Pro Football Reference Blog

I KNEW the Broncos wasn't outplayed by Indy last year. As soon as I saw what was going on, I went straight to see if that game was there. I feel vindicated.

I feel the need to point out that the first game Manning lost but played well enough to win, the one vs the Jets, is where he, and almost all the rest of the Colts starters, were pulled in the 3rd quarter to rest up for the playoffs. I have to wonder how QBs being pulled from games affects this sort of list, as there may be similar cases of QBs playing "well enough to win" but ending up with a loss because they were yanked early for their own safety (Or, on the other side, QBs barely playing at all in a game, and perhaps having that game not be counted).

Scott,

Are you thinking about Rivers having a 71% p(W) in a game where the Chargers never threatened to win?

Jerry, that would be a great example. As would be the playoff game against the Steelers from 08.

Anthony, really? "If SD's D got 1 stop, there is no doubt SD would of won that game..."?

We must have been watching different games. In fact I know we were since the score was 35-24, not 35-28.

Steelers weren't in much danger of losing in the second half according to Brian Burke's win probability graphs: http://wp.advancednflstats.com/nflarchive.php?year=2008&team=PIT&gameid=54462

The two games are a bit different. The playoff game was close at halftime, and then San Diego ran one (!) offensive play in the third quarter. Their win probability, according to Burke, never exceeded 10% in the fourth quarter. The regular season game, which is what I had in mind, involved the Steelers jumping out to an early lead and then letting the Chargers rack up a lot of yards. Again using Burke's win probabilities, San Diego was never above 50%, never above 11% after the first quarter, and never above 9% in the second half.

I should add that this isn't meant to denigrate Philip Rivers. It's just that I remember the 2009 regular season game well enough that I know a 71% estimate of win probability has no relation to what happened on the field.

I am a major colts fan and as such, i really despise the chargers, but even I cannot deny how heroic philip rivers has played this year. I still can't understand how he isn't mentioned in the category of the best qbs. Is it really a matter of playoff wins? go look at any of the great qbs, and you will find their history litered with poor performances but still coming out with the win. I simply will never buy into the fact that a qb should be judged on win/loss. Its the same bs excuse that manning detractors have used his entire career.

It's safe to say that when you lose a game despite those numbers, it's probably the fault of somebody other than the QB.

"Safe"? Maybe not.

---

bit.ly/9Or5R2

October 10, 2010

After two blocked punts resulted in a safety and touchdown, the Chargers began to move the ball, only to have Mike Tolbert fumble at the goal line and Philip Rivers lose the ball on a blindside sack on successive first-quarter drives.

[Then] the Raiders drove 97 yards and 73 yards for touchdowns on their final two possessions and then all but iced the game with a 64-yard return of a sack-fumble with 58 seconds remaining by Tyvon Branch after Michael Huff knocked the ball away from Rivers.

---

Cf. Sacked When Passing [ bit.ly/bdTDBk ]. Rivers' sack-rate-per-attempt doesn't necessarily jump out (though it's on the high end of those listed by STATs.com); but his sacked-when-passing rate suggests Rivers gambles on protection or would rather take a sack than throw it away even when he's aware of pocket-collapsing pressure (if River were "unaware", his sack-rate-per-attempt would presumably be higher).

Guess what?

I'm not questioning your data (nor Chase's for that matter). I am questioning its usefulness and the perception of ex ante reasoning to justify your conclusions (e.g., Rivers is "probably" not responsible for SD's underperformance, then looking for exculpatory data which hardly exculpates). Especially when, in the case of sack-fumbles, all fumbles/sack-fumbles are treated alike (-50 yds) -- including those that result in opposing game-winning or margin-of-victory TDs. n.b. I absolutely concde the difficulty of mining "meaning" from fumble data without outcome data for the ensuing (opposing) possession.

I also thought your earlier question was provocative and astute: Should Rivers/the Chargers be winning more (more games, more often) based on his numbers?

I don't follow Rivers or the Chargers but the question intrigued enough to look at other numbers. STATS.com revealed, among other things, that Rivers currently leads the league in sacked-when-passing.

Similarly, I had NO idea the Chargers lost on 10/10 in no small part because of Rivers' sack-fumbles. It was your assertion that SD losses relative to Rivers' numbers were "probably the fault of somebody other than the QB" that prompted my Google search [ bit.ly/bn2ABQ ] and at least 1 returned hit that pointed directly at Rivers and his game-affecting turnovers (the story linked to was the 1st and only hit I've examined thus far).

Shorter: I'm only following the path led by your original question. I don't think GOAT data is any more or less useful than actual game synopses absent similarly useful context.

... [share of] culpability.

Those drive stats look off. Football Outsiders have them and the Chargers have never been above 2.6 for any of the last 3 seasons.

For as bad as SD's special teams have been, scores like 21-14, 27-20, 35-27, 20-17 and 23-20 aren't even close to looking like "DEF/ST fall apart" days. Those are manageable games. Only the Raider score looks bad, but they scored a huge TD when Rivers fumbled the ball late and was complaining to a referee instead of making a tackle attempt. Nothing to this point is going to get me on the "woe is Philip Rivers 2010" bandwagon. He needs to turn more of those yards into points. And there's more to it than dropping 38-41 on teams like Jacksonville and Arizona.

Scott.... If you had the time to watch every Chargers game and follow every play, it would be blatantly obvious that the only reason the Bolts were in those games that they lost this year was solely because of Rivers' astounding play. The Bolts have had the worst ST through 9 games in NFL history. Their TO differential is their worst in a decade (Rivers is but a small part with only 8 picks). I guess when they win their last 9 games of the season perhaps you will warm up to him...

I posted this on the forum that linked here:
Let me start by saying I have always hated the concept of "QB wins" as a stat. Now, please correct me if I'm wrong here, but if I understand this right, he is using YPA as a predictor for wins?? Think about the absurdity of that for a second. Any QB facing a prevent defense is in a situation where he's likely going to lose ("preventing a win" jokes aside) but also rack up a ton of YPA. Further, consider that an eight yard completion on 3rd and 10 is a great way QBs inflate YPA and QB rating, despite the fact the play was an unsuccessful endeavor.

Personally, I think a much better predictor would be Football Outsiders' DVOA or DYAR--far more telling QB efficiency stats than simply YPA.