2014 Fantasy Football in Review: Quarterback Consistency

People like to complicate things -- myself included.

And one of the most confounded, bebothered, and confusticated aspects of our lives -- fantasy football -- comes down to two things: player selection and strategy. They aren't mutually exclusive, but they are different.

If you can peg guys such as Odell Beckham or C.J. Anderson without being fooled into wasting your waiver wire claims on Donald Brown, then you don't really need strategy because you're omniscient. If you knew to draft Antonio Brown over Calvin Johnson, then you're a smarter guy or gal than I am. Might be a sorcerer, too.

I'm not against changing my ways, admitting I'm wrong. But I don't have the foresight to do anything but read into available information and make the best decision I think exists. So, for me, a guy who is playing a game that doesn't change much -- I mean, really, how different are league settings now than they were 10 years ago? -- all that I can do is adjust to the trends.

It's a passing league, now -- right? -- so drafting wide receivers with high ceilings is actually a smarter play than a decade ago. But while most leagues remain head-to-head, is opting for a more volatile receiver over a running back really the right choice? Do wideouts really offer a better weekly advantage than running backs?

That's what I'm going to delve into: consistency. And it'll be a long journey, but I think we'll learn a lot. It's the offseason, after all, so I'm sure you have about 20 hours a week freed up now that you aren't doing as much homework.

Defining the Process

I'll foreground some stuff, mainly shortcomings of my research because I'm aware of them.

I looked only at 2014. I don't really have the means to dig much deeper back into history and see if these trends hold up year after year, so I did what I could.

I also am not including Week 17, which I think is a positive. I factored out games missed, too, which does limit the data in a sense (if you want to apply them to season-long strategy), but giving players zeroes will trash consistency scores even though you weren't playing them.

For this installment, I took the top-24 quarterbacks in terms of total fantasy points after Week 16 and found their weekly fantasy points -- not their average, but their actually weekly scores. I chose 24 because few leagues even go near 24 quarterbacks in a single week, and because guys who made the top 24 had to play in more than just a handful of games.

With that, I calculated some stuff. Stick with me. I'm an English Lit major, so I'm sure you can handle this math. If not, it's very easy to learn about the findings either way.

I found the standard deviation of each player's weekly fantasy output. This number tells us basically how likely a player is to stray from his average in a given week. The standard deviation is the crux of the research because I use to to find a floor, ceiling, and coefficient of variance for each player in each scoring system.

I'll just give an example so we're on the same page. Drew Brees finished fifth in fantasy points among quarterbacks this year and averaged 19.59 points per game. The standard deviation of his game-by-game production was 5.83 points. Subtracting and adding this to his average gives us a range from 13.76 to 25.41. This means that 68% of his games will fall in this range. That's just how standard deviations work. If we subtracted and added two standard deviations from the average, then 95% of the outcomes would land in that range. (I'm not convinced that math isn't a work of Melkor, but that's for a different study.)

Clear enough? I hope. It gets more fun now.

Quarterbacks

Quarterback is an overvalued position in fantasy football -- but to say it's not important is lying.

I'm actually quite lucky with how the data ended up. There's a great example of how this information can be of use that just glared me in the face when I finished.

Sitting here now -- sometime after the 2014 season has ended -- which quarterback would you have preferred to roll out every single week this year: Ben Roethlisberger, Matt Ryan, or Tom Brady?

Did you say Roethlisberger? I would have, too, honestly. Dude had 12 touchdowns in two games! That stuff sticks with us. I would have taken Ryan last because he didn't really stand out to me.

But after seeing the data week by week, I would have wanted Ryan. Why? Consistency.

Player	Total	Average	St Dev	68% CI Low	68% CI High	Variance
Ben Roethlisberger	289.2	19.28	10.39	8.89	29.67	0.54
Matt Ryan	277.0	18.47	6.40	12.06	24.87	0.35
Tom Brady	274.2	18.28	8.30	9.98	26.58	0.45

So, all three guys, who ranked sixth through eighth in terms of weekly average, are within 1.00 point per game. That makes for nice comparing. But that average isn't equal, meaning it wasn't equally likely that they'd meet that average in a given game.

Sure, Ben and Brady had higher ceilings, but Ryan had the highest floor, which I like. You can just as easily disagree with me. The other two have the higher ceilings, and could "win a week," with which I don't disagree. But over the course of the season, I'd have preferred Ryan's more consistent point totals. Ryan was the third-most consistent (i.e. had the third-lowest coefficient of variance of the top-24 quarterbacks).

Ben was actually the 22nd-most consistent quarterback this year -- which really resulted from his huge games. Don't misunderstand me though. Ben had a high volatility because of his monstrous ceiling. But he also had two games with fewer than 11.0 fantasy points, which don't do much good for your squads.

Applications: To Stream or Not to Stream?

That is the question. Whether 'tis nobler in the mind to suffer the slings and arrows of outrageous misfortune (of getting Eli Manning'd) or to take (safer) Arms against a Sea of streamers, and by taking an early-round quarterback, end them?

"Okay, Polonius," you say, "Get to the point."

Well, you can stream quarterbacks. Why? Ceilings. And floors. But mostly ceilings. Of course, I mean the 68-percent-likely ceilings and floors.

Check out the "realistic" ceilings and floors for the top-24 quarterbacks this year, sorted by ceiling. Look how many had ceilings of at least 20.00 points compared to those with floors of 10.00 (in bold).

QB	68% CI Low	68% CI High	Â	QB	68% CI Low	68% CI High
Andrew Luck	14.08	30.86	Â	Ryan Tannehill	11.15	24.46
Aaron Rodgers	13.38	30.58	Â	Ryan Fitzpatrick	5.81	24.16
Russell Wilson	12.22	30.41	Â	Jay Cutler	11.13	23.84
B. Roethlisberger	8.89	29.67	Â	Joe Flacco	8.70	23.54
Peyton Manning	12.46	28.12	Â	Matt Stafford	7.47	23.17
Tom Brady	9.98	26.58	Â	C. Kaepernick	8.32	22.08
Cam Newton	9.02	26.25	Â	Kyle Orton	8.60	21.83
Philip Rivers	8.98	25.64	Â	Andy Dalton	6.58	21.47
Drew Brees	13.76	25.41	Â	T. Bridgewater	5.89	19.88
Matt Ryan	12.06	24.87	Â	Alex Smith	9.79	19.53
Eli Manning	8.52	24.86	Â	Derek Carr	6.00	18.63
Tony Romo	11.06	24.43	Â	Blake Bortles	7.41	16.81

20 of those 24 quarterbacks have a ceiling greater than 20.00 fantasy points (greater than 21.46, actually). In the right matchup, those guys can come close to that number, which is why there is a podcast dedicated to that kind of thing.

Only nine guys had a floor in the double-digits. So, that means that guy who drafted Peyton Manning in the first round this year and said, "Hey, he's a lock for 20 points a game!," was wrong. But if he happened to say, "Rationally, over the course of the season, he's going to average 20.29 points and also have one of the best chances to put up a 30-point game of any quarterback," well, then he was right.

But what about later choices in the draft? Or during the season? Would playing Matthew Stafford over Alex Smith all season be more advantageous by season's end? Yes. Stafford outscored Smith per game (15.32 to 14.66), but with such a minute difference per game and considering his lower floor (and much higher draft cost), you don't need to be stuck to Stafford week after week when guys such as Smith, who have similar realistic ceilings, are in good matchups.

So, 20 quarterbacks can hit 20 points without too much of a double take. Since only four quarterbacks averaged at least 20 points per game, you can make up that gap with a properly streamed quarterback who performs at a higher level (a standard deviation above his average) -- especially if you are playing an opponent who is clung to a player like Stafford.

If you're playing against Peyton, Peyton drops 31 (his season high), and you're using Eli, you likely aren't going to compete with that. But Peyton didn't score 31 each week, and Eli scored at least 20 in six games. The right call at quarterback can, according to the numbers, help limit the edge that the quarterbacks who have the highest ceilings offer.

Daily Fantasy Applications

While I opt for streaming in traditional leagues, I'm less inclined to save salary cap when playing cash games in daily fantasy football. Why? Well, the floor.

The players with the highest floors (in the table above) are, generally, the most expensive quarterbacks for a reason. They're consistent, and they also offer a high ceiling, too. But in cash games -- head-to-head games or 50/50s -- trying to secure that floor is helpful in racking up a safe score.

Why do I like safety in the quarterback in daily head-to-head but not traditional? The cost isn't the same at all. Allotting more of my salary to the position to secure Aaron Rodgers isn't the same as passing on a running back in the second round of a draft to get him. In, say, Week 8, I know which cheap running back will have a workload for a single game. I don't know that in August, and I know that I can typically rely on a streaming quarterback to narrow the gap.

But the value of running backs is something we'll discuss in the next iteration of this series (it's where I think things get most interesting).

How can the higher ceilings apply to daily fantasy? Well, those are nice for tournaments. For example, Eli's ceiling (24.86) wasn't far off from Brees' (25.41). The discount on a player such as Manning -- or in more extreme examples, of course -- who offer similar ceilings can be crucial in creating a high-scoring lineup.

What It Means for the Position

For a series on consistency, I realize that this was less about consistency and more about realistic floors at ceilings -- not exactly the same topic, but close.

That's because the numbers suggest that treating quarterback as a season-long commitment isn't worthwhile. There are too few guys worth holding onto for an entire year because considering waiver wire options have nearly the same realistic outputs, allowing you to play quarterbacks in friendlier matchups. That is to say that you don't need to worry much about week-to-week consistency because, if you're in most normal-sized (10-, 12-, or 14-team leagues), you can find viable options each week.

Also, the lack of high floors from even the elite guys -- only Andrew Luck had a floor of 14.00 points or higher -- means that you really aren't getting guaranteed points each and every week if you go with an early quarterback, though I'm not trying to say that having Aaron Rodgers on your team is a bad thing.

The other three positions will be more telling in terms of actual, weekly consistency, but taking a guy like Peyton or Brees for his "guaranteed production" isn't necessarily true or worth it.

That's especially true if you miss out on running backs, which is the focus of the next article in the series.