« In the NO Pod Ep. 305: Davis on the Warpath

In the NO Ep 306: Jrue the Terror and 8-0 »

# A Different Look at the Remaining Schedule, Part 2

*Punchline: The Pelicans and Nuggets are the teams in the top-8 with the most potential to fall out of the playoffs.*

The New Orleans Pelicans have 21 games left to play, with one tonight against I hate the Mavericks. A few weeks ago, I looked at the schedule trying to identify the importance of games by looking at clusters of teams. Since then, as always happens, some of the structure changed. With the schedule burning in the fire that is time, there is less fog, but remaining schedule quirks become ever more important.

**Approach**

We are trying to answer the simple question: Will the Pelicans make the Playoffs?

This involves various questions about likelihood that I don’t want to get into since others will. So, I ask the related question: What games are most important for the Pelicans making the Playoffs?

Clearly, any game involving the Pelicans is important. From there, it immediately gets to a point where many lose focus. This is because the top-8 teams don’t matter (to the top-8). They’re in. It’s the *ninth* team that matters the most. They are the line. The teams nearest them, either side have less importance, and importance drops as you move away from ninth. This may seem like no big deal. if so, adjust your perspective. The eighth position matter to the team in ninth, of course, or any team below them, but that’s not the focus of this analysis. Reflection on this should illustrate why a cluster-based approach should yield some useful information (relatively) simply, especially compared to detailed predictions based on matchups, rest, fit, and more.

The structure of the Playoffs requires teams to be among the top-8 in the West at the end of the season. This means how well teams in the East do overall does not matter, save the games involving teams from the West. It also means the overall record is not important, nor does the overall rank, as long as it is 8-or-higher in terms of making the Playoffs (all that can affect seeding otherwise). Also, we can reasonably eliminate teams in the West from our analysis since their fate is sealed, either in or out of the Playoffs, at least without drastic and prolonged issues. If those happen, we’ll re-visit those assumptions as we re-visit the standings in the normal course of events.

So, we’ll look at some clusters of teams and see what we learn.

**Groups**

For now, we just focus on 8 teams: Pelicans, Timberwolves, Spurs, Trail Blazers, Thunder, Nuggets, Clippers, and Jazz, as they are ranked 3-10 in the West, each having a reasonable chance both to make and to miss the Playoffs. We lump all the rest of the teams into the group “Other.” Those teams are not important when they play each others (likely, at least for this particular effort), but they are important when they play a team of interest, of course.

As it stands now, the Clippers and the Jazz are ranked 9th and 10th, so they are out of the top-8. These have to be the focus of the inquiry. As the teams in the top-8 slide past each other in the standings they are just trading playoff seeding (an important question in its own right, but one completely subordinated to whether you make it or not, and it’s much harder to answer at this point to boot). They are not pushing each other out of the Playoffs, even if they are contributing to that in some way. It’s the teams out of the Playoffs passing you that push you out; you have to change groups, not move around within one. I treat these two each as their own group (for now).

For reasons that become clear later, I end up looking at the Nuggets and Pelicans as their own groups (for now), then the Spurs, Timberwolves, Thunder, and Trail Blazers as a group (for now), though I report the details of each team for completeness.

**Calculating Risk**

Given the assumptions above about the “Other” teams being immaterial or likely unchanging, the jostling of teams in the top-8 in the West does not matter. So, the bar is the ninth ranked team. However, it is not the current standings that matter directly; they are an influence that becomes stronger as the season progresses. Rather, it is the ranking at the end. So, the potential for a team to finish ahead of another needs to be quantified to account for the unknown. Depending on the context, this is often called a team’s “Magic Number” (for something good) or “Tragic Number” (for something bad). The difference is perspective on the consequences. The basic idea is the same.

There are many ways to arrive at the calculation, but the following explanation makes good sense to me. Consider two teams, A and B, and someone wants to know if B will finish ahead of A in a season where teams all play an equal number of games by the end and no game ends in a tie (like basketball). The best either team can do is finish with a win total of

*(Current Wins) + (Remaining Games)*

since it’s possible to win out. B can not finish ahead of A if

*(B Current Wins) + (B Remaining Games) < (A Current Wins)*

at some point in the season since B simply lacks the number of chances left to pass A. This simple observation is the key to the analysis. Each time B plays a game, they either win or lose (duh). With a win, the potential to pass A is unchanged. Since a B Remaining Game became a B Current Win, nothing changes with the total. With a loss, however, they lose potential (a Remaining Game), and their ability to pass A diminishes, as the sum reduces by 1.

When A plays a game, A losing changes nothing in the calculation since it literally doesn’t appear in the calculation. However, when A wins, the A Current Wins increases by 1, and it is more difficult for B to pass A. So, what we need to consider here are each team’s current win totals, remaining games, and, depending on our perspective, one team’s wins and the other’s losses or vice versa.

Because our situation is slightly different, we modify the calculation slightly. I’m just going to worry about getting teams to tie at this point. Once the tiebreakers are settled to a practical level, we can sharpen things. For now, we aim for the tie and still get something out of the analysis. Also, since we are dealing with more than two teams, we do the pairwise calculations and aggregate the results appropriately. This includes looking for schedule quirks that can derail Tragic Number type analyses.

The following was produced prior to the March 4, 2018 games starting.

Let’s walk through it.

In the upper table, we see the teams of interest and the Other group. The number in each cell is the number of games remaining between the listed teams (or Other). The Clippers and Jazz games against the other teams of interest are highlighted in red, including the one game against each other. Below that, we see the teams’ current records and remaining games. Below that, the pair-wise calculations described above appear for the Clippers and Jazz against the other other teams of interest (not each other, as they aren’t trying to pass each other . . . that may appear to be necessary, but it’s incidental). The number there is the number of Clippers / Jazz losses plus the number of wins for each other team to tie that team. The larger the number the better for the Clippers / Jazz, which makes smaller numbers better for the teams currently in the top-8. Since the Clippers / Jazz need to just pass one team in the top-8, it’s the maximum of those numbers that matters. The current state of the primary tie-breakers are listed. The main takeaway is that these are not fully-determined where the action is. Thus, they are presented in order to show that they do not matter yet.

*This makes the Nuggets and Pelicans the most vulnerable teams.* Currently, the Pelicans are in the 5th seed and the Nuggets are in the 8th seed, but that’s immaterial. These are the teams with the most potential to fall out of the Playoffs. They need the most “to happen” to stay in, if you want to look at it that way.

The current standings suffer from the following distractions:

- Tiebreakers that may or not be fully-determined
- Win% being lower for teams that are the same number of “games back” but where one team has played more games
- Consider 2 teams. One team has one more win, one more loss than the other team. This team would be the same number of “games back” from any other team (tie-breakers aside). However, if the teams have winning records, then the extra 2 games in a 1-1 split lower the winning percentage, moving it toward 0.500 just by the nature of averaging. This percentage is immaterial for the end-of-season question, however, since all teams will have played an equal number of games by then. It’s a slightly obscuring illusion, which is harmful to casual analysis.

This is the punchline here. For anyone looking ahead, keep in mind that the battle is behind you. *Watch your six.* We’ll review this every week or so to see if there are some significant changes.

**Schedule Details**

- I’m not going to address likelihood of wins. That stuff is being taken care of elsewhere. Those considerations are necessary, and this work complements it. Both are needed to really understand risk, but what is contained here is under-represented.
- One particular hiccough in this kind of Tragic Number type analysis is when opponents play each other. For example, if a team needs two other teams to lose out for them to win out and capture the top spot in some league, that may seem possible looking at pairwise numbers. However, if those two opponents actually play each other, they can’t both lose out, so the team needing both to lose is in fact already eliminated. In this case, the Nuggets and Pelicans have completed their season series, so this is not an issue at this point. The 0 is italicized for this reason.
- The one game between the Clippers and Jazz (April 5 in Utah) has unknown importance at this point; it’s just too far out. In the end, root for the team with the least potential to knock the Pelicans out, of course.
- The Clippers play a number of games (8 of 21 remaining) against the teams they are threatening to knock out of the Playoffs. They’ll get good punches from those teams
*ceteris paribus*. The Jazz only have 5 of 19 against such teams. Looking only at Nuggets and Pelicans as opponents, the Jazz have just 1 (Pelicans) while the Clippers have 3 (2 are against the Pelicans). These games end up counting for 2, potentially, in the Clippers / Jazz quest for the Playoffs since the Pelicans and Nuggets are the easiest targets and drive the Tragic Numbers (for now). - Because the Pelicans have 21 games remaining and the Nuggets only 19, the Nuggets are slightly worse off from this perspective. By that logic, going into tonight, the 3-10 rankings in the West in terms of “making the Playoffs” are: Timberwolves, Trail Blazers, Thunder, Spurs, Pelicans, Nuggets, Clippers, Jazz.
- This will depend on what “the list” is on a given day, but the general rooting structure should be: Pelicans, anti-Clippers, anti-Jazz, then against the lower of the remaining teams when the teams of interest play each other. Clearly, if another team is playing one of these teams of interest, excepting the Pelicans, root against the teams of interest. This is based on the assumption that we are trying to optimize the chances to make the playoffs, not the position in the Playoffs IF you happen to make it. That would be different. I’m not getting into that at this point.
- The Pelicans and Nuggets each play a total of four games against the other four of the top teams of interest. The top teams of interest play five games among themselves. Hopefully those 5 don’t annoy me analytically. I’d hate to have to account for those.
- We’ll dig more into the tie-breakers as things evolve. They will involve common records, division records, and conferences records, depending. I don’t see the return in looking at all that this far out. The Cost-Benefit is out of whack compared to the return on just waiting to see how the primary determiners evolve.
- In the last 10, each of Nuggets, Pelicans, and Clippers are 7-3, and the Jazz are 8-2. You can ask yourselves if that performance will continue or not, but it goes to show to that small gaps in the records can be made up quickly when 3 of 4 teams are performing like that and one starts to play 0.500 ball. Just about every kind of volatility is present in this Playoff race. It’s great, and it’s infuriating.
- The Pelicans do themselves a massive favor if they can handle the Clippers. Also, that Jazz game Sunday would be another I’d hate to lose if I were them. Big week. They need to nail it.