OK, this may sound like a WDIS, but it's really more about Norton's rankings of two LB's, J Anderson and Pace.
The Jets face TB, who is ranked 26th in giving up FF point to LBs. Car faces NE, the 5th MOST generous team in the NFL when it comes to FF points for LB's.
In 4 out of the last 5 weeks Pace has failed to get into double digits in my tackle heavy league. As evidenced against TB, Anderson did little, but TB usually doesnt give up many FF points to LB's. Anderson did have 16 points against the run heavy Jets 2 weeks ago.
Yet every week, Pace is in Norton's top 20. He's 18 against the stingy TB offense this week too. Then D Harris is at 7 against TB as well? With Bart Scott, where are all these tacles going to come from? Beason against the friendly LB NE offense is ranked at 13, only 5 spots in front of Pace? Seriously? Anderson is projected at 83 with a very favorable matchup. Willis at 4 against the pass happy AZ offense? With 9 tackles? The Cards are the 3rd stingiest team in the NFL when it comes to surrendering FF points to LB's.
The more I look at the rankings, the less I have much faith in them. They just don't make much sense to me, and frankly, I'm having a hard time trusting them. I get the impression that they are just thrown together without much thought or research. It didnt used to be that way.
It was recently suggested that another staffer also do rankings. I would like to see that. These rankings now look more like they are based on a player's reputation rather than tackle and points opportunites.
In general, FBG is the best site, and I'm being honest, but the weekly projections just aren't my cup of tea, from IDP to offensive players. Using an algorythim to calculate 0.3 sacks or 0.4 TD is not very useful. It also makes it safe... you can't be wrong.
Rant over.
I'm going to play a little point-counterpoint here, Rovers, mostly because I know you won't mind.

Calvin Pace has four or more solos in six of eight games this season. He's projected to four solos this week. It's clearly not a great overall tackle opportunity matchup, but Norton's only got the four Jet linebackers projected to 17 total solo tackles. I don't think that's unreasonable. The total group of linebackers (some 3-4 and some 4-3) are
averaging a little over 16 solos per game this season. It looks like both Harris (only projected to six solos) and Pace are getting a bump in the rankings/projections this week because of their big play expectation.
Pace has admittedly been inconsistent in the line scores -- five sacks in eight games with three coming in one game, but he's got 12 pressures in the last three weeks according to profootballfocus. So he's getting to the passer but not finishing in the pocket. Though the Bucs are middle of the pack in sacks allowed per game (2.08/gm), they rank fourth in pressure allowed (quarterback hits from NFL.com per pass drop). It's definitely arguable that Pace's full sack expectation in the projections is high. Drop that to 0.5 expectation and Pace drops all the way down to a tie for #42 in the rankings.
I agree that Anderson's matchup is strong. I'd probably project him to five or six solos myself; I'm not sure what John's thoughts behind the 4-2 projection may be there. It'd be a toss-up to me on who to start -- Anderson's relatively safe with questionable upside, Pace is a boom-bust play.
I've no problem with Willis' projection. The Cardinals may project poorly to number of points given to the linebacker position as a whole, but Willis is likely to get 50% of that projection (or better).
There is a rare stinker in this group of inside backers vs Arizona, but Laurinaitis sixteen solos, Hawthorne ten, Henderson, Durant and Beason had eight, Tatupu and Pierce seven. OLB Briggs had nine. In W1, Willis himself had eleven. I think nine solos is well within range this weekend.
With regard to the methods, I'd probably use an algorithm of sorts rather than a gut feel projection as well. I know that John uses seasonal solo and assist norms, runs a weekly spreadsheet with tackles and big plays per defensive position allowed by each offense, double checks the Matchup Analyzer for close calls and injects a little bit of gut feel based on playing time changes, expectation of how the game might go, etc. Without the fractional big play projections, you'd essentially see seven or eight clumps of tiers based on who was projected for five solos vs six solos and who got projected to a sack or not.
Given the issues above, I'd much rather consider the seasonal averages, matchup analyzer data, etc to look deeper into the non-specific SOS by positional data. I believe the NYJ tackle stats are in line with expectation, as are the Willis numbers. Both, I think, correctly tease out the numbers within the larger numbers. The Anderson projection I'd differ with a bit, and you may be correct that John's natural bias toward conservatively projecting players who have yet to prove themselves over a longer time frame might be a hole in the data.
Two things might be helpful with the projections.
Ranges would be helpful. A guy like Pace has an expectation between 1-1-0 and 6-3-2 this week, with those lines closer to his ceiling not necessarily more likely than those near his floor. A guy like Anderson has a much tighter expectation. His expectation is probably between 4-1-0 and 7-1-0 with the 5- or 6- line carrying a very high percentage. To do that for every player would keep a single set of projections from ever getting finished.
Fractionating the tackle projections might be helpful in better delineating tiers, but doesn't do much to help the range issue. In theory, it should make things more precise, but it just puts a more specific number within the larger range.