2009 Statistical Rankings: 49er Players Through Week 2

On Wednesday, we here at Niners Nation (NN) introduced a new weekly post presenting Football Outsiders' (FO) team rankings according to defense-adjusted value over average (DVOA). Hope you enjoyed it, and hope you continue to read it for the next 15 (or more?) weeks. It turns out that FO also publishes player DVOA stats, as well as advanced stats measuring the performance of each team's OL, DEF front seven, and secondary. These stats are what I'll be presenting today, and will continue to present each Thursday or Friday during the regular season (and playoffs?).

If you don't know what DVOA is or how to interpret it, consult Wednesday's post. It's got my English-language answers for these relatively complicated questions. For today's post, and beware it's a long one, I'll be explaining FO's player stats as I present them.

After the jump, I'll discuss the player stats for the 49ers' QBs, RBs, WRs, TEs, OL, DL, and DBs...

Before I begin, I just want to alert everyone to something that I'm sure would come up in the comments were I not to address it in the article. Basically, I'm only going to present stats for individual 49ers that have enough passes (15), carries (16), or targets (6 for WR, 3 for TE) to qualify for FO's most valid individual statistics, DVOA and DYAR. Below the relevant "yardage opportunity" thresholds, these statistics can be wildly erratic from week to week, and - frankly - aren't as trustworthy or enlightening as I'd like them to be. Let me make clear though that this would be my opinion about any statistic, as all statistics are negatively impacted by extremely small sample size. Furthermore, I'd still take FO's small-sample stats over the popular NFL stats because, although both may be flawed when sample size is low, at least FO's stats are adjusted for opponent (after Week 3), game situation, and other important factors.

With that said, it's on to the stats and rankings...

QUARTERBACKS

Here are Shaun Hill's stats through 2 games (bold = top 8 in the NFL; italics = bottom 8):

DYAR	Rank	Passes	Actual Pass Yards	Yds Per Pass	Rank
-20	31	65	296	4.55	32
DVOA	Rank		EYds	EYds Per Pass	Rank
-16.5%	30		278	4.28	31

Before I discuss this table, let me explain what Defense-adjusted Yards Above Replacement (DYAR) and Effective Yards (EYds) mean. I've done this before on NN, but I'll do it again for the sake of clarification and NN newbie introduction.

For DYAR, FO first determined - via statistical analysis of NFL historical data - a "replacement level" of performance for each position that represents the number of yards an average NFL "backup" gains at that position. DYAR compares a player's yardage total to this "replacement level," and, as such, represents the difference between a player's yardage total and the yardage total that would be expected if he was replaced by an average backup at his position. In other words, DYAR answers the question that's often thrown around when it's time to argue about the MVP, "What would happen to Team X's offense if Player A got hurt and was replaced by Player B, who's an average backup?" In this sense, DYAR is a measure of a player's value to his team's offense, with positive numbers meaning his offense would be worse off yardage-wise without him. Now, obviously some backups are not average at all (See Sproles, Darren), so I'm using the MVP analogy in the name of education, not exact methodological precision. The purpose of DYAR is to compare players at the same position in a statistically valid way, not to settle the MVP debate. Therefore, it's important to realize that FO's use of a replacement level - something that's also done in sabermetrics - is aimed at creating a trustworthy statistic (by standardizing, and providing a baseline for starter performance) rather than providing fodder for sports talk radio.

For EYds, FO simply adjusts each players yardage totals based on the same factors for which it adjusts their DVOA statistic (i.e., opponent, game situation, etc.). In other words, EYds is basically DVOA without the whole "play success" determination that I described on Wednesday. This is intuitive because both are grounded in the concept that some yards (e.g., rushing yards on 3^rd and 20 against DET) are easier to gain than others (e.g., rushing yards on 3^rd and 2 against PIT), and therefore that high-difficulty yards should be worth more (i.e., are more efficient) than low-difficulty yards. Because this has the effect of standardizing (aka evening out) game situations for all players at the same position, EYds tells you how many yards a player would be expected to have if he played in average game situations. When a player's EYds is higher than his actual yards, it means that (a) he's been playing in game situations that are more difficult than average, and (b) he's actually been more efficient than his actual yards would seem to suggest. In contrast, the opposite is true when EYds is lower than actual yards.

Just to make EYds as clear as possible, here's a quick example:

Thomas Jones runs for 10 yards on 3^rd and 1 against TEN this week, while Willis McGahee runs for 10 yards on 3^rd and 10 against CLE this week. Both Jones and McGahee ended up with 10 rushing yards according to the NFL. However, we all know that Jones' 10 yards were more valuable than McGahee's because TEN is the top-ranked run DEF, whereas CLE is the 31^st-ranked run DEF. Now, let's say that, on 3^rd and 1, TEN's run DEF is 25% better than average (i.e., each rushing yard on 3^rd and 1 against TEN's DEF is worth 125%), whereas CLE's run DEF is 25% worse than average (i.e., each rushing yard on 3^rd and 1 against CLE's DEF is worth 75%).

Given these defensive situational stats, Jones' 10-yard run would have been expected to go 12.5 yards against an average run DEF (10 x 12.5 = 12.5), while McGahee's run would have been expected to go 7.5 yards (10 x .75 = 7.5). In other words, Jones' EYds for his run was 12.5, whereas McGahee's EYds for his run was 7.5. Therefore, the fact that Jones' EYds (12.5) is higher than his actual yards (10) accurately shows that (a) Jones' run was in a more difficult game situation, and (b) Jones was more efficient than McGahee. Notice also how going only by actual yards, which is what the NFL provides and pundits always refer to, our conclusion about which RB is better would have incorrectly been, "they're equals." This is precisely the major benefit of relying on EYds instead of actual yards.

OK, enough with the statistics lessons...

By any valid measure, Hill's value and effectiveness through 2 games compares pretty unfavorably to the other starting NFL QBs. With respect to his DYAR, the 49ers' pass OFF would actually be 20 yards better through 2 games if he had been replaced with an average NFL backup. Unfortunately, the Niners don't appear to have an average NFL backup on their roster, at least that's what I think. The kicker is that, even if you consider Alex Smith to be an average backup, Hill's DYAR suggests an improvement of only 20 yards were Smith to be elevated to the starting lineup. And one more thing about Hill's DYAR just to drive the point home: The only starting QBs that are less valuable (i.e., have a lower DYAR) than Hill so far this year are Brady Quinn (aka "1 TD guy"), Jake Delhomme (aka "6 turnovers guy), and Matt Stafford (aka "8 days on the job" guy). That's quite impressive company, indeed.

With respect to EYds, Hill's are lower than his actual passing yards, which means that his meager production has unfortunately come in situations easier than average. Furthermore, if we compare Hill to a starter who has similar actual passing yards through 2 games, we find that he's actually been less effective than - gasp - Marc Bulger, who has 66 more EYds than actual passing yards (355 vs. 289).

Of course, all of the above stats are affected by the fact that Jimmy Raye is not Sean Payton and Isaac Bruce is not Marques Colston. In other words, SF's OFF doesn't require much in the way of passing yardage, and SF's WRs aren't exactly Jerry Rice and Terrell Owens when it comes to yards after catch.

The latter factor is something we can't really account for, but the former is accounted for by the per-pass stats in the table. As you can see, Hill has thrown for 4.55 yards per pass, which is pretty abysmal when you consider that the average is 6.25 among the 35 QBs with 15 passes or more. What's worse though is Hill's 4.28 EYds per pass, which is over 30% lower than the qualified-QB average. Not to mention that the #1 QB, Drew Brees, has an EYds/Pass that's triple that of Hill's (12.83).

RUNNING BACKS

Among Niner RBs, only Frank Gore qualifies for the most valid FO stats through 2 games (> 14 runs). Here they are:

DYAR	Rank	Runs	Actual Run Yards	Yds Per Run	Rank
24	13	38	237	6.24	5
DVOA	Rank		EYds	EYds Per Run	Rank
7.6%	15		169	4.45	17

Looking at DYAR, Gore has been more valuable than all but 12 RBs, the most valuable of which has been Marion Barber (DYAR = 75). If we translate Gore's DYAR into DYAR per run, we see that replacing him with an average backup would cost the Niners' run OFF over half of a yard per carry. Of course, the question is whether or not Glen Coffee is an average backup. Based on the preseason, you'd think he's above average. However, because (a) preseason stats don't mean anything (See Clayton, Thomas; sorry Ninjames), and (b) he's a rookie, you can't be so sure of that above-average status.

Regarding EYds, Gore's are considerably lower than his actual run yards, which means that his production has come in easier-than-average situations. I have a sneaking suspicion that this is due to his pair of 80-yard TDs against the Seahawks, which both came in a relatively easy running situation (1^st & 10 with the Niners up 3 points). I can't know this for certain though because I'm not privy to the situation-specific NFL averages that FO uses to determine whether the difficulty of a particular game situation. However, it's interesting to note that the RB immediately ahead of Gore in actual run yards, Chris Johnson with 254, also seems to have had his EYds (145) take a massive hit thanks to 2 long TD runs in Week 2, albeit in (intuitively) easier game situations. Specifically, Johnson's 57-yard TD run came on 3^rd & 19 with the score tied at 0, and his 91-yard TD run came on 3^rd & 10 with the score tied at 24. Running the ball on 3^rd & Long with the score tied against HOU's run DEF sure seems like about as easy a game situation as you can get to run the ball. Although this is most likely why his yards-to-EYds difference is even larger than Gore's, I still think both of their differences are due in major part to the length of their TD runs because, subtracting 25%, for example, from an 80-yard "easy" run (or 20 yards) does much more damage to a RB's EYds than subtracting 25% from a 20-yard "easy" run (or 5 yards); 15 more yards of damage to be exact.

WIDE RECEIVERS

Among Niner WRs - and this shouldn't surprise you - only Isaac Bruce qualifies for the most valid FO stats through 2 games (> 5 targets). Incidentally, you should know that, for WRs and TEs, FO analyzes all plays in which a player was the intended receiver of a pass, not just the passes he actually catches. For your convenience, I've included "Catch Rate" in the table, which is simply Receptions divided by Targets. OK, here are Bruce's stats:

DYAR	Rank	Targets	Actual Rec Yards	Yds Per Target	Rank
-2	55	16	109	6.81	46
DVOA	Rank	Catch Rate	EYds	EYds Per Target	Rank
-14.3%	55	50.0%	99	6.19	54

According to DYAR, Bruce could be replaced by an average backup and the 49ers wouldn't lose much in the way of receiving yards. Unlike the situation with Gore, the 49ers likely do have an average backup or two on their roster. The question becomes, "Does Bruce's experience, leadership, and rapport with Hill make him more valuable than the average WR backup with which you'd replace him; say Brandon Jones or - gasp - a signed Michael Crabtree?" At this point, I'd say the answer is probably, "Yes, Bruce is more valuable right now."

I think the take home message of this table is that the 49ers' pass OFF doesn't have a #1 WR. Alternatively, you might say that anyone who thinks the 49ers have a #1 WR is sorely mistaken - at least through 2 games. I say this because Bruce, the Niners' nominal #1 WR, actually ranks as a middle-of-the-pack #2 WR according to every stat in the table. Now, as was the case with Hill, he's being affected by Raye's offensive philosophy. As evidence for this Raye effect, I'll echo something Mike Sando reported recently: Only twice has Raye been the offensive coordinator (OC) of a 1,000-yard WR, and only twice has one of his WRs had 65 catches or more. Interestingly enough, Bruce is on pace for a 64-catch, 872-yard season. Coincidence? I think not.

TIGHT ENDS

Among Niner TEs, only Vernon Davis qualifies for the most valid FO stats through 2 games (> 2 targets). Here they are:

DYAR	Rank	Targets	Actual Rec Yards	Yds Per Target	Rank
0	31	13	72	5.54	34
DVOA	Rank	Catch Rate	EYds	EYds Per Target	Rank
-7.7%	31	53.8%	61	4.69	32

All hail your 49ers' 2006 1^st-round pick! The 31^st-most valuable TE in a 32-team league! Seriously, though, there are two OFF positions that were poised to benefit when Raye replaced Mike Martz as OC: RB and TE. After all, Tony Gonzalez made his bones when Raye was in KC. Not to mention the fact that Martz had the perennial nobodies like Jeff Robinson, Roland Williams, Ernie Conwell, Brandon Manumaleuna, and Dan Campbell manning the TE position during his tenures in STL and DET. And yet, VD's 2-game "renaissance" consists of vaulting from 37^th in DYAR last year to 31^st this year.

To be fair, ARI and SEA have two of the better sets of LBs and Ss in the NFL, so it might be the case that VD's stats will improve when FO incorporates opponent adjustments after Week 3. Maybe, maybe not; we'll just have to wait and see. I wouldn't hold my breath though. The bottom line is that the 49ers have a seemingly above-average backup in Delanie Walker, so they might actually benefit from starting him over VD. Whatever the case may be, it's getting awfully close to "fish or cut bait" time for the guy in the fuchsia shirt.

Here he is soliciting your vote for the poll:

via blogs.orlandosentinel.com

OFFENSIVE LINE

For the sake of brevity, I'm not going to explain FO's methodology behind their stats for OLs and DEF front sevens. You can find their thorough explanation here. Rather, I'll just tell you that, in the next table, the following stats, are interpreted in the following ways:

• Adjusted Line Yards (ALY) - Run yards per RB carry (up to 10 per carry) that are attributable to the OL. This is a measure of OL run-blocking performance, with higher numbers meaning better OL play.
• Power Success Rate - % of runs that are successful on 3^rd/4^th Down & 2 or Less (i.e., ones that result in a 1^st down or TD). This is another measure of OL run-blocking performance, with higher %s meaning better OL play.
• Stuffed Rate - % of RB carries that gain 0 or negative yardage. This is another measure of OL run-blocking performance, with lower %s meaning better OL play.
• Adjusted Sack Rate (ASR) - Sacks per pass attempt. This is a measure of OL pass-blocking performance, with lower %s meaning better OL play.

Each of these stats has been adjusted for game situation, with all stats except ASR also having been adjusted for OFF formation because some formations produce more OL-reliant rushing yards than others. As with other FO stats I've presented over the past 2 days, opponent adjustments won't kick in until after Week 3. OK, here are the 49ers' OL stats through 2 games:

ALY	Rank	Power Success Rate	Rank	ASR	Rank
2.51	31	50.0%	25	10.0%	29
Actual Yds Per RB Carry	Rank	Stuffed Rate	Rank	Actual Sack Rate	Rank
5.20	5	24.0%	23	12.1%	30

Let's start with pass-blocking. Yeah, not so good. According to their actual sack rate, Hill is getting sacked about once every 8 times he drops back to pass; good (bad?) for 30^th in the NFL. According to ASR, the picture's a little rosier, but still not pretty: game situations have accounted for between 1 and 2 total sacks thus far this season, meaning that the OL is actually pass blocking better than what basic NFL stats would otherwise indicate (ask me in the comments if you want to know how I came up with this). Of course, surrendering 6 or 7 sacks in 60+ dropbacks isn't going to cut it in the long run (long pass?...sorry, had to).

In terms of run blocking, it sure looks like the Niners' OL isn't doing it very well. To begin, only half of SF's yards per RB carry can be attributable to the OL, a stat that's no doubt affected by Gore's twin 80-yard TD runs. In other words, outside of those 2 runs, in which the lion's share of the yardage was thanks to Gore's breakaway speed, the SF run OFF has been the proverbial 3 yards and a cloud of dust. More proof of the OL's lackluster run-blocking performance through 2 games is their relative inability to (a) pick up 1^st downs on short-yardage runs, and (b) prevent DEF penetration for stops at or behind the line of scrimmage. These conclusions are based on the OL's power success rate and stuffed rate.

Below is a chart displaying the OL's directional run-blocking performance (click to enlarge):

Looking at the top of the chart, you can see the 49ers are running the ball predominately up the middle and behind RT. Keeping in mind that the NFL standard for teams with a right-handed QB is to have their better run blockers on the right side of the line and their better pass blockers on the left side of the line, the Niners' preference for middle-right runs is in the top quartile of an already middle-right run-heavy league. It's also interesting to note - not to mention very peculiar - that SF hasn't the run the ball to the outside right at all in their first 2 games.

With respect to directional ALY, notice how inept the 49ers' OL has been on runs behind the tackles. This is especially troubling behind RT for the reason I just described: it's where their OFF prefers to run the ball. Hey. At least they're only below average when it comes to running up the middle. Of course, these stats probably confirm what we've been seeing in the run game from our 49ers. Namely, 3 yards and a cloud of dust; keep running it up the middle for minimal gains regardless of whether the DEF is stopping it or not - and they have been.

DEFENSIVE FRONT SEVEN

The same stats for OL can be applied in reverse to the DEF. However, rather than these stats measuring the performance of the DL, it's more accurate to say that they measure the performance of the DEF front 7 because LBs - and sometimes Ss - are involved in run-stopping. So here they are:

ALY	Rank	Power Success Rate	Rank	ASR	Rank
3.24	8	100.0%	32	5.8%	14
Actual Yds Per RB Carry	Rank	Stuffed Rate	Rank	Actual Sack Rate	Rank
3.13	6	32.3%	2	4.5%	13

Once again, we see statistical proof that the DEF is carrying this team. Except for being perfectly unsuccessful in short-yardage 3^rd/4^th down situations, they've been stout overall stopping the run. Although perhaps not apparent at first glance, the Niners' middle-of-the-road ASR is an important development. For one, they were a dismal 26^th last season, with an ASR of 5.0%. Also - and this is a nuanced point, so try to keep up with me here - the 2008 49ers' DEF had an ASR that was lower than its' actual sack rate, whereas this season the reverse is true. In other words, last year's pathetic sack total was - even more depressingly - a statistical mirage due to favorable game situations. Think "losing by 20 points on 3^rd & 20 in the early 4^th quarter." In contrast, this year's pass rush has actually been hurt, rather than helped, by the game situations in which they've played. Aside from this being yet another vindication of the validity of FO's stats, it more importantly suggests that the Niners pass rush might have more to show for itself once game situations even out a little bit as the season progresses.

Below is the DEF front 7's directional run-stopping chart (click to enlarge):

As you can see, the 49ers are a top-8 DEF through 2 games when it comes to stopping runs to the outside. Is this a speed thing? Is this a Nate Clements and Shawntae Spencer thing? Is this an overall aggressive philosophy thing? You decide.

One other thing I'll point out about the directional stats is a bit of good fortune the DEF front 7 seems to have experienced in their first 2 games. Namely, they allow over 7 adjusted yards per RB carry when the opponent runs behind their LT, but - thankfully - ARI and SEA have basically ignored that major weakness in their play-calling. If I were MIN, or any other future 2009 opponent, I'd be looking to exploit this weakness much more than 3.2% of the time.

DEFENSIVE BACKFIELD

Finally, FO puts out DVOA stats for each pass DEF that tell you how they've performed against their various types of opposing receivers. As is the case with all passing stats, these look at every pass attempt, not just those that are completed. Again, I've made a reader-friendly chart for you (click to enlarge):

Although it's quite difficult to figure out, a thorough search of the internet, NFL game books, and Fooch's DVR led to the pass DEF formations displayed above. In nickel, Manny Lawson and Parys Haralson move to DE, Dre Bly comes in as the 3^rd CB, Mark Roman - ugh - replaces Michael Lewis at SS, and Lewis moves down to replace Takeo Spikes at WLB. In dime, either Lewis moves back to S (with Dashon Goldson and Roman) or Tarell Brown replaces Lewis as the 6^th DB.

As for the stats, the chart shows that Spencer and Bly have actually played more efficiently than Clements. Of course, it's once again important to remember that the above stats haven't been adjusted for opponent yet. Obviously, Larry Fitzgerald and T.J. Houshmandzadeh are 2 of the best #1 WRs in the NFL, so Clements' stats stand to improve simply by virtue of the adjustments. I don't anticipate much of a change in the stats vs. TEs, because SF has faced one really good TE (John Carlson) and one invisible TE (Anthony Becht), which pretty much averages things out anyway. And one more thing I don't envision changing much is the pass DEF's poor performance against RBs, but that might just be - at least in part - by design.

BOTTOM LINE

OK, so through 2 games, we can draw the following conclusions about the 2009 49ers based on their team statistics thus far:

1. Hill has produced pass-unfriendly numbers despite playing in pass-friendly situations. But having Jimmy Raye as an OC is a superceding pass-unfriendly situation.
2. Gore's been a valuable and efficient RB, though not quite as much as what those two 80-yard runs might suggest. Sorry for the reality check, but don't expect him to be gaining 6+ yards per carry for the remainder of the season.
3. Bruce may be the Niners' best WR, but he sure hasn't been a #1 WR, which means Morgan hasn't been a #2 WR, and so on.
4. VD...can win without him.
5. The OL hasn't been run-blocking very well, which is surprising, what with all the practice they've been getting. They've been pretty bad compared to the rest of the league at pass-blocking too, but still better than what their 8 sacks allowed might suggest.
6. The DEF front seven has been stout against the run, especially when the opponent has run to the outside. Also, their pass rush through 2 games is statistically better than last year's, and may improve even more once their game situations even out.
7. The secondary is defending well against #2 WRs, #3 WRs, and TEs, but not so well against #1 WRs and RBs. Clements' numbers will probably improve once FO starts making opponent adjustments, but the poor pass DEF against RBs probably won't.

OK, that's it for this week. If the length of these 2 posts has rubbed you the wrong way or made you feel like you just survived being hit in the back by Patrick Willis, don't fret. Next week, I won't be trying to teach everyone how FO's stats were developed, what they mean, how they're interpreted, etc. That'll make the posts a whoooooooole lot shorter. So, see you again next week, hopefully after a Niner win!

**DVOA, DYAR, ALY, and ASR statistics used to produce this article were obtained from Football Outsiders.