Filed under:

# The Playoff Threshold in the West

I have been a huge fan and recommender of sportsclubstats.com since before I had a blog. In fact, I started blogging because I wanted to put all of my favorite useful links on one page for my own private use, and SCS was among the first of those links. These days, SCS is the go-to site for people referencing NHL playoff odds, as was demonstrated by Jim Fox quoting the site on a recent Kings broadcast.

However, people keep saying that SCS tells you the odds of team x making the playoffs. It doesn't. It runs 10 million simulations of the remainder of the season (that's 10 million each day) and tells you where each team finished, giving percentages for each final standings position for each team, using -- and this is the key -- either a virtual "coin toss" to determine the outcome of each game, or a "weighted" algorithm which is a random coin-toss slightly skewed to favor teams with better records, and taking into account home-ice advantages. In other words, the simulation assumes that the outcome of every game is totally random or nearly random with the assumption that better teams and home teams are favored over worse teams and away teams.

The problem is, of course, that the outcomes of games are not random, and every season we see numerous examples of teams that sucked in the first half coming on strong in the second half (e.g. Anaheim last year) and teams that were on fire in the beginning of the season crapping out in the second half (e.g. Dallas last year). The seeds are so tightly packed, especially in the West (where the difference between 3rd and 12th is usually no more than a handful of points), that just one such cold-hot or hot-cold rollercoaster throws off the whole "matrix." And we usually see more than one. The Kings last season were a pretty good example of this, having gone on a historic hot-streak the last three months of the season.

I don't feel comfortable using SCS to quote odds of a team making the playoffs. But I love using their numbers for another, related, purpose. That is, to estimate the number of points a team will need to qualify for the playoffs. Not "will team x make it?" but "will this number be good enough to make it?"

The relevant section of SCS is each team's individual page, where it shows the outcomes of the 10 million simulations, and shows -- for each final point total -- in what percentage of simulations that specific point total was good enough to qualify for the playoffs for that particular team.

I will look at the numbers for each team -- not just any one team, but every team in the conference -- to see where each team's final point total crosses the 90% and then the 95% threshold for playoff qualification. For example, as of today, for the Kings, 94 points makes the playoffs 91.8% of the time, and 95 points makes the playoffs 96.4% of the time. 96 points is good enough 99% of the time. 99 points, 100% of the time. For (to pick another team) Detroit, they cross the 90% threshold at 93 points, 95% at 94 points, etc., and 100% at 98 points. That is, their threshold, on average, is a point lower than the Kings'. While I haven't looked at every team's numbers today, last I checked all the Western teams crossed the 95% threshold between 94-96 points. (And it doesn't matter if you use the 50/50 or weighted model; you still are very, very likely -- in 19 out of 20 simulations -- to qualify for the playoffs with 95 points.)

What is the target number for playoffs? " LA Kings Insider

With 58 points through 50 games, the Kings are on pace for 95 points. Is that enough to make the playoffs? In recent seasons, yes, for the most part. Since the start of the 2005-06 season, and the advent of three-point games, the eighth-place team in the Western Conference has finished with the following point totals: 95, 96, 91, 91, 95 and 97. The eye-opener is last season, when Dallas finished with 95 points and finished in ninth place, two points behind Chicago. For the Kings to get to 97 points, they would need 39 points in their last 32 games, the equivalent of a 19-12-1 record. And, as Darryl Sutter points out, it’s not simply a race to a certain number. It’s also about holding off teams, from ninth to 12th place, that are currently very much in striking distance of the top eight.

I only have one thing to add to Rich's observations, and that has to do with his comment that since 2005, the eighth place team finished with 95, 96, 91, 91. 95 and 97 points. What's interesting about that is that there is a difference between where the 8th place team finished and the point total they needed to make the playoffs. Looking at the 8th and 9th place totals makes this clear(er).

• 2006 -- 8th: 95 / 9th: 92.
• 2007 -- 8th: 96 / 9th: 95.
• 2008 -- 8th: 91 / 9th: 88.
• 2009 -- 8th: 91 / 9th: 89.
• 2010 -- 8th: 95 / 9th: 90.
• 2011 -- 8th: 97 / 9th: 95.
So, for example, look at 2010. The 8th place team finished with 95 points. But they could have finished with 91 points and still made it, because the 9th place team only had 90 points. They might even have gotten to 8th place with as few as 90 points, depending on how the tie-breaker played out (and no I didn't go back and calculate several hypothetical tie-breakers). Similarly, just because the 8th place team made the playoffs with (again, using 2010) 95 points, doesn't mean that the 9th place team would have made it with 95 points. The teams on the outside looking in always have exceed the current 8th place team's point total or else tie them and prevail in the tie-breaker.

I'm on the edge of confusing myself. But there's a real point here. When Rich says "95, 96, 91, 91, 95, 97" were the playoff numbers since 2005, that actually depends on whether you are looking at the playoffs from the point of view of a 1-8 seed or a 9-15 seed. 9-15 needed to (and failed to) get at least the numbers Rich mentioned. But the 1-8 seeds needed to get at least (and succeeded in getting at least) 93, 96, 89, 90, 91, 96 points in those same years. Those "magic numbers" are 2, 0, 2, 1, 4 and 1 points lower than Rich's, respectively.

On average, the 9-15 teams would have needed at least 94.167 points to rise to at least 8th.
On average, the 1-8 teams needed to stay above 92.5 points to stay at least 8th.

### Bonus inter-blog editorial content

A Real Look At The NHL Standings – Don’t Tell Hammond. " Surly & Scribe's L.A. Kings Blog

Never mind how pissed I am that Hammond butchered a perfectly fine idea for a full article by using incomplete stats in a superficial subject-skim, a full 11 days after my original article posted here at S&S. He has a right to publish whatever he wants, and who is to say he actually saw my article [...] complete with all he posted and more, especially the chart of teams’ games left. It’s not an issue; I’m over it. Actually, I have no idea why you would even bring it up; it doesn’t concern me in the least.

Anyway, here is a more in-depth update to the subject of remaining games, playoff chances and how the schedule (home vs. road games left) impacts those chances of various teams making the playoffs. The Kings have 12 home, and 20 road games left, having so far taken 55 points [actually 58 points -- Q] of 98 possible [actually 100 points possible -- Q] for a winning percentage of .5612 [actually, it's .580 -- Q].

Given the blogger's outrage, I was interested to check out the blogger's original post to see if it bore any similarity to Rich's post.

What’s that? Playoffs? Don’t talk about — Playoffs? You kidding me? Playoffs? " Surly & Scribe's L.A. Kings Blog