Finals Home-Road Format Is Close To Irrelevant

In last night's Game 1 between Los Angeles and Boston, ESPN/ABC's Jeff Van Gundy proposed changing the NBA Finals format from 2-3-2 (two home games for Team A, three home games for Team B, two home games for Team A) to a 2-2-1-1-1 format, which is used in all preceding playoff series. There are good reasons for the existing scheme: The 2-3-2 format requires only two inter-city journeys rather than, potentially, four. When East meets West, inter-city travel is generally further than in the intra-conference playoffs, and there are also many more journalists covering the NBA Finals compared to earlier rounds, so the airfare costs and hassle of 2-2-1-1-1 could be enormous. So the NBA's logic is patent.

But what of Van Gundy's claim? He seemed to argue that it is difficult for the team with the middle three home games to win all three and thereby fully exercise its home-court entitlement. But under his scheme, Team B would get a home date for Game 3, Game 4, and Game 6. Why should we think that the probability of winning Games 3, 4, & 6 at home (after a Game 5 on the road) is higher than the probability of winning Games 3, 4, & 5 consecutively? To be sure, few Team Bs in the last couple decades — only Detroit in 2004 and Miami in 2006 — have won the middle three games at home. But the sample size is low: only 20 series in those two decades. How many teams, in tightly-contested series (thus, consider only the conference finals) have won games 3, 4, & 6 at home? I doubt much more than four in the last 40 conference final series.


What's more, from the perspective of Team B, the 2-3-2 format is more forgiving. Assume Team B has a 2/3 chance of winning any given home game, and assume each game result is independent of other games. Assume further that Team A is 99% dominant at home and already won both of Game 1 and 2 on Team A's floor. If you win two of your three middle home games (probability = 12/27, or 44%), or all three of your middle home games (probability = 8/27, or 30%) then you can extend the series to at least Game 6 and make a respectable showing. (Thus, your chance of getting to Game 6 is 74%.) In a 2-2-1-1-1 format, unless Team B wins both of Game 3 and Game 4 at home (probability = 4/9, or 44%) , then Team A can finish off the series in its arena in Game 5. Given this disparity in likelihood, why not take the comfort of the 2-3-2 format?

Now, the dominant Team A might not like this, but they're going to win the series anyway, and lengthening the series means more ad revenue for everyone, so what's the harm?