Commentary

The Playoff Predictor: Explaining how it works

Originally Published: December 6, 2007
By John Hollinger | ESPN Insider

Welcome, everyone, to our latest creation: The Hollinger Playoff Odds. The idea behind it is to try to predict a team's final regular-season record by having a computer play out the remainder of the season's schedule, allowing us to figure out which teams have the upper hand in the playoff chase. Then we have it play out the playoffs and the draft lottery and report back the results.

Better yet, because the games are being played on a computer and not in real life, we can repeat it 5,000 times and get a more definitive picture of the odds of certain events happening.

For those who are wondering, this tool is completely automated. It will greatly disappoint some readers to learn that my "obvious bias" against your clearly championship-bound team has nothing whatsoever to do with where they appear in the projected standings.

In all seriousness, fans of, say, the Nets or Bulls may feel slighted since this tool isn't very fond of their chances at the moment. But all it's saying is that neither team is likely to make the postseason if they continue playing the way they've played thus far. I doubt fans of either team would argue much with that assessment.

That's the key assumption here: that teams will continue to play the way they've been playing. Based on that idea, every day, the computer plays out the season based on each club's Hollinger Power Ranking on that day. There's only one caveat: To limit the highs and lows, there is a function that regresses performance to the mean.

What regression to the mean says, in layman's terms, is that even if the Celtics stay awesome, they're unlikely to be quite this awesome for a full 82 games. The regression term does most of its work in November and December, diminishing in importance as the season goes on. But it's especially important early because, for instance, it prevents us from declaring a team will go 82-0 simply because they won their first few games.

While it's interesting as heck to play with even now, this is a tool that becomes more valuable the later in the season we get, because that's when schedule discrepancies between playoff contenders really become important. You'll be able to tell how important it is for a team to have, say, three extra home games in that final month compared to another club it's tied with in the standings. And for those of you in places like Minnesota and Philadelphia, you can see how your team's quest for Ping-Pong balls is going.

With that said, let's deal with a couple of key questions people are likely to ask about their teams:

So the Spurs have only a 9 percent chance of repeating? Remember the key phrase here: If teams keep playing the way they've been playing. That's what the entire projection is based on, because that's the only information the projection "knows." Based on what's happened so far in 2007-08, if Boston and Orlando keep playing the way they have been, then whoever wins the East will be favored in the Finals. (Side note: When was the last time you heard that phrase?) As the year goes on, the projection will "know" a lot more about these teams and will adjust accordingly.

So the Jazz are a mortal lock to win the Northwest? To this method's eyes they are, because it doesn't know that the Nuggets are fielding half a team right now. If and when the Nuggets get their lineup intact, their Hollinger Power Ranking should improve, and based on that so will their projected finish. That said, by then it's possible they will have lost too much ground, because the Jazz look awfully strong so far.

So what's up with the rest of the East? This tool really shows how the East has split into the Big Three and the Little 12. Orlando, Detroit and Boston all show up as locks to win their respective divisions, while Atlanta limps into the No. 8 seed with just 37 wins.

What about injuries? That's another key factor here. This method doesn't know that Sacramento just lost Kevin Martin for a month, or that Cleveland hasn't had LeBron James while it's gotten its brains beat in this week. As the season goes on and teams go through more peaks and valleys, the impact of this will be less extreme, but this early in the season it has a major effect on the projected finish, so some teams may be in a different position than you expected.

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