Brian Fremeau is back for his 10th season of college football advanced stats at Football Outsiders, with a lesson on average teams (like Arkansas State) and updated FEI projections for 2016.
21 Dec 2011
by Tom Gower and Mike Kurtz
Mike: Far be it to bite the hand that feeds me, considering my history of general fantasy football success, but head-to-head matchups are really, really stupid.
Tom: It's bad enough that the NFL uses it to decide games, but when you may have hundreds of dollars riding on your fantasy team and it costs you games, then it's serious business. Especially because you might actually have the power to make a change.
Mike: I tried willing the NFL to switch over to DVOA, but sadly the force is not strong enough in me.
Tom: I didn't even try.
Mike: Probably for the best.
Tom: But I may have the power to adjust the structure of a fantasy league I play in next year.
Mike: Next thing you know, you'd be killing younglings and me and my awesome beard would cut your arm off.
Tom: Do me a favor and make it my right hand. I'm a lefty.
Mike: It's the least I can do.
Tom: So how would you organize your league if you wanted to minimize head-to-head results?
Mike: I would probably give out something like "quality" points every week. Instead of matching up teams, go by raw scoring and give points based on a baseline, while perhaps taking points away if they are below the baseline.
Tom: How exactly, though? Do all teams get quality points, or just good teams? Just good teams that lose, or all good teams? My idea is a hybrid model, where teams get a binary X/0 score based on head-to-head results, and then X to 1 points based on their ordinal rank in scoring among the league.
Mike: Teams don't lose. That's the point. There are no head-to-head matchups.
Tom: Oh, you want to abandon head-to-head completely?
Tom: I think it actually adds some interest level.
Mike: The main problem with rotisserie is that eventually the bottom half of the league loses interest. My solution to that problem is to allow a team that gets hot late some extra method to garner points and thereby get into contention. I would probably keep head-to-head for the playoffs because I still think it works very well for that.
Tom: I think that's kind of unavoidable, though.
Mike: Well, let's say that quality points are some multiplier added on to your score. So even if you're near the bottom if the standings, if you crush the rest of the league for two-to-three weeks, it could be calibrated so that moves you back into the pack and thereby into contention.
Tom: Well, one of the issues with standings, in general, is that you want there to be late-season drama without completely devaluing the early season.
Mike: The problem with pure points is that eventually you're stuck behind, due to having fewer games to gain points on relatively slim margins.
Tom: So you need to end up giving out as many points the first half of the season as you do the second half of the season.
Mike: A team could build up a huge lead early by this system, but it's unlikely the entire top half of the league does so. You get to reward overpowering teams at all phases. It just makes more teams contenders than otherwise would be.
Tom: I feel like my system does the same, and also makes each individual week more interesting. I think your idea of abandoning head-to-head completely reduces the sheer competitiveness.
Mike: I might agree if rotisserie weren't a very popular format in hockey and baseball.
Tom: Winning my first head-to-head championship the way I did, on the play that ended Monday Night Football, was awesome.
Mike: And I like head-to-head for the playoffs, as I said. If anything, making the playoffs special in that fashion would help.
Tom: I also submit that the NFL game schedule, which is much more compact than NHL or MLB's, makes it inherently more head-to-head friendly.
Mike: Not really. How many teams have finished the season out of the playoffs but near the top in points? That's just frustrating.
Tom: There are a few.
Mike: Tons, in my experience. And you still have a head-to-head aspect. How many weeks have we mentioned our points total compared to the rest of the league, win or lose? Nearly every week. In fact, you're competing against everyone, every week under this scheme, and outperforming your fellows makes your points count for more.
Tom: Well, I've mentioned it a lot only because I've been at or near the top of the league so often. I really like my hybrid system.
Mike: I don't like giving points based on ordinal rank. If the top and bottom team are separated by five points, the top is unjustly rewarded and the bottom is unfairly punished for basically identical performances.
Tom: See, I think ordinal rank will create enough variance to make the league interesting while still allowing relatively wide fluctuations in the standings. The ability to add 18-20 points to your rank in any given week makes the standings more dynamic.
Mike: I also think you're just adding another thing to complain about. People already feel cheated when they score second-best but get a loss.
Tom: There will always be something to complain about, and if not we'll invent something.
Mike: My point is that you're compounding the problem rather than solving it.
Tom: If nothing else, my system would create more randomness.
Tom: Or at least perceived randomness.
Mike: And randomness is bad. The game is already absurdly random.
Tom: I don't really mind randomness in my trivial pursuits, and think this randomness is based on interesting things.
Mike: All the developments over the years, particularly PPR, have been to undercut that randomness.
Tom: Alternatively, give X/2 points for head-to-head wins, just so they’re less overwhelming.
Mike: It might be interesting to you, but I just see it as another frustration on top of loads of random frustrations for players. Players who, in many cases, are putting up money on the game.
Tom: One man's feature is another man's bug. I suppose I should mention at this point that I've never played in a league where I had a financial stake riding on the outcome.
Mike: My league has been extremely low-stakes, when there have been stakes, but I'm not playing fantasy football to gamble. I'm playing a game at least partially based on skill, so adding randomness to it is, in my mind, going in the wrong direction.
Tom: Wait, I thought you didn't like teams in the bottom half of the league being out of it and getting bored? The only way to keep them interesting is to allow for wild fluctuations that look like randomness.
Mike: No. I just explained how amplifying success compared to the norm can allow them to stay in it. You can go on a late-season tear and work your way into the playoffs.
Tom: That's what I'm trying to do, too. I just don't quite understand how you're trying to do that.
Mike: No you're not, you're doling out points based on ordinal ranking, not actual relative performance. You're also simply adding points to the total instead of, as I suggested, using some multiplier, which means that you still hit a wall where it becomes so improbable that you could score enough points to overcome an early-season deficit.
Tom: Fine, here's an example for you. Team X gets 100 points every week. Team Y scores 50 or 150 points every week. Which team is better? Assume the average team scores 90 points.
Mike: They're basically equal.
Tom: I think that's where we disagree. I think the more consistent team is better, and I'm happy rewarding them for that.
Mike: Why is the consistent team better?
Tom: They're an above-average team all 14-ish games. The other team is only above-average seven-ish times.
Mike: But you like head-to-head systems, and Team X is probably going to lose just as many as Team Y. I'd go so far as to say more.
Tom: I like both. That's why I proposed a hybrid system.
Mike: Which would reward Team Y more than Team X, because it will get many more points for being well above-average every other week than X will get from being consistently above-average every week.
Tom: I see Team Y as very likely to end up at .500, whereas Team X is likely to end up over .500.
Mike: I think you're wrong there, and I don't think your system really does what you want it to do, anyway
Tom: I think we're now starting to see why fantasy leagues end up head-to-head: because they can't agree on a separate system that diminishes its importance.
Mike: Almost certainly.
Tom: Like democracy, it may be the worst form of determining the standings, except for all others that people have to agree on.
Tom: Highlighted by Matthew Stafford's good game, my fantasy opponent exceeded his projection for this week by 60 points. Unfortunately for him, I had Reggie Bush. And Drew Brees. And LeSean McCoy. I exceeded my projection by 55 points, and he still lost by 50. I love having a great fantasy team, and I am taking my rightful spot in the finals.
Mike: Yes, you finally are living the life.
Tom: And the good news, for dramatic purposes, is that my final opponent is the person who tried to trade me Matt Ryan and Roddy White for Brees earlier in the year, so the championship will be decided by Monday Night Football.
Mike: Dun dun duh DUUUUUUUUUUUUN.
Tom: My other fantasy team won as well, so I'm in the loser's bracket finals in that league.
Mike: Have we discussed how lame losers' brackets are, yet?
Tom: More or less, and I freely admit it's lame.
Mike: As long as we're clear on that point. My nemesis Norv Turner attempted to derail my fantasy dreams.
Mike: Hester is not actually a nemesis, and Grossman has been vanquished, languishing as ironic starter in Shanahan's Washington debacle.
Tom: Grossman, then, and NOOOORV is more of a whipping boy.
Mike: Fair enough. Anyway, Norv isn't even reliably useless. I expected the Ravens defense to clamp down and San Diego to enjoy its usual level of futility. Instead, the Chargers blew the place up.
Tom: Yeah, that was weird.
Mike: Which put me in the unfortunate position of having my WR1 playing on the winning side of a blowout and my RB1 playing on the losing side. And, of course, due to Ben Roethlisberger's well-intentioned-but-quite-stupid heroics on Monday night, the upset didn't really mean anything. But that is neither here nor there.
Tom: So did Norv actually succeed in his attempt to derail your fantasy dream?
Mike: No, but it was close. I also suffered the wrath of the Lions. 76 of my opponent's 150 points came from Stafford and Calvin Johnson, so a game that should have been a blowout ends up with me sweating David Akers on Monday Night Football. This, I should note, is a kicker-friendly league.
Tom: 3/4/5/7 friendly, or the more standard 3/4/5?
Mike: Three points per field goal up to 29, four for 30-39, five for 40-49, and six for 50-plus.
Tom: I presume that made you the only Steelers fan rooting for the 49ers to actually score a touchdown in the red zone.
Mike: Nah. I would take the fantasy loss if it meant the Steelers won. Unfortunately, the Steelers lost. Fortunately, my team eked it out, 156.48-150.34.
Tom: Congratulations on your glorious triumph.
Mike: Thank you, thank you. The other league was a laugher.
Tom: In your favor, I presume?
Mike: My lowest score was Rob Gronkowski's 7.3, and seven of my 10 slots put up double digits. Even with Chicago's abysmal showing, I won 144.88-120.5, and set up a rematch in the finals with the team that temporarily usurped my place at the top of the standings during the regular season. Interestingly enough, both of my teams next week will feature Jacksonville and Tennessee, so I'm betting on a really terrible game!
Tom: There's a chance you'll get what you're looking for.
Mike: Oh, I do not doubt it.
Tom: That said, the Titans did allow a season-high 27 points to the Colts this past week, and have the sort of anemic pass rush that even the most skittish quarterback might not find bothersome.
Mike: Jacksonville has the worst passing offense in the league by a mile. St. Louis is second at -22.1% DVOA. Jacksonville is at -40.7%.
Tom: Blaine Gabbert's name appears elsewhere in this column. I've also seen the Jaguars play a number of times. I'm well aware of what they (can't) do.
Mike: The Jaguars have a merely slightly below-average rushing attack, but Tennessee has a decent run defense.
Tom: You're just not falling for my evil scare-mongering.
Mike: Haha. I like how things are going, but we'll have to see.
Reverse Jinxes (No. 4, Elias, 6-5) 88 def. Dyscalculia Plus Ones (No. 1, Will, 8-2) 77
The only upset of the playoffs thus far! Elias did a good job of picking his starters, considering all but one of his bench came it at zero points or below, principally James Starks and Davone Bess at 0 and Green Bay DST at -3. Overall, though, this was a very underwhelming game, involving 10 single-digit slots out of the teams' combined 20. The difference was largely Tom Brady for Elias. Even so, the Jinxes need to up their game, because their championship opponent is in incendia.
That's Great Hustle! (No. 2, Sean, 8-3) 139 def. Known Chumpsky (No. 3, Rivers, 6-5) 71
It's not enough that Sean put up one of the highest points the league has seen, take a look at his bench: Matthew Stafford, 29; Aaron Hernandez: 19; Pierre Thomas, 14; and PHI DST, 20. His bench, as a team, would have defeated Elias' entire squad with just seven players! There really wasn't anything Rivers could have done against this buzzsaw, but disastrous performances by Eli Manning (four points), Shonn Greene (seven), Larry Fitzgerald (six), and Mason Crosby (one) certainly did not help.
(Editor's note: Thanks for the kiss of death last week, Mike!)
Mike: This is like a rejected Conan O'Brien segment.
Tom: Personally, I'd say that should be virtually all of them, but that may be a little extreme.
Mike: Yes, consider what is rejected while thinking about what actually gets through.
Tom: I haven't actually watched Conan more than once since my sophomore year of college. Then, I think my roommate may have watched Conan partly because he knew I didn't like it, but I woke him up every morning with my alarm because nothing wakes me up, so that probably just made us even. (Sorry, Steve, I know I sleep like the dead.)
Mike: I'm amazed you actually had a roommate for more than a few weeks.
Tom: I may talk about things like shooting a man in Reno just to watch him die, but I never do them. Plus, I've never been to Reno.
Mike: A likely story. Anyway, why does this cat need a human mouth? Cats already have mouths! With a little CGI, you can even make them move in vaguely speech-like ways!
Tom: This commercial is just a terrible idea from start to finish.
Mike: Instead, Wal*Mart decided to replace the entire mouth with an extremely creepy human mouth with human teeth. I'm not sure it's the worst idea we've seen turn into a national commercial -- It is definitely the worst execution that we have seen from a major player. Again, why the human mouth? Why?!
Tom: Why the cat? Why the yodeling? Why the human mouth? Why not verbal mention of what the commercial is for?
Mike: The cat also does a weird eye ... thing ... halfway through the commercial which isn't helped by the weird mouth -- it makes it look like a sloppy drunk. But yes, that warrants further consideration.
Tom: I spend little enough time around cats I don't know their normal eye movement when they're yodeling, but that does indeed seem unusual.
Mike: Why yodeling? Why would anyone associate yodeling with Christmas, much less the carol performed in this commercial?
I just noticed there is a "click here to get the ringtone" link underneath the video. My brain asploded.
Tom: R.I.P. Mike Kurtz, 1982-2011.
Mike: This commercial is so terrible there's no real reason to continue. Even Friskie’s World of Wonder was awful in a whimsically screwy kind of way.
Tom: I have a kind of soft spot for Friskie's acid cat food.
Mike: Oh, yeah. It's insane, but it was trying. And mostly failing, but there was some real (awful) thought that went into it, whereas this is a transparent attempt to manufacture a meme.
Tom: I guess. No chance it works, but whatever.
Mike: You cannot manufacture memes. They spring forth like a terrible Diana from the head of Anonymous.
QUARTERBACK: Blaine Gabbert might have been a successful NFL rookie quarterback on a different team. Unfortunately for him, he's on the Jacksonville Jaguars. 5 points.
RUNNING BACKS: In the blowout win over those Jaguars, Jacquizz Rodgers somehow got enough carries to avoid the penalty but not enough yards to gain more than 1 point. Behind him at 2 were LeGarrette Blount and Joseph Addai, who had fewer successful carries than Delone Carter did despite eight more attempts.
WIDE RECEIVER: Pierre Garcon and Percy Harvin each put up 0 points. Yes, somehow Percy Harvin had three receptions for eight yards.
KICKER: Mason Crosby and Billy Cundiff both earned their 0 points not by being on bad offenses, but by making two extra points and missing on a field goal.
KEEP CHOPPING WOOD: Is it unfair of us to mention players in this space like Jaguars cornerback Ashton Youboty and Bears quarterback Caleb Hanie when they’re backups playing in adverse situations? Maybe a little, but they still didn’t play well and contributed to their teams losing this week.
MIKE MARTZ AWARD: The Oakland Raiders scored a touchdown to take a 26-14 lead with 7:47 to play in Sunday’s game against the Detroit Lions. Coach Hue Jackson then decided to kick the extra point and take the 13-point lead despite it being overwhelmingly beneficial to go for two and the 14-point lead. He did not, and the Lions scored two touchdowns and two extra points to win.
COLBERT AWARD: There are obvious bold calls, like Belichick’s fourth-and-2, and less obvious bold calls, like running a shotgun draw from your own end zone against the league’s best rushing defense. Whether Mike Tomlin or Bruce Arians made the unconvential playcall, Mewelde Moore picked up 21 yards and the Steelers moved out of the shadow of their own goalposts, giving themselves a chance to get points at the end of the first half.
jipanick: Championship week is coming up and I have a chance to beat the highest scoring team and win a first title after playing with the league six years. Scoring is 1 point per ten yards from scrimmage (fractions retained), half that for passing yards, 4 points for passing tds, 6 for rush/rec tds, and -1 per turnover. No ppr. Start QB, 3WR, 2RB, Flex, TE, K, DEF. I'm set with no questions at QB, K, and DEF. So, for WR, I get three of Marques Colston (v ATL), Jeremy Maclin (@ DAL), Brandon Lloyd (@ PIT), Percy Harvin (@ WAS), Torrey Smith (@ CLE). I think Colston and Maclin are automatic, but the third slot is a problem. All three matchups are brutal.
Technically, Jason Avant is still on my bench but I have no intention of playing him. At RB, I get three of LeSean McCoy (@ DAL), Willis McGahee (@ BUF), Ryan Mathews (@ DET), Reggie Bush (@ NE). McCoy is a no-brainer, and I'm leaning towards McGahee for one other slot, but Mathews/Bush is tricky. Mathews is on a hot team that is likely to win, but is on the road against a solid fantasy run defense and has Mike Tolbert the touchdown vulture lurking. Bush has been red hot but I have a sneaking suspicion that NE is going to blow Miami away and they may not run much. Technically, I could start a fourth receiver instead. I also have Lance Ball handcuffing McGahee.
At TE, I have Jason Witten. However, the Cowboys could conceivably have nothing to play for. I can ditch my backup quarterback or Avant and grab Tony Scheffler just in case, but honestly I'm not sure that a full game of Scheffler would be better than a pre-season type of effort for Witten. There's also no guarantee Garrett will sit his starters even if NYG wins. Thoughts?
Tom: Lloyd, Harvin, or Torrey Smith? Ouch. Lloyd is out to me because of matchups. Flacco's too up and down, but I like Smith's upside potential. Cleveland's secondary is Haden and a bunch of things I think you can exploit. I think you're clearly better off with three running backs, and personally would go with McGahee and Mathews in addition to McCoy.
Mike: I'm actually going to suggest Bush, here.
Tom: Over who?
Mike: Probably over Mathews. New England's defense in general is so bad that it would take a lot of points by the Patriots offense to take the running game away, and it's just too tempting a match-up. And you know I would never recommend Bush lightly. I would go with Witten.
Tom: I can see that, but I'm always worried about him in non-PPR situations. I wouldn't think too hard about not starting Witten.
Mike: It's been a rough year for the Cowboys, so even if they are out of contention, Garrett needs his players to play hard for appearance's sake.
Flores: I somehow survived a terrible week by Wes Welker, Gronkowski, and Eli Manning to make the championships! So.... Q1) QB dilemma is killing me. Eli vs Jets is terrible, but he is playing well (of course his receivers hate him, so....). Ben Roethlisberger is hurt, STL is bad but their run-D is so awful PIT might just run it all day. Trying to get Matt Moore off waivers - NE pass-D is atrocious, so if I get him, would he be better than Eli or Ben? If I don't, who do I gamble on?
Q2: 2 RB and flex spot. I have no idea who to play. Choices are: Darren Sproles (ATL made him disappear last time), Steven Jackson (that whole team is terrible but he somehow manages productivity, when not vs SF), Mike Tolbert (keeps losing snaps to Mathews even at goal line, though I'm sure Tom is happy), Santonio Holmes (ugh. Even against a bad bad Giants D, Mark Sanchez. Ugh.), Donald Brown (my team is The Goddamnit Donalds, and starting him last week was great - is a repeat possible?), and Kahlil Bell (Caleb Hanie is horrid, but Packers are bad vs run and RBs as receivers).
Q3. SEA DEF vs SF or TEN vs JAC? SEA DEF is better, but JAC offense is atrocious. Thanks again for all your help this season. Seriously, if you guys are ever in NYC, I'll be happy to buy you guys a round of drinks!
Tom: Ben should not play. The Steelers can probably run the ball successfully against the Rams. Trust him at your own peril, even with Eli's lousy matchup.
Mike: If Ben is healthy and plays then I see no reason to avoid him. The Steelers are a pretty bad running team, so they need to throw enough passes to keep the opposing defense honest. This means Mike Wallace, and Mike Wallace has the potential to make St. Louis' secondary look silly.
Tom: Will he be healthy? I think they'll try to hide him some.
Mike: I think he will be healthier than he was on Monday, which should be enough.
Tom: If you get Matt Moore, I think I'd play him. I think he could be reliable.
Mike: I see Moore and Roethlisberger as equal propositions. I suppose Moore might be a slightly more reliable play, but he has lower upside.
Tom: I'd agree with that.
Mike: Ben could just come out firing to prove Monday was an aberration and hang four touchdowns, so we'll have to wait and see.
Tom: Start Sproles, and uh?
Mike: Jackson. Pittsburgh is merely average against passes to running backs and only slightly above-average against the run. St. Louis will stick with the run because it's the only reliable method they have to matriculate the ball down the field, and if things get hairy, there will still be a lot of dumpoffs to Jackson because Pittsburgh's secondary is lights-out.
Tom: True. Given the other options, Jackson was clearly my second choice among his options. Brown is actually first in rushing DVOA. Houston is not great against the run, and the Colts are a surprisingly effective running team at times. I was favorably impressed by what I saw Sunday, so I'd take a flier on him.
Mike: Brown is great at getting successful plays, not touchdowns. "Great."
Tom: He's a boom-or-bust back. Great DVOA, lousy success rate. I hate the other options.
Mike: Yep. Go with Tennessee DST.
Tom: I'm guessing you didn't watch Thursday Night Football: Thursday Night Edition. Play Tennessee DST.
Mike: I would've gone with them in my other league if my opponent hadn't preemptively grabbed the defenses with good matchups two weeks ago.
Aaaand we're spent. Mostly. Some of you crazy kids have Week 17 championships, and your scramble writers are here to help you with your rosters and to tell your commissioner what a fool he is. Send your questions to scramble-at-footballoutsiders.com or our handy dandy forum thread!
50 comments, Last at 31 Jan 2013, 7:55am by Luciarssix