As part of a new series at For The Win, we are examining who should win major NBA awards based on what we can learn from advanced analytics.
Of course, numbers aren’t the only thing we should consider when making these predictions. But they do remove some bias we have about narratives, and while using one stat may not tell the whole story, various formulas that yield similar results can provide a notable context in these decisions.
We used a methodology originally suggested by Owen Phillips, who now works in analytics for the Knicks.
The first step was finding which metrics to use. For this survey, we included the metrics deemed as the most trustworthy by NBA executives when asked by HoopsHype. (We’ve included more information on where you can find each metric at the bottom of the page.)
Based on a tip from another individual who works in a front office for an NBA team, because each of these metrics is graded on different scales, we adjusted for playing time by multiplying their impact contribution on each metric by the percentage of possible minutes they have played for their team so far this season.
We then standardized each score by finding the Z-score for each player in each metric. The Z-score measures how many standard deviations from the mean for each data point. If a Z-score is 0, that means the player is exactly league average.
As the last step, we took the average Z-score that a player had across each metric. Because these metrics are not perfect, however, we removed the outliers of the highest overall Z-score and the lowest overall Z-score. Finally, this gave us our results you will find below.
If all of this sounds like a lot of numbers, don’t worry! Just take a look at the “score” for a clean metric that summarizes overall productivity.
For comparison, we’ve also included the latest betting odds for each award. Hopefully, the advanced metrics can provide context as you consider which future bets to place.
NOTE: The metrics pulled directly from the HoopsHype survey included Daily Plus-Minus (DPM), Estimated Plus-Minus (EPM), RAPTOR (FiveThirtyEight), Real Plus-Minus (ESPN), Box Plus-Minus (Basketball-Reference), Player Efficiency Rating (Basketball-Reference), Win Shares (Basketball-Reference), kWPA (Inpredictable).
We also added the model of Box Plus-Minus from Backpicks.com, Daily-Updated Rating of Individual Performance (DRIP), Stable Player Impact (SPI), and Global Rating (HoopsHype).
LEBRON (BBall-Index) was not included because the data has not been published yet this season. However, if it’s publicly available by our next update, LEBRON will also be calculated.