Based on their recent play:

Consider: The Warriors were 26-35 on March 4 after Don Nelson's technical foul at the buzzer cost them the game against Washington. Since then they're 17-6, including a surprisingly easy Game 1 win at Dallas. It's not a coincidence that this is also when guard Baron Davis returned to the lineup, giving the Warriors a healthy backcourt for the first time all season.

Yes, a couple of these games get an asterisk -- particularly those in the final week, which consisted of a blowout of Dallas' scrubs and routs of the Tankerwolves and Tank Blazers. On the other hand, Davis missed a game in that stretch which accounted for one of the five losses, and he only lasted 11 minutes in a squeaker over Memphis.

Let's throw all of those out and just focus on the games Davis started and finished since he came back, and games played against teams that actually were trying to win. That gives us 17 contests to work with, and the Warriors' record in those games was 13-4.

MAVS' FULL SEASON VS. WARRIORS DURING STREAK
MAVERICKS WARRIORS
W-L 67-15 13-4
Offensive Efficiency 108.5 108.8
Defensive Efficiency 100.6 102.3
Scoring margin +7.2 +6.6

Surprisingly, the small-ball Warriors have become competent defenders in that stretch -- even with Al Harrington masquerading as a center. Golden State's defensive efficiency in those 17 games is 102.3, which would rank ninth in the NBA. This tends to be obscured by the fact that they play the league's fastest pace, so occasionally they'll give up big point totals -- Golden State allowed 110 or more points five times in those 17 games. But the return of Davis, who is an awesome defender when motivated, clearly has pumped up the D.

And, like Denver, the Warriors have put together a devastating offensive attack. Golden State's offensive efficiency in this stretch is 108.8, which for the season was bested only by Phoenix. That's doubly impressive considering the 17-game stretch included several elite defensive teams: two games against the Spurs and Mavs, for instance, and one each against the Rockets and Pistons.

Golden State's victory margin in that time is even more impressive than the Nuggets': An average of +6.6 points per game -- nearly matching the +7.2 that Dallas rang up during its 67-win season. And as with Denver, it was a slightly road-heavy slate, with nine contests away from home.

Throw those numbers into the machine, and suddenly the Warriors-Mavs series gets a lot more interesting. The Warriors entered the series with a 37.8 percent chance of winning, and after the Game 1 upset -- get this -- there's a 59.6 percent chance of Golden State taking the series. In other words, the 67-win team is now the underdog.

So with that said, here's the big question: How much faith do we put in these late-season surges, and the revised odds that flow from them? Certainly the exercise above puts both the Nuggets and Warriors in the best light possible, evaluating them only with 100 percent healthy lineups and only during their most recent runs of success.

Nonetheless, it does seem to have some relevance. In both cases, the evaluation period is nearly a quarter of a season, so it's not like our sample size is absurdly small. Additionally, there's a valid reason for evaluating these teams when fully healthy -- they're both fully healthy right now, and we're trying to determine how they'll fare in the immediate future.

All told, then, I would venture to say that these numbers are a lot closer to reality for the Warriors and Nuggets than their original puny odds. And if that's the case, one has to consider another set of odds. Multiply the probabilities above and you'll find a 1-in-4 chance that both Dallas and San Antonio -- the two favorites to win the title -- fail to make it out of Round 1.

Who said the first round wasn't going to be interesting?