Let’s Stream Defenses: Week 11 (Fantasy Football)
This week, I thought I’d talk about projecting points allowed. I use Vegas lines for the points allowed component of my projections, because it’s pretty much impossible to do better. With millions of dollars riding on the Vegas lines, any possible improvements to them would already be in place. That doesn’t get us all the way to projected fantasy points, though. Here’s how fantasy points due to points allowed are scored here on FantasyPros.
|Min PA||Max PA||FPTS|
The difficulty comes from the fact that this is weirdly disjointed. Fantasy points don’t go down smoothly as points allowed go up – rather, fantasy points take big jumps at a handful of thresholds. The Broncos are projected to score 19.75 points this week. We could just apply the scoring and say that that would be worth 1 fantasy point, but that implies an unrealistic view of the accuracy of our projections. The projection of 19.75 means the Broncos have about a 50% chance of scoring less than that number, and a 50% chance of scoring more than that number. So really, our expected fantasy points should probably be somewhere between 0 and 1.
Another problem is that even though the projected points allowed is in the center of a range of outcomes, it doesn’t tell us how spread out that range is. It could be the case that they are guaranteed to score between 17 and 22 points, and it could be the case that less and 10 or greater than 40 are are realistic outcomes. This is where we have to make an assumption: that NFL scoring is described by a Poisson Distribution. Describing NFL scoring with a probability distribution gives us the ability to calculate the probability that the Broncos’ score will fall into each range of values (using the cumulative distribution function). Because football scoring happens mostly in chunks of 7 and 3, this assumption isn’t perfect. The Poisson Distribution is only a perfect descriptor of situations where events happen one at a time, and have roughly the same probability of succeeding with each attempt. Completed passes by a quarterback are a great example of this. Still, this gives us much better results than just plugging into our scoring rules. Using this method, here’s the probability that the Broncos will score in each range:
|Min PA||Max PA||FPTS||Odds|
Now that we’ve calculated the probability of each scoring range, we get our projection for fantasy points due to points allowed by multiplying the probability of each range by the value of that range and then adding together the results. In this case, our projection of 19.75 points scored by the Broncos translates into 0.75 fantasy points for the Chargers. That means the other 6.81 points of the Chargers’ projection come from sacks, turnovers and defensive touchdowns.
Show Me The Ranks
There are six teams on bye this week. This creates some challenges for streaming defenses not because the teams on bye are good, but because 5 of the 6 are offenses we like to target (Buffalo, Cleveland, Miami, San Francisco, and the Jets).
|The Start Them With Confidence Tier|
|The Still a Fine Choice Tier|
|The Maybe You Can’t Find Something Better Tier|
|The Surely You Can Find Something Better Tier|
*Baltimore being favored by 4 over Cincinnati is a real line, but there is currently no over-under, because of the uncertainty about whether or not Joe Flacco will play. I guessed the line at 47. It’s not clear to me that Robert Griffin or Lamar Jackson would be a downgrade from Flacco, so I don’t expect the line to be particularly different whatever happens.
**Technically, this is a home game for the Rams. Practically, the game is in Mexico City, so it’s not really a home game for either team, so I calculated projections as though they were both on the road. Neither D/ST was going to be in the conversation for fantasy anyway, so it doesn’t make a huge difference.
Tell Me About The Starters
- ARI vs OAK: Despite their offensive woes, Arizona’s defense has been quietly effective. They’ve been good enough to take advantage of good matchups (20 and 9 fantasy points in their two meetings with SF, 12 points against MIN), and they tied Jacksonville for the best fantasy performance against Kansas City with 5 points, including 5 sacks. We’ve said all there is to say about how bad Oakland’s offense is – there’s no defense I wouldn’t start against them.
- LAC vs DEN: Case Keenum finally snapped his 11-game interception streak before the Broncos’ Week 10 bye, but that was in just a 17-point effort against Houston. To be fair, Houston is one of the best real-football defenses in the league, but Denver was similarly bad against Kansas City, and the Chargers defense is much better than KC.
- PIT @ JAC: Jacksonville has been very nice to defenses all year. The return of Leonard Fournette should help the team, but that’s already built into this projection. Without him, Pittsburgh would be even more favored. Either way, they’re a great start this week.
- WAS vs HOU: This rank might surprise you. While both teams are ranked as starters, the underdog, Washington, gets the higher rank. This is because the difference in sacks outweighs the difference in points allowed. Over the past year, Houston has recorded an average of 2.2 sacks/game, and Washington has allowed that same average of 2.2 sacks. On the other side, Washington has recorded 2.9 sacks/game, while Houston has allowed 3.3. That results in projections that differ by about 0.8 sacks, and therefore 0.8 fantasy points due to sacks. Even though Houston is the favorite, they are only favored by 3 points, and when we translate that into projected fantasy points due to points allowed, the difference is only about 0.5.
- OAK @ ARI: With a total of just 40.5, this is a game where both defenses should be fantasy viable. Projecting Oakland is an interesting exercise because there’s a big discrepancy between what Oakland has done against other opponents, and what Arizona’s offense has allowed their opponents to do. Oakland is averaging 1.4 sacks/game over the last year, compared to 3 sacks/game allowed by the Cardinals. Oakland has recorded just shy of 1 turnover/game, while Arizona has allowed just shy of 2 turnovers/game, which is second-worst in the league behind the Giants.
- CHI vs MIN: Chicago has been the best real-world defense in the league this year. Minnesota is not a team I’d start just anyone against, but they’re certainly in range for the Bears.
- HOU @ WAS: This is another game where both teams are viable fantasy starters. Houston ranks 3rd in Defense-Adjusted Value Over Average (DVOA), and are certainly set up for success as road favorites.
- IND vs TEN: Tennessee was suddenly the team we’ve always wanted them to be last week against the Patriots, but one game doesn’t wipe away their track record of being sack- and turnover-prone, and struggling to score points. We need to see it a couple more times before we start fearing Tennessee as a matchup for fantasy defenses.
- ATL vs DAL: The Cowboys are right up there with Cleveland and Buffalo among the teams allowing the most sacks, and this is one of the lowest over-unders the Falcons have seen all season. Both of these things point to a good opportunity for the Falcons to put up good fantasy numbers.
- CAR @ DET: The Panthers were thoroughly trounced by the Steelers last week, but that game was the outlier. They’ve actually been pretty good this year, and are in a good spot against a Lions team that’s likely to be missing Marvin Jones.
- NO vs PHI: New Orleans is very heavily favored at home against an Eagles team that has kind of fallen apart since benching Super Bowl MVP Nick Foles for Carson Wentz. (That’s not actually the reason, but they have been bad.) The Saints aren’t a top-tier defense, but they’ll have some turnover upside in the likely event that the Eagles are playing from way behind.
How Was Last Week?
The top-ranked Jets were a huge dud, allowing the Bills to almost triple their point projection with a practice squad quarterback. Fortunately, the rest of the top tier did very well, with only the Eagles failing to hit 10 fantasy points, for a Tier average of 8.3. The second tier did not fare so well, with only 2 points/team. This was actually worse than the non-starter Tier, which averaged 4.7 points thanks to stellar performances from Buffalo, Washingto, and Pittsburgh, and relatively few teams that got really crushed.
|The Start Them With Confidence Tier||8.3|
|The Still a Fine Choice Tier||2|
|The Surely You Can Find Something Better Tier||4.7|
Jacob Herlin is a Senior Data Analyst for FantasyPros. For more from Jacob, follow him @jacoblawherlin.