In this week's Varsity Numbers, Bill Connelly revisits some measures and concepts: Adjusted Scores, Covariance, and momentum (or whatever you choose to call it).
30 Aug 2011
by Danny Tuccitto
This feature has taken quite a journey over the years. It first appeared back in the Pro Football Prospectus days as a book essay entitled, "Fantasy Risers and Fallers," in which we highlighted players whose fantasy points were either primed to explode or primed to fall off a cliff. Later on, it left the printed page, and became a piece for the website under the same premise. After a couple of years, the focus shifted toward players who KUBIAK thought were going to be much better or much worse than the conventional wisdom indicated as per average draft position (ADP). That's not really a risers-and-fallers concept anymore, so a name change was in order.
One final change we've made to complete the metamorphosis is to consider ADPs within positions rather than overall. For instance, say we told you that KUBIAK ranks Giants wide receiver Hakeem Nicks as the 34th-best player, whereas ADP says he's being taken, on average, with the 18th pick. Wow, a 16-pick difference! People must be overvaluing Nicks! No, not really. It turns out KUBIAK has Nicks as the eighth-best wide receiver, and ADP has him as the sixth-best, so that 16-pick difference is just an artifact of people taking wide receivers much earlier overall than KUBIAK says they should.
The rest of the piece will involve us identifying several players who are undervalued according to positional ADP ranking, as well as several others who are overvalued. In other words, players who KUBIAK feels are being taken either too late (i.e., undervalued) or too early (overvalued) at their positions. For each player, we'll give the raw rankings, and then do our best to explain why KUBIAK and ADP diverge.
If you haven't purchased KUBIAK yet, you can do so here. It's more than just a list of names on a page. It's a customizable Excel application that's constantly updated throughout the preseason to reflect the latest injuries, transactions, and changes to the depth chart. Once you buy it, you can download it an unlimited number of times.
QB Rank per KUBIAK: 14
QB Rank per ADP: 22
If you believe positional ADP, there's a decent chance McCoy might not even get drafted in your league. Being a Cleveland Browns quarterback might have something to do with that. Luckily, KUBIAK has never heard of Tim Couch or Spergon Wynn. On the other hand, KUBIAK also doesn't know about McCoy having a winner's moxie. No, he's ranked as the 14th-best quarterback because quarterbacks usually improve in their second season, and we project Cleveland to have the second-easiest schedule in the league. Last year, McCoy's stats pro-rated to 209 fantasy points over 16 games, and that was against one of the tougher schedules in the league. If he duplicated that, he'd be 26th in KUBIAK among quarterbacks this season. Adjust it upwards for the factors we just mentioned, and he moves up to a KUBIAK projection of 250, good for 14th.
QB Rank per KUBIAK : 5
QB Rank per ADP : 10
In fantasy circles, Big Ben suffers from what we'll call Steelers Football Syndrome (SFS). People tend to associate their offense with smashmouth football, when in reality they've been more of a passing team for quite some time now. That's why, outside of his rape-allegation-adjusted ADP last season, Roethlisberger perennially ends up outperforming his draft standing. This year appears to be no different.
Just like McCoy, Roethlisberger is likely to benefit from playing in the AFC North, which reaps the bonanza of a non-conference schedule against the NFC West: Pittsburgh has the third-easiest projected schedule this season. Given that his pro-rated stats would have ranked him seventh among fantasy quarterbacks last season, a KUBIAK move up to fifth this season seems more likely than an ADP drop to 10th.
RB Rank per KUBIAK: 13
RB Rank per ADP: 27
Here at Football Outsiders, we spend most of our time talking about play-by-play metrics. In fantasy football, though, performance per play is less important than the sheer number of opportunities a player gets to perform. Barring some touchdown percentage anomaly, a running back who has a 3.5-yard average over 300 carries will score more rushing points (105) than one who has a 4.5-yard average over 200 carries (100) even though the former was far less efficient. In fantasy football, opportunity is king.
That's the main reason why KUBIAK projects Benson to be much better than where people are drafting him. With rookie Andy Dalton at quarterback, Cincinnati backs will be getting the ball early and often. And with no competition for his job, Benson will receive the vast majority of those carries barring injury.
We should also note that KUBIAK incorporates regression to the mean for Benson in terms of yards per carry. It's unlikely he'll repeat last year's dismal 3.5 average. So, if you spread 3.8 yards over 300-or-so carries, you end up with a pretty good RB2, not the RB3 that ADP suggests.
(Note: If you do draft Benson, be aware he may miss a game in Week 8 due to his recent Texas plea bargain.)
RB Rank per KUBIAK: 22
RB Rank per ADP: 32
Addai has a few things going for him from a KUBIAK vs. ADP perspective. First, simply duplicating his pro-rated 2010 stats would make him worthy of being the 20th-ranked running back in this year's projections. Second, if Kerry Collins has to start a few games (or more), that can only help Addai, both for Benson-esque reasons as well as crap-I-don't-know-the-offense-yet-so-let-me-check-down-to-Addai-every-play reasons.
Third, fantasy football owners are human beings (See SFS above). When they get burned because a player is unreliable from week-to-week, he immediately goes on their Do Not Call list. Furthermore, when a player misses a ton of games due to injury, human fantasy owners just assume he's going to miss a ton of games the next season. In reality, injuries are, of course, unpredictable. It's the good ol' availability heuristic at play. Maybe he won't play all 16 games, but it's unlikely he'll miss half the season like he did last year.
WR Rank per KUBIAK: 36
WR Rank per ADP: 76
Roberts is another example of the importance of opportunity in fantasy football. Last season, with Steve Breaston nursing a sore knee, Roberts saw most of the action at No. 2 wide receiver in Week 16, and his best game of the season. Obviously, one game is not a reliable predictor of anything, so KUBIAK isn't projecting upwards of 250 fantasy points for him in 2011. However, with Breaston now in Kansas City, Roberts at least gets the opportunity to show what he's got over a full season.
Some might argue that Roberts is in a battle with Early Doucet for the receiver job opposite Larry Fitzgerald. Aside from arguing that Roberts is the better player, we'd also remind everyone that the Cardinals led the league in four-wide-receiver frequency last season. That doesn't appear to be changing this year given the addition of an actual NFL-caliber starting quarterback.
Also on the positive side of the ledger for Roberts is that Arizona has the fifth-easiest schedule in the NFL this season according to our DVOA projections.
WR Rank per KUBIAK: 10
WR Rank per ADP: 21
Don't get us wrong. KUBIAK doesn't foresee Lloyd duplicating his breakout 2010 season. It's not that clever. However, it's also not as sanguine about Lloyd's 2011 season as the general fantasy football public seems to be. Although it's true that Lloyd won't be in a high-flying pass offense, he's still the No. 1 wide receiver for the Broncos, and there's a wide margin between him and the rest of the receiving corps. He may not be targeted 152 times again, but he's still likely to be targeted plenty barring injury.
Furthermore, as we pointed out in Wisdom of Crowds last week, there's no evidence to suggest that a receiver who breaks out at Lloyd's age will necessarily be a one-hit wonder. You can probably get good value with Lloyd if you take advantage of the Chicken Littles in your league.
QB Rank per KUBIAK: 17
QB Rank per ADP: 12
If Lloyd is this year's victim of widespread panic, Stafford is the beneficiary of widespread ease. Ask just about anyone who the sleeper quarterback is this year, and they're likely to say it's Stafford; hence his status as a low-end QB1 according to ADP.
The reason KUBIAK sees him as more of a QB2 is that, for all his potential, Stafford's constant injury woes have prevented him from posting QB1-level fantasy stats thus far in his career. Therefore, we have him listed with yellow risk, meaning we've discounted his fantasy projection by five percent in order to account for the uncertainty of his situation. Not having a solid footing on what to expect from him, it's probably wiser to take a Stafford as your Plan B than to take him as your Plan A, lest you be forced to rely on a worse quarterback for Plan B if Stafford gets hurt again. Again, KUBIAK isn't saying Stafford will be bad, just that you'd get better value waiting as long as possible to take him.
QB Rank per KUBIAK: 18
QB Rank per ADP: 15
Kolb is in the same boat as Stafford. Many fantasy players and experts are assuming that he's going to turn into the fantasy reincarnation of Kurt Warner. Perhaps he will, but there's just not enough prior statistical evidence to back it up. So, again, like Stafford, he's got yellow risk, so you'd do better than banking your quarterback position on a player with such a wide range of potential outcomes.
RB Rank per KUBIAK: 43
RB Rank per ADP: 24
Apparently, the fantasy football intelligentsia thinks a healthy Grant in 2011 will return to a points ranking about one depth chart spot below the healthy Grant of 2009. According to KUBIAK, that's a long shot. That's because a lot has changed since 2009, most of which involves Grant's opportunity for touches. James Starks will be stealing early-down carries, John Kuhn will probably be stealing goal line carries, and even rookie Alex Green might end up stealing some third-down responsibilities now that Brandon Jackson has left town.
The projected stat line for KUBIAK's 24th-ranked running back, Beanie Wells, is about 1150 yards and 5 touchdowns, and that's with basically no competition for carries. If Grant is able to reach those benchmarks, it'll probably be due to injuries that KUBIAK doesn't (and can't) foresee.
RB Rank per KUBIAK: 33
RB Rank per ADP: 21
Surprisingly, the downward pressure on Williams' fantasy prospects this season isn't Jonathan Stewart, as it has been throughout his career. Instead, this year's problem that hurts Williams is the team's running quarterback. If the Panthers have any sense whatsoever -- and we're assuming that they do -- they'll take advantage of Cam Newton's running ability, especially around the goal line. Therefore, KUBIAK expects Williams' touchdowns to remain in the single digits; potentially in the low single digits.
Another reason why KUBIAK has him 12 spots lower than ADP in the running back rankings relates to our Establishment Clause. Namely: We're projecting the Panthers to repeat as doormats of the NFC South, so second-half carries are likely to be the fantasy football equivalent of Elvis sightings.
WR Rank per KUBIAK: 23
WR Rank per ADP: 8
Don't get us wrong, Jackson has undeniable talent. He's also the Blake Griffin of the NFL: seemingly every one of his touches ends up on that night's highlight reel. The problem for KUBIAK is that he doesn't touch the ball enough to warrant WR1 status. As we pointed out in Football Outsiders Almanac 2011, Jackson tends to have far fewer targets than other No. 1 wide receivers, he only catches about half of those targets, and his drops aren't necessarily due to the fact that he runs a lot of deep routes.
As an example, compare Jackson's numbers to KUBIAK's ninth-ranked wide receiver, Mike Wallace. Just like Jackson (96), Wallace (98) had about 100 targets last season. Just like Jackson (24 percent), Wallace (27 percent) was near the league lead in terms of the percentage of targets that travelled farther than 25 yards in the air. However, Wallace's overall catch rate was 61 percent, which was much better than Jackson's 49 percent.
Finally, from a regression-to-the-mean perspective, Jackson's 22.5 yards per catch last season is such an outlier that it would be shocking if he duplicated it this season. If you give him the same catch rate and prorate his targets to 16 games, but cut his receiving average to an awesome-but-potentially-repeatable 18.0, he ends up with about 100 fewer yards.
WR Rank per KUBIAK: 16
WR Rank per ADP: 5
We'll leave you with yet another player whose positional ADP ranking seems to be the result of people falling in love with his potential, and ignoring the fact that he's never been as good a fantasy point scorer as ADP projects. First, he suffers from the same target-siphoning that DeSean Jackson does. His first two seasons, Vincent Jackson had about 100 targets. Last season, his pro-rated targets come out to only 92 even if you set aside his zero-target first game back.
Second, we currently have Jackson set at a red risk level, which may or may not change between now and the beginning of the season. However, even if we improve it to yellow for KUBIAK purposes, that still only makes him the 10th-best player in the wide receiver rankings. So, as was the case with Stafford and Kolb, KUBIAK isn't saying Jackson will be bad; just that he probably won't be worthy of the pick he's likely to go at in your fantasy draft.
38 comments, Last at 02 Sep 2011, 5:00pm by Alternator