Don't count out Cavs and Nuggets yet

Let's start with the obvious and work our way down: It's not looking real good for the Cavs and Nuggets.

They're still alive heading into this weekend, but teams that trail 3-2 in a best-of-seven series are 16-81 from 1990 to now, a 16.5 percent success rate.

(Note: This article is referring only to series in the 2-2-1-1-1 format, which is the format used in all rounds before the NBA Finals.)

Based on that precedent, it seems probable that we'll be in Los Angeles on Thursday watching the Lakers and Magic tip off for the NBA crown.

But hold up a second. The oft-quoted numbers on the unlikelihood of a 3-2 comeback seem a bit skewed, don't they? In a lot of those series, we knew pretty darn well who the better team was, and most of the time, the team that led 3-2 ended up pounding their lesser rival in one of the final two games. For example, when the Lakers walked out of Game 5 against Houston with a 40-point victory a couple of weeks ago, I don't think a lot of folks were rushing to bet on a Rockets comeback.

But the two conference finals series aren't shaping up that way. Through five games on each side, I don't think anyone can definitively say that the Lakers have outplayed the Nuggets or that the Magic have outplayed Cleveland. They're ahead on games more because of timely plays than because they played better overall, and I don't think anybody would be shocked if at least one of the two teams came back.

And it turns out that when trying to project outcomes of 3-2 series, parity counts.

While Denver and Cleveland are in similar situations, they have provided different kinds of evidence that they can compete with the team up 3-2 in the series.

Denver has outscored the Lakers in the series (by a total of five points), and Cleveland is the higher seed in its series (and is trailing Orlando by two points in the series, the margin of victory in Orlando's Game 4 overtime win).

That makes these two series unusual, because in more than half the situations in which a team had a 3-2 lead, the higher seed has been ahead 3-2 and has outscored its opposition over the five games.

Teams in that exact situation virtually never lose. When the higher seed is up 3-2 and has a positive scoring margin, it is 48-3 over the past two decades (94.1 percent). The three exceptions are San Antonio coming back against New Orleans last year, the Lakers knocking out Sacramento in 2002 and Houston overcoming Phoenix in 1995. So basically, unless you're playing a defending champion, you're guaranteed to advance if you're the home team and you're ahead on both games and points.

But that's not the circumstance in either series this year, so the odds shift quite a bit.

In situations in which the home team isn't ahead on both points and games, the team leading 3-2 still fares quite well: 33-13 (71.7 percent). But 33-13 is a far cry from 48-3, so the Lakers and Magic aren't exactly automatic propositions.

Delving further, we see that the Lakers' situation is actually a fairly rare one. It's only the 14th time a team has been up 3-2 but down in total points in the past two decades, and the team ahead 3-2 has won 10 of those series (76.7 percent). One of the exceptions happened two weeks ago, when Orlando beat Boston; the others were Utah downing Houston in 2006 and the Knicks knocking off Miami in 2000.

Thus, the Nuggets' odds are much stronger than they first appear to be. The overall record of higher seeds that lead 3-2 is 58-6, but history says Denver has about a 1-in-4 shot of winning the series when scoring margin is taken into account. This makes total sense: If the two teams are basically even -- which was my premise entering this series, and one I haven't found any reason to discard since -- Denver's chances would be almost exactly 1-in-4, which is essentially like winning two consecutive coin flips. (They'd be exactly 1-in-4 if we ignored home court, but factoring in home court means their odds are more like 22.5 percent.)

Similarly, Cleveland's chances improve substantially once you look at the historical context of teams in their particular position.

Lower seeds such as Orlando are 23-7 when they have a 3-2 lead since 1990, a 76.7 percent success rate, leaving Cleveland at 23.3 percent. You'll quickly notice that this is almost exactly the same as in the coin-flip scenario presented above for the Nuggets.

The Magic have the superior scoring margin through the five games, but in this case we can't find much evidence that it matters. Lower seeds ahead on both points and games are 15-4 in this scenario (78.9 percent), while those that are ahead 3-2 but down in points are 8-3 (72.7 percent). So the teams that have scored more than their opponents have a slightly better record, but no statistician would accept this as a relevant conclusion given the relatively small samples and the closeness of the two percentages.

Certainly the Lakers and Magic can head into Friday night feeling fairly optimistic that they'll be seeing each other next week. If we assume a 22.5 percent chance for the Nuggets to advance based on the logic above, and a 23.3 percent chance that the Cavs will advance, then a little more than half of the potential outcomes produce a Lakers-Magic Finals: 59.4 percent, to be exact.

A Denver-Cleveland pairing (Melo-Bron, anyone?) remains highly unlikely at just 5.3 percent, but Nuggets-Magic (17.3 percent) and the much-anticipated Cavs-Lakers matchup (18.1 percent) are still on the table.

So when they tip off for the Game 6s on Friday and Saturday, and you see a graphic reporting the odds are stacked against teams trailing 3-2, take it with a grain of salt. True, the Nuggets and Cavs aren't in great shape, but their odds are far better than the overall record of teams trailing 3-2 would have you believe.

All in all, there appears to be about a 41 percent chance that either the Cavs or the Nuggets, or both, will be playing for rings next week.

John Hollinger writes for ESPN Insider. To e-mail him, click here.