Monday, April 25, 2011
A better way to look at teams in the clutch
By Tom Haberstroh
Fire up the sirens. The Heat lost another close game on Sunday night. As a team, the Heat are now 6-15 in games decided by five points or fewer. That has been the Heat's defining "close game" statistic this season and they've been saddled with a bad record since October.
And here's why that stat is incredibly misleading. It's a rear-view mirror stat that doesn't tell us what we think it does.
Take Thursday's Game 3 for example. The Heat won by six points. Not a close game? False.
The Heat's lead with 43 seconds remaining stood at just four points after Lou Williams hit a 3-pointer. The Sixers came up short on their next possession and a pair of free throws by James Jones with three seconds left pushed the Heat's final lead to six points. If Jones missed his second free throw, should we characterize the game differently?
Go back to the Heat's Game 1 victory. Miami held a one point lead with 2:02 remaining in the fourth quarter. Close game, right? Not by the conventional measure, which states that the final score should be the end-all-be-all. The Heat stomped on the Sixers throats down the stretch and won by eight points. Nope, not a close game according the industry standard.
It seems long overdue but we need a new measure. The Heat play themselves out of the conventional "close game" parameters because they often play very good in close games. What we really ought to ask is, when the game gets close in the final minutes, do the Heat pull out the win, regardless of the final score?
So let's keep it simple and hash out the parameters. When the game gets within five points anytime during the last five minutes of the game, we'll call that a close game. What is the Heat's record in those games?
The answer: 24-21.
No, not great by any means, but above .500, and certainly better than the flawed 6-15 record that everyone has blindly accepted as gospel. Their 24-21 record ranks as the 12th best in the league whereas the 6-15 record placed them second-to-last next to the lowly Timberwolves.
The Heat's close game win percentage jumps from .286 by the old measure to .533 in the new measure, which is by far the biggest difference in the NBA.
Here are the teams that are most improved by the adjustment.
NBA team record in close scores vs. close games
San Antonio Spurs
Golden State Warriors
Source: Data courtesy of NBA StatsCube. Note: "oWins" indicates wins in the "old" measure of close games. "nWins" indicates wins in the "new" measure.
Doing some simple math, the Heat pushed 18 games that were close in the final minutes outside the old 5-point boundary. Those wins need to be accounted for, if we're being honest about what's a close game. On the other side, the Heat lost six games that were close in the final minutes but the losing deficit exceeded five points on the final scoreboard. Once we account for those games, the Heat are 24-21. (Note: you can find the entire rankings here, sorted by best close game record).
I should stress -- and stress three times over -- that this isn't an attempt to pump up the Heat. This is an attempt to improve upon a flawed model. The Heat's 6-15 record doesn't tell us how they do in close games; it merely tells us how they do with a close final score. The Heat lost a close game on Sunday, but they pulled out Games 1 and 3.
Let's not allow an imperfect stat cloud our judgment.