Even For a Team Down 3-1, the Rangers Are in Bad Shape


The New York Rangers and Los Angeles Kings will meet in Los Angeles on Friday night for Game 5 of the Stanley Cup Final. Before Game 4, when the Rangers avoided the sweep, I said that simple probability was not on the Rangers’ side after they fell behind 3-0. Even if we assumed they were evenly matched with LA, New York had just a 6.1 percent chance of winning four consecutive games before the Kings could win one.

Now that the Rangers have a win under their belt, their chances are better — improving to about 11.1 percent (if we again assume the teams are evenly matched). But the odds are still long: New York still needs to beat Los Angeles three times in a row, including two games on the road.

We were wondering, though, if the sequence of wins and losses matters in a 3-1 series. For instance, by winning Game 4 and snapping a three-game Kings winning streak, do the Rangers have more momentum than, say, a team that won Game 1 and proceeded to lose three straight?

The simplistic way of looking at this question would be to count the number of times each possible four-game sequence (e.g. LLLW, LLWL, LWLL, WLLL) occurred, and to determine how many times playoff team down 3-1 ended up winning the series after each sequence. From 1968-2013, here are those numbers:

rangers_sequence_paine

This seems to suggest the Rangers picked the worst possible sequence of wins and losses for a team down 3-1. But, of course, selection bias could be rearing its ugly head here: Teams trailing 3-1 that won in Game 1 or Game 2 are more likely to have been at home for the first two games of the series (historically, the NHL has used a 2-2-1-1-1 format for the majority of its playoff series). This doesn’t confer any special advantage — there’s no format under which teams would have more home games than their opponent through the first four games of a series — but it does tell us they were seeded higher, meaning they had the better record and were probably the better team. All else being equal, those are the types of teams you’d expect to come back more.

Here’s what happens if we restrict to looking at teams that started the series on the road:

rangers_sequence_paine_2
The difference between teams like the Rangers — who lost the first three games of the series but won Game 4 — and teams that lost the first two games, won Game 3 and lost Game 4 is negligible after we remove teams that were on the road in Games 3 and 4. In that sense, the sequencing of the Rangers’ home wins and losses thus far doesn’t matter.

However, there remains a rather large effect for teams down 3-1 who managed to get their win on the road, especially in Game 1. Teams who take Game 1 in the other team’s building and then proceed to lose three straight tend to win the series at a rate more than three times that of teams who trail 3-1 via any other sequence of events. Even in a relatively small sample of series, this is a statistically significant result.

The question is, why? We’ve removed the bias of including teams that were at home for Games 1 and 2, so this sample strictly deals with teams who, for seeding reasons, did not have home-ice advantage (and were therefore less likely to be the better team). And all still had to win three consecutive games, including two more on the road. My only hypothesis is that it must matter for a team to show it is capable of beating the opponent on the road early in the series. But it’s possible I’m missing a theory, so please leave yours in the comments below.

At any rate, those numbers aren’t of much consequence to the Rangers, because they didn’t get their win in Los Angeles. The history of teams trailing 3-1 offers little hope for teams whose lone win came at home, as New York’s has.