Streaming defenses in fantasy without any football knowledge

(“Today’s Crazy Fantasy Stat” is an occasional offseason offering from PFF that highlights something that catches our eye and aids in our preparation for the 2017 fantasy season.)

Every year, all the fantasy smartypantses out there recommend waiting on taking a defense in your draft. Don’t jump early. Trust your research and matchup ability in the season to stream defenses, adding and dropping based on all the research and knowledge you can accumulate during a given week.

That’s hard, though.

The truth is, not everybody has the time and wherewithal to spend all the time studying and researching every last bit of defense and special teams fine details to figure out exactly the right streaming candidate in a given week. And for those who don’t want fantasy to overtake their entire life, I can understand the temptation to simply draft what should be a trustworthy defense in, say, round 10 and be done with things until a bye week.

So this week, I want to find a way to convince even those people who can’t spend hours on defense research that no, you should still wait on defense.

(Caveat for everything that follows: Defense/special teams scoring is wildly inconsistent across the fantasy landscape. I’m using the PFF scoring system for all references to D/ST fantasy points and rankings. Your specific hosting site might vary, but the variances shouldn’t be significant or that meaningful.)

The penalty

First off, you always miss out on lottery tickets when you pounce on defenses early. Per Fantasy Football Calculator, the Broncos defense was first off the board in 2016, late in the eighth round. Not long after that unit were Jay Ajayi and Derek Carr. About a round later came the Arizona and Seattle defenses, just before Kirk Cousins, Bilal Powell, and Tevin Coleman. Houston’s defense, fifth off the board, went two spots before Michael Thomas, four before Jimmy Graham. Even a trustworthy defense isn’t worth those guys.

Join PFF Elite for access to PFF Signature Stats across all positions.

Now, of course, you know that. And Ajayi, Cousins, and Thomas could just as easily have been Charles Sims, Travis Benjamin, and Kamar Aiken. The point isn’t that you are passing up guarantees. The point is that you are passing up chances. If you said the drafting public was 100 percent accurate in its assessment of the defenses, that might be worth it. But it isn’t. The chart below shows the top 11 defenses off the board in FFC 2016 ADP, along with their end-of-season ranking in fantasy points:

Overall, the drafting public is pretty good at figuring out the defenses. But only pretty good. And “pretty good” isn’t enough to justify passing on the potentially elite options in that range of the draft.

Now, how to choose a defense when you want to turn your brain off? Well, let’s take this step by step:

Pick a bad opposing offense

The bottom seven offenses of 2016 (worst to best: Rams, Browns, Jets, Bears, Texans, 49ers, Giants) played a combined 54 games against defenses owned in under 40 percent of leagues (in other words, defenses that were generally available as streaming options). Breaking those up to weekly averages, and simply choosing a lowly owned defense going against an awful offense would have given you 141 fantasy points on the season, the same as the No. 9 2016 defense, Carolina. That Frankenstein unit put up at least an average of 5.0 fantasy points every week but Weeks 5 and 6.

Fantasy total: 141

Pick a defense going against the prior week’s worst offense

The problem with that first method is that it used end-of-season scoring totals to pick options in season. I’d argue the lowest-scoring offenses weren’t exactly a surprise, but I’ll grant the ex-post-facto nature of it makes it dodgy.

So let’s take the same list of defenses owned in under 40 percent of leagues and apply that to the season’s schedule as we went along. The strategy was simple: Take the 15 lightly owned defenses, and pick the one whose opponent scored the fewest points of the prior week. If a team scored 0 and faced an easy-to-add defense the next, you used that defense.

(For all the remaining strategies, since we’re working off of year-to-date statistics, and Week 1 has none of those, I’m using the leaguewide Week 1 average D/ST fantasy score of 5.8.)

That strategy was imperfect. It led to only 93 fantasy points on the season, a bottom-six fantasy total. There were some good stretches (Weeks 8-12, this strategy yielded 8.4 fantasy points per game; Weeks 14-17, 8.5), but overall, it wasn’t worth it.

Fantasy total: 93

Pick a defense against one of the prior week’s worst offenses

The obvious flaw in the above strategy was an absence of nuance. It didn’t allow for you to think or manipulate anything. It led to, among other issues, starting the Browns defense against the Patriots in Tom Brady’s Week 5 debut. Let’s just say that any strategy that says that was a good idea is flawed.

So let’s adjust it. Take that same 15-defense sample of easy-to-add units, and instead of picking the single worst unit, just make sure you select an opposing offense that struggled the week before (defining “struggled” as scoring 19 or fewer points).

This one did better. Averaging out the defensive points per week, and this strategy led to 113 fantasy points, good for 19th in the league. Still not great, but an improvement.

Fantasy total: 113

Pick a defense facing a generally struggling offense

A one-week sample of any offense has a great chance of being misleading (see that scoreless Week 4 Patriots game for an example). So instead, take the three-game average of every offense and choose defenses facing any team that averaged under 20 points per game over its prior three games. For Weeks 2 and 3, before teams have three-game samples, simple use the biggest sample possible.

There are a lot of such offenses. The Rams alone averaged under 20 points a game for the sample 12 of 15 possible times. The only teams that never had a qualifying sample across the season were the Falcons, Packers, Saints, and Raiders.

The weekly average using this method yielded 143 fantasy points for the season. That’s slightly better than the first suggestion and good enough to be 2016’s No. 9 defense, just above the Panthers.

Fantasy total: 143

Pick a defense facing the most struggling offenses

Like I said, the above method gave a lot of offenses. Over 16 weeks (every week but Week 1), the choices were so voluminous that you’re not really that much better off than throwing a dart, even if that dart throw yielded a top-10 defensive performance.

So narrow it down. Instead of selecting from any offense averaging under 20 points a game over the prior three weeks, select from the three worst offenses over the last three games. Some weeks that’ll be offenses averaging 17 or 18 points; some weeks that’ll be offenses averaging 8 or 9. The point is to just look for the most struggling offenses entering a given week, and to limit our sample.

This one does even better, getting us to 160 fantasy points. Every single week offered at least 5.7 fantasy points from the defense/special teams, and for the season would have been the No. 4 fantasy defense, behind only Minnesota, Kansas City, and Denver.

Fantasy total: 160

So there it is. Using nothing more complicated than your cell phone calculator, you could theoretically have not drafted a defense at all and ended up with a Frankensteined unit that was better than all but three D/ST units in 2016. That’s before using any real brain power, and no scouting at all. That’s without even considering anything on special teams, anything to do with turnovers, anything to do with injuries. I literally didn't mention a single defensive player's name anywhere above this. It’s literally “who is available?” and “which offenses are bad?” You’d have to assume that, using any intelligence at all to the selection process, you could only improve that standing. Turn your brain on, and you could have the No. 1 defense in the league.

But no brain at all, and you could still have the No. 4.