There was just no doubt what would happen here.

Having gifted the Eagles with three non-offensive touchdowns, we'd battled back to take the lead. Our O had looked good, our run-game especially, but the comeback belonged to a truly amazing show by our D. Facing a timely mix of rush and coverage, the Eagles' O hadn't crossed midfield. And now, desperately needing to finish this game, we were all set up to put it away.

Third-and-three, from the Eagles' 11.

When the Niners broke the huddle, the play-clock showed about 10 seconds left. They spent half that time getting set at the line. With three seconds left, Colin Kaepernick tapped his helmet to signal an adjustment. And he called for the snap, a half-beat after the clock ran out.

Now it was third-and-eight.

A couple years ago, Kaepernick had a golden chance at instant immortality. He let it slip away, largely because of a sequence like this. And yet, as we've discussed, he doesn't learn from past mistakes. Even with something so totally basic—when you break the huddle, look at the play-clock—he simply doesn't seem to learn.

By all accounts, Kaepernick's an intelligent guy. (For what it's worth, his Wonderlic score was 10 points higher than Peyton Manning's.) It's outrageous how often he just plays dumb.

Of course, making him learn is the job of his coaches, who might not be much better themselves. Which leads us back to third-and-eight.

Two weeks ago, the Niners faced third-and-six from the Bears' seven. A touchdown might've put it away. But instead, Greg Roman called for a run up the middle and settled for three. And as you know, the Niners lost.

So here was a chance for Roman to show how much he'd learned. And apparently, what he'd learned was nothing. Another run, and another field-goal.

Last week, we reminded ourselves of one characteristic of the prior regime: its stubborn refusal to stick with what worked. This week, let's remind ourselves of another: its stubborn refusal to play for the win. You might recall one game in particular, against the Vikings in Minnesota. On a late third-and-six where a conversion would've iced it, we ran the ball, punted away, and gave Brett Favre a chance to win as time ran out, which is precisely what he did. Now Roman had given that chance to the Eagles. And as I put it five years ago, "as it always has since the dawn of time, playing not to lose set us up to do exactly that."

Our D had been magnificent, but the Football Fates pounce on decisions like these. And even though the Eagles were starting out at their own nine, there was just no doubt what would happen here.

The Eagles were inexorable. A pass for 11, a run for 5. Passes for 15, 11, and 14. A sack set up a third-and-14, but they just went ahead and picked up 17. One more pass for 11 more—after Riley Cooper had let the game-winner go right through his hands—and they were set up with a first-and-goal at the Niners' six.

Whatever god Greg Roman believes in, you can bet he was praying hard.

And praying well, for the D delivered. Great coverage forced a throwaway, Antoine Bethea made the play of the year by tackling LeSean McCoy just shy of the goal line, and then he blitzed to force another incompletion. And finally, on fourth down—"Huge play here," said Troy Aikman, perceptively—Aaron Lynch's pressure, plus more great coverage, effectively put an end to things.

In the NFL, that's the difference between life and death. And mercifully, the Niners live.

The defense seems to have found something, a way to make do 'til the troops get back (though the special teams—aside from Phil Dawson—were abysmal). But the questions on offense just won't go away. And those questions, for the most part, remain philosophical.

When Roman arrived, the question of offensive identity came up immediately. Were we a pass-first WCO? Or were we a smashmouth ground-and-pound? We seemed to drift from one to the other and back again, almost indiscriminately. Eventually there seemed to be a method to the madness: we were trying to be equally capable of both, and it seems like that's still the idea today.

It sounds like a really good idea, too, literally the best of both worlds, and Roman's Niners have likely come closer to success than anyone else. But the Niners now might be too versatile for their own good. By almost constantly shifting modes—not just game to game or quarter to quarter, but sometimes even series to series or play to play—they seem unable to get any rhythm, and the whole operation just seems to bog down. If it weren't for Kaepernick's ad-libbing—both TDs on Sunday were essentially on busted plays—things would look even worse than they do.

The conventional wisdom is that we should stop dicking around with this pass-happy crap and get back to doing "what we do best," which allegedly is running the ball. Naturally, I disagree; his clock-management issues aside, Kaepernick is still our best weapon, and he's at his most dangerous out of the spread, where he can crush a D with his arm if he can and his legs if he must. Plus, of course, as we've discussed forever, it's a passing league, and passing teams win. But though I'd prefer the passing identity, the crucial thing now is to pick some identity: pick it, commit to it, and get into the flow of executing it.

On the other hand, as Sunday proved, the identity issue is just half the battle. The Niners' refusal to play for the win has now reached epidemic proportions. This time we got away with it, barely. But as a long-term philosophy, it's obviously doomed.

Because, of course, that's what history shows. But Greg Roman, much like his quarterback, isn't a student of history.