New User? Sign Up / Forgot Your Password?
All Live
Projections
    All Live Projections

    Behind the Schedule, Part V (08/19/16)

    By Sean Pyritz @srpyritz
    Welcome to the fifth installment in what will likely be a six-part series on the NBA schedule. If you missed any of the previous installments, I encourage you to go back through the archives because we've been digging up artifacts each week. Today we will finish up our exploration of the past with our grandest scope yet – the final stepping stone before taking the leap into the future next week.

    In Part II, we introduced the idea of win transfers. With a magnifying glass on the Atlanta Hawks, comparing each direct matchup across seasons, we got an unencumbered look at how wins are transferred across the league from year to year. Expanding the scope to the whole league requires trading the magnifying glass in for the Hubble Space Telescope, a very expensive proposition. However, we can take a more affordable and insightful route on the fringe of the NBA schedule universe.

    Much like the NFL, the NBA has limited the number of games, so there is not a uniformity to the schedule. Each season, four non-division teams within conference get a fourth game with a particular team. While the majority of the schedule is set, there is a guaranteed four game change to cycle through additional games versus different opponents. The table below captures how this cycle has impacted the wins of each team from the 2015 season to the 2016 season.



    As we demonstrated last week, even a one game swing is significant in the NBA, making this table very intriguing. Certainly there are lots of factors in why a team would perform differently in a four game stretch from season to season, but we are here to investigate those factors stemming from the schedule. Take the Indiana Pacers for instance. They dropped three wins from their total – an amount that would have put them tied for 3rd in the East. This is especially surprising because the 2016 season saw healthy seasons from Paul George and George Hill, both of whom missed significant time in 2015. Peeling back the layers we can see that the teams Indiana had to face for a fourth time in 2016 were substantially better than those in 2015. Combined the four teams Indiana faced in 2015 had an average winning percentage of .347 and Net Rating of -5.5 (a mark that would have placed in the bottom five this past season). In 2016, faced with two playoff teams – Atlanta and Toronto – the combined winning percentage jumped to .488 with the Net Rating following suit to -0.22 (close to league average). Thankfully, they faced each opponent on equal rest or else they might have dropped all four games, but more on that in a moment.

    Until now, we've been viewing win transfers on a matchup basis, each team versus each team. There is another dimension in the schedule universe though – time. If we compare each game chronologically, it makes for interesting results because each team is not directly trading wins with another team, but with itself. For example, the Orlando Magic played the Cleveland Cavaliers in their 16th game of the 2015 season, losing on the road on a third game in four nights. Fast forward to 2016, the Magic faced a much weaker Bucks team at home in the 16th game slot and came away with the victory. Framing the entire league through this dimension, a few patterns emerge that speak to the overall impact of the schedule. In these win swapping situations, the wins are much more likely to come at home, so flipping from home to the road alone in a particular game slot spells trouble, regardless of opponent.

    Differences also emerge when we introduce the concept of “upsets.” Not all teams are created equal, not all matchups are created equal, therefore, not all wins are created equal. Sometimes, the odds-on favorite takes the loss. For our purposes, I've simply defined an “upset” as when a team with a higher net rating on the season loses to one with a lower net rating – note that this is a quick and dirty method that does not consider home court advantage, players missed, or even changes in play over the course of a season. On average, over the past two seasons, roughly 31 percent of games played result in an “upset” as I've classified it. Yet, when we compare to just the sample of games in the slots where wins and losses are flipped between seasons, we see that the games that have flipped have had both a higher probability of upset to signal a flip and a higher probability of upset to complete the switch the next season – for both wins and losses. In other words, it is likely that if a team completes more “upsets” than is typical, it is also more likely to get “upset” in a similar position in the season the following year. I want to reiterate that I am choosing to slice the schedule in this manner to explore, not to discount the realities of the players, coaches, referees, injuries, etc. on the outcomes of games.

    Before we wrap up and take a look at next season, it is appropriate to put a bow on what we've discovered so far and what we might expect to help us when glance through the crystal ball next week. With the focus on Net Rating (the measure of efficiency we introduced in Part IV), the effects of the schedule factors league-wide can shine. For reference, the average margin of victory in terms of Net Rating in all games over the past two seasons is 11.7. Benchmarking the effects of win transfers – either matchup based or game slot based – there is not much difference to really give any type of indication which games might be vulnerable to flip (this an area for continued research for schedule enthusiasts everywhere). Compare that win figure to the average home court advantage over the last two seasons: 2.7, almost 25 percent of the total.

    Jumping back to rest days for a moment, we can see that the theories on performance due to rest bear true in the Net Ratings across the league. The average Net Rating for zero days of rest, one day of rest, and two or three days of rest ascend from -1.7 to 0.3 to 1.2, respectively. Most of the improvement in is a reflection of consistent increasing stinginess on the defensive side of the ball with more rest while the offense flattens out. Entering the cave, a bit deeper, the schedule quirks echo the same sentiment. Three games in four nights earned a -0.6 Net Rating, while four games in five nights brought it all the way down to -2.1 – a major disadvantage, no wonder the league is eliminating them. Finally, combining rest days between teams, the team with the rest advantage has been truly fortunate, receiving a 1.9 Net Rating boost. The NBA could and should do a better job reconciling these differences – the average rest advantage in terms of days is 1.3, by far the greatest of any of the scenarios mentioned. With all that said, we are now prepared to look ahead to the 2016-17 NBA season and see who might have a head start thanks to the schedule alone.
    Print This Article
    PREVIOUS ARTICLE
    Behind the Schedule, Part VI - 2016-17 Breakdown (8/27/16)
    NEXT ARTICLE
    Behind the Schedule, Part IV (08/12/16)
    Want More Tips and Information? Click To Register!

    COMMENTS

    SEARCH BLOGS

    ARCHIVE

    The Predictalator

    05/10/2017 Highlight: With baseball season now into its second month, we can leverage our ResultsFinder tool to profile our season-to-date numbers. MLB sides and totals continue to remain extremely profitable despite a recent run of bad variance to start May. The Predictalator's record of 272-250 on sides and totals this season doesn't look all that impressive, but that's because the simulations have sided with underdogs more often than favorites. Through May 9, the Predictalator has generated $547 for an average $50 bettor. "Normal" or better sides, totals, and run lines (our strongest edges) are 14-10 (+$118) season-to-date.

    Check out the Shop now to learn more.

    Recommended Features

    LOGIN

    LATEST ARTICLES

    NHL Stanley Cup Preview
    We simulate the Stanley Cup Finals 50,000 times to provide each team's probability of hoisting the Cup, and break down the strengths and weaknesses of each team.

    MLB - The So What?
    We use 50,000 simulations of the rest of the MLB season to analyze major personnel changes, including the returns of Josh Donaldson and Troy Tulowitzki to the Blue Jays lineup.

    FIND US ONLINE


    SportsRadar


    All registered accounts are subject to our General Terms Of Service. Content is intended for informational purposes only. This is not a gambling website.

    FSTA Member 2015
    Home | The Predictalator | Blog | Privacy |  Media Contact |  B2B Solutions |  Sitemap |  Customer Support | 
    © 2010-2015 PredictionMachine.com, LLC. All rights reserved. | Site Designed by Randombell.com