully, you're starting to see a pattern emerging. So far, we were horrible at predicting division champs, we all basically got the same playoff teams correct, we hammered our Super Bowl champ picks, and we either hit the double bullseye or missed the dartboard entirely with teams like MIA, CIN, and TB. In fact, if I had more space, I would have shown you that Drew, like pretty much everyone else, also missed the barn on the Bears, Cowboys, and Chiefs.
Where Drew differs from the pack was in his clairvoyance regarding PHI. Indeed, he was the only one of us who correctly predicted PHI to make the playoffs. I'm not sure whether or not his crystal ball told him Michael Vick would take the starting QB job from Kevin Kolb, but one can reasonably infer that he didn't think the loss of Donovan McNabb was going to affect the Eagles all that negatively; and he was right.
However, a famous quote by Harvey Keitel's Pulp Fiction character, "The Wolf," - which I won't repeat here for NSFW reasons - comes to mind right about now (for those unaware, it's basically a euphemism for prematurely congratulating each other for a job well done). Or, to cite a more wholesome film, Drew's picks were like Titanic. In the beginning, everyone was in awe of a great human feat, and all would have ended well if not for the whole shoddy-engineering-plus-iceberg-equals-hell-on-earth thing that everyone ignored. Sure PHI was a great pick, but Drew's ship sunk when he picked NE to win only 8 games and PIT to win only 6. It's one thing to underestimate the 2 teams that ended up being the top 2 seeds in the AFC. It's quite another altogether to have the winningest team in the NFL and the AFC's Super Bowl representative both having non-winning seasons, thereby missing the playoffs entirely. If only the voyage had ended before that Celine Dion song.
#4 - David Fucillo (Fooch)
MAE: 2.69
Hits: 4
Near Misses: 9
Barn Misses: 9
Division champs: 25.0% (Patriots, Colts)
Playoff teams: 50.0% (Patriots, Colts, Ravens, Falcons, Saints, Packers)
Conference Championship game teams: 25.0% (Packers)
Super Bowl teams: 50.0% (Packers)
Super Bowl champ: 0.0%
Best picks (Prediction/Actual): Ravens (12/12), Bills (4/4), Browns (5/5), Chargers (9/9)
Worst picks (Prediction/Actual): Buccaneers (2/10), Bengals (9/4), Bears (6/11), Panthers (7/2)
As you can see, Fooch's results pretty much fit the 2010 NN pattern, so I'll again just point out the differences. First, Fooch's MAE was slightly better than Drew's even though Drew had just as many hits and 5 more (!!!) near misses. This is because Fooch's other misses weren't as bad as Drew's. In short, Fooch's picks were less erratic overall than Drew's.
Second, although I've listed some teams among his best and worst picks that you haven't seen yet, where Fooch stands out is in his pick for TB, a team that we all had problems forecasting. Specifically, his TB pick was even worse than Ninjames's! Luckily for Fooch, this was his only really atrocious pick, which can't be said for our resident Golden Nuggeteer.
Aside from these 2 idiosyncrasies, Fooch's picks didn't really stand out from the crowd. The old saying goes, "slow and steady wins the race." In Fooch's case, however, it looks like "slow and steady won 4th place."
#3 - urnext
MAE: 2.56
Hits: 3
Near Misses: 10
Barn Misses: 8
Division champs: 37.5% (Patriots, Colts, Falcons)
Playoff teams: 50.0% (Patriots, Colts, Ravens, Falcons, Saints, Packers)
Conference Championship game teams: 0.0%
Super Bowl teams: 0.0%
Super Bowl champ: 0.0%
Best picks (Prediction/Actual): Packers (10/10), Saints (11/11), Redskins (6/6), Cardinals (6/5), Raiders (7/8)
Worst picks (Prediction/Actual): Buccaneers (3/10), Panthers (8/2), Cowboys (12/6), 49ers (11/6)
urnext's picks stand out in a few ways. First, his MAE matched the accuracy rate for our "mindless prediction" of 2010 duplicating 2009. Of course, this just means that only 2 of us actually had picks that could be considered better than mindless (golf clap). Second, urnext was the only NN writer - cue sarcasm - to get a whopping 3 of 8 division champs correct. The real genius of this was being the only one to not pick NO to win the NFC South. Finally, and most impressively, urnext was able to finish in 3rd place despite being the only one besides Ninjames to post a goose egg with his Conference Championship game and Super Bowl picks. Ninjames was drunk enough to pick 14 wins for SF, but even that level of inebriation didn't lead to a "Niners win it all" pick ala urnext. And for that, it is hereby decreed that, under penalty of ban-hammering, no one is to question Ninjames's alcohol tolerance or urnext's Niner fandom in...Any. Comment thread. Ever.
#2 - wjackalope
MAE: 2.50
Hits: 3
Near Misses: 12
Barn Misses: 10
Division champs: 25.0% (Patriots, Colts)
Playoff teams: 58.3% (Patriots, Colts, Ravens, Jets, Falcons, Saints, Packers)
Conference Championship game teams: 25.0% (Packers)
Super Bowl teams: 50.0% (Packers)
Super Bowl champ: 0.0%
Best picks (Prediction/Actual): Browns (5/5), Saints (11/11), Raiders (8/8), Cardinals (6/5),
Worst picks (Prediction/Actual): Buccaneers (4/10), Bengals (9/4), Cowboys (11/6), Chiefs (5/10)
It's true that wjackalope's MAE was atypically above the mindless zone. However, his overall results earn him the title, "Dilettante of Niners Nation": jack of all trades, master of none. Specifically, his hits, near misses, and barn misses were about as good as urnext's, his correct playoff picks up through the Super Bowl champ were as good as Andrew Davidson's, he had GB losing the Super Bowl like Fooch did, and his best/worst picks were populated with the same teams as everyone else's. Literally, the only thing exceptional about wjackalope's predictions is that he was the only one to perfectly gauge the power of Tom Cable. Come to think of it, this fits perfectly, as the full phrase is, "Jack of all trades, master of none, though often times better than master of one." Congratulations, wjackalope. You are the master of Tom Cable.
#1 - Florida Danny
MAE: 2.47
Hits: 5
Near Misses: 12
Barn Misses: 10
Division champs: 12.5% (Colts)
Playoff teams: 58.3% (Steelers, Colts, Ravens, Jets, Falcons, Saints, Packers)
Conference Championship game teams: 50.0% (Jets, Packers)
Super Bowl teams: 50.0% (Packers)
Super Bowl champ: 100.0% (Packers)
Best picks (Prediction/Actual): Ravens (12/12), Colts (10/10), Jaguars (8/8), Chargers (9/9), Titans (6/6)
Worst picks (Prediction/Actual): Panthers (9/2), Bears (6/11), Patriots (9/14), Rams (2/7), Buccaneers (5/10)
Although well within my rights as champion, I'll refrain from gloating (in the first sentence). Instead, I'll just point out that (a) I was the only one to correctly pick PIT as a playoff team, and (b) unbeknownst to me prior to the season, I'm some kind of savant when it comes to the AFC South. So yeah, I had the Patriots missing the playoffs and winning only 9 games, and sure I might have thought CAR drafted Sam Bradford and traded Matt Moore to STL, and yeah, I technically was only slightly more accurate than Pick Bot 2009. Fine, bring all that up if you want. All I have to say is, "Boom, crush. Night, losers. Winning, duh."
Just for your own edification, here's a table showing how we did on average as a group, using our average error for each NFL team, and the consensus of picks for division champs, playoff teams, etc. (NFL teams are sorted by most to least accurate average NN writer prediction):
|
|
NN Consensus |
Actual |
Error |
|
BUF |
4.14 |
4 |
0.14 |
|
CLE |
4.71 |
5 |
0.29 |
|
NYG |
9.43 |
10 |
0.57 |
|
NO |
11.71 |
11 |
0.71 |
|
WAS |
6.71 |
6 |
0.71 |
|
BAL |
11.14 |
12 |
0.86 |
|
NYJ |
10.14 |
11 |
0.86 |
|
SEA |
6.00 |
7 |
1.00 |
|
MIA |
8.14 |
7 |
1.14 |
|
DET |
4.86 |
6 |
1.14 |
|
GB |
11.29 |
10 |
1.29 |
|
TEN |
7.43 |
6 |
1.43 |
|
SD |
10.71 |
9 |
1.71 |
|
OAK |
6.29 |
8 |
1.71 |
|
ARI |
6.86 |
5 |
1.86 |
|
IND |
12.00 |
10 |
2.00 |
|
JAX |
5.86 |
8 |
2.14 |
|
PHI |
7.86 |
10 |
2.14 |
|
DEN |
6.86 |
4 |
2.86 |
|
ATL |
10.00 |
13 |
3.00 |
|
HOU |
9.57 |
6 |
3.57 |
|
PIT |
8.29 |
12 |
3.71 |
|
STL |
3.29 |
7 |
3.71 |
|
NE |
10.14 |
14 |
3.86 |
|
MIN |
10.00 |
6 |
4.00 |
|
KC |
5.57 |
10 |
4.43 |
|
CAR |
6.86 |
2 |
4.86 |
|
SF |
10.86 |
6 |
4.86 |
|
DAL |
11.00 |
6 |
5.00 |
|
CHI |
5.57 |
11 |
5.43 |
|
CIN |
9.57 |
4 |
5.57 |
|
TB |
3.14 |
10 |
6.86 |
|
MAE |
|
2.61 |
|
|
Hit |
|
7 |
|
|
Near Miss |
|
15 |
|
|
Barn Miss |
|
8 |
|
|
AFCE |
NE |
NE |
YES |
|
AFCN |
BAL |
PIT |
NO |
|
AFCS |
IND |
IND |
YES |
|
AFCW |
SD |
KC |
NO |
|
NFCE |
DAL |
PHI |
NO |
|
NFCN |
GB |
CHI |
NO |
|
NFCS |
NO |
ATL |
NO |
|
NFCW |
SF |
SEA |
NO |
|
DIV% |
|
25.0% |
|
|
PLAY1 |
NE |
NE |
YES |
|
PLAY2 |
CIN |
PIT |
NO |
|
PLAY3 |
IND |
IND |
YES |
|
PLAY4 |
SD |
KC |
NO |
|
PLAY5 |
BAL |
BAL |
YES |
|
PLAY6 |
NYJ |
NYJ |
YES |
|
PLAY7 |
ATL |
ATL |
YES |
|
PLAY8 |
MIN |
CHI |
NO |
|
PLAY9 |
DAL |
PHI |
NO |
|
PLAY10 |
SF |
SEA |
NO |
|
PLAY11 |
NO |
NO |
YES |
|
PLAY12 |
GB |
GB |
YES |
|
PLAY% |
|
58.3% |
|
|
AFCC1 |
BAL |
PIT |
NO |
|
AFCC2 |
IND |
NYJ |
NO |
|
NFCC1 |
DAL |
CHI |
NO |
|
NFCC2 |
GB |
GB |
YES |
|
CC% |
|
25.0% |
|
|
SB1 |
BAL |
PIT |
NO |
|
SB2 |
GB |
GB |
YES |
|
SB% |
|
50.0% |
|
|
CHAMP1 |
GB |
GB |
YES |
|
CHAMP% |
|
100.00% |
|