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

    Behind the Schedule, Part III (08/05/16)

    By Sean Pyritz @srpyritz
    This week we continue our exploration into the NBA schedule and how it shapes wins throughout the league. In parts one and two, we focused primarily on wins and losses and the different ways to view where they come from and where they go. But wins and losses are merely byproducts of on-court performance, providing little by way of context as to how they occurred. Therefore, in this edition, we will pull back the curtain and reframe our previous analyses using more precise instruments – continuing with the Atlanta Hawks as our test subject.

    Performance Metrics

    For many years, economists and mathematicians alike have been writing papers and crunching the numbers in an attempt to explain and predict what happens any given night in the NBA – with no signs of stopping. However, some of the best measures have been established for over a decade now. We've seen some of these ideas rise above the surface of the ocean in previous articles, but we've yet to identify them properly. Beyond the simple box score, these metrics tell the story of a game – or in our case a season.

    With an assist from the work of the legendary Dean Oliver – I said it once and I'll say it again, check out Basketball on Paper if you would like learn straight from the horse's mouth - we can break the story of NBA performance into two parts: offensive and defensive rating (efficiency); and the “Four Factors.” In the table below, each of the past two Atlanta Hawks seasons are broken into sections for efficiency and the “Four Factors” for offense and defense. Starting with efficiency, the idea is to measure scoring in a controlled manner so we can compare teams that play at different speeds – simple points scored and points allowed do not account for this. Offensive Rating (OR) is the points scored by a team per 100 possessions (close to the average number for an NBA game). Defensive Rating (DR) is the points allowed per 100 possessions. The difference between the two is known as Net Rating – the best predictor of future wins on the market today.

    Where efficiency metrics more appropriately address what happened, the “Four Factors” tackle the question of how it happened. Effective Field Goal Percentage (eFG) is an upgrade over traditional field goal percentage by taking into account the additional value of three pointers. Turnover Rate (TOr) measures the percentage of possessions that end in a turnover – the lower the better. Offensive Rebound Rate (ORBr) calculates the percentage of available offense rebounds captured. Finally, Free Throw Rate (FTr) is the number of free throws per field goal attempt. Taken together, these four capture all the endpoints of a possession and thus success in winning basketball games – made/missed shot, turnover, offensive rebound, or free throw. The interpretation of these factors on defense is reversed – lower eFG/ORBr/FTr and higher TOr are better. In the table below, the optimum value for each metric is bolded.



    The numbers in the table tell a coherent story of the 12-win drop for the Hawks between the past two seasons. Representing the average performance in each season, the table makes it easy to distinguish where the Hawks faltered last year in comparison to 2015. Thanks to a decline in each of the offensive “Four Factors,” the 2016 Hawks saw over four points shaved off their Offensive Rating. In the next two sections, we will use these tools to extend our analysis of previously covered schedule topics – rest advantage and win transfers.

    Rest Advantage

    As we've seen in the previous installments, the distribution of days off in between games on the NBA schedule is not fair – creating opportunities for competitive advantages for more well-rested teams in any given game. The table below breaks down the performance of the Atlanta Hawks over each of the past two seasons in each of the three rest advantage scenarios: Atlanta has more rest, equal rest, and opponent has more rest.



    If there is a competitive advantage in having more rest than the opponent, a given team should perform best when at a rest advantage and worst when at a rest disadvantage. The Net Rating of the 2015 Hawks team in the table above bears this theory out. However, the pattern is reversed in the 2016 season, as the Hawks performed the best with less rest and worst with more rest than their opponent.

    Looking at the “Four Factors,” Atlanta's defense performed best with a rest advantage in both seasons – the Defensive Rating was better in these situations in both years as well. It would follow that if defense performs best versus more tired offense, then offense performs best versus more tired defense. However, this story does not play out in this example. In 2016, the Hawks offense had its best “Four Factor” values with a rest disadvantage - with the exception of eFG, where equal rest produced the highest performance. As a result, the Hawks soiled their opportunities with a rest advantage, accounting for seven more losses between 2015 and 2016.

    Win Transfers

    Last week we introduced the idea of win transfers as a means of explaining the change in wins for a team from season to season. For the Hawks to drop 12 wins off their record, for example, those 12 wins had to go somewhere; and, as we saw, the 12-win drop is a net effect, as the Hawks accumulated and gave away wins in individual matchups from 2015 to 2016.

    In the table below, the performance of the Hawks in the 2015 season is broken into categories. The first category is all 60 wins in order to approximate how the Hawks performed in winning efforts. The second is 2015 wins that turned into losses in 2016. I connected the two based on sequential matchups versus like opponents. For example, the Hawks won their first matchup with the Pistons in 2015, but lost their first game with the Pistons in 2016. Categories three and four are losses in 2015 and the losses from 2015 that became wins in 2016, respectively. The idea is that maybe there is something in the performance one season that hints the team may trade a win for a loss or vice versa in the next season.



    The table shows the Hawks net 12-win drop came about by giving 22 wins away and gaining 10 wins back. Starting with the win categories, the wins that became losses the next season were nearly identical to average wins in 2015, with notable eFG exceptions. The Hawks performed nearly 20 percentage points better on offense and worse on defense in the games they gave away the following season than on the average. It should be noted that Effective Field Goal Percentage is the strongest of the “Four Factors” in terms of predicting future outcomes.

    On the loss side, the samples aren't quite as large, but the defense tells the story. The Hawks performed over four points better in Defensive Rating during games they would win back the following year than in typical losses. In fact, if the Hawks performed on offense like they did in average wins, they would have had a positive net rating and likely won a few more of those games – making them prime targets for win reclamation in the following year.

    Based on this one example, it is impossible to say whether there are signals hidden in a loss/win in one season that might predict a win transfer in the next. Nevertheless, the scent is there, and we will expand upon this topic going forward. It is worth considering that win transfers do not have to be opponent dependent, but could be viewed as simply a factor of the sequence in the season – i.e. game one versus game one in each season and so forth. There are lots of factors to consider here, but the potential we've seen thus far warrants further exploration.

    I find this concept of win transfers eternally fascinating because they get at the heart of how the league changes from year to year. Whether there are explanations hidden in the schedule or signals buried in the box scores from previous seasons will require an expansion beyond a single case study. Now that we are equipped with the tools and lenses to properly analyze these queries, we will expand our analysis to the entire league next week.
    Print This Article
    PREVIOUS ARTICLE
    Behind the Schedule, Part IV (08/12/16)
    NEXT ARTICLE
    Behind the Schedule, Part II (07/29/16)
    Want More Tips and Information? Click To Register!

    COMMENTS

    SEARCH BLOGS

    ARCHIVE

    The Predictalator

    03/27/2017 Highlight: March Madness is winding down and it was another solid month for the Predictalator. There were a total of 69 "normal" or better college hoops positions (sides and totals) in the month of March, and the Predictalator posted a strong 42-27 (60.9%) record, generating $704 in total profits. NBA has also followed suit with "normal" or better positions winning at a 58.1% clip since the beginning of February, and at a 57.9% clip for March.

    Recommended Features

    LOGIN

    LATEST ARTICLES

    MLB Win Totals - Over/Under Picks
    We take a look at the 2017 MLB over/under team season win totals. Highlighted, "normal" or better MLB Win Total Picks are 54-19 all-time. See free picks on the Cubs, Dodgers, Phillies and more.

    Updated Bracket Odds - Final Four
    The Predictalator simulates the Final Four 50,000 times to see how likely each remaining team is to win the NCAA Tournament

    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