My Latest Project

Several years ago I put together a project in which I ranked every player in NBA history statistically, and for my work on that project and the resulting formula and presentation I was awarded First Place at the BYU-Idaho Undergraduate Research Conference. But even with the success of that formula and the recognition I earned, I have always felt that it was slightly flawed. Over the past year I have been developing a revision of my player ranking formula, and my MVP rankings this year were based loosely on that new formula.

In the past few weeks I feel I have finally found the final piece to the puzzle, a way to make make winning count more naturally in favor of each player. I used to use winning percentage as a multiplier in order to make winning a part of the formula, but overall winning percentage means nothing if you lose in the first round of the playoffs (like this year's Spurs). That's when I realized that the best way to give players credit for winning when it matters is to add in their playoff statistics as "extra credit," so that the longer a player keeps his team in the playoffs, the better his overall stats look. I also decided to count missed games as 0's, since you are not helping your team if you are not on the court, no matter the reason.

Another twist that is new to this formula is that different positions are judged differently. All players were sorted by position, then averages for each statistical category were calculated by position. Guards were given more credit for doing well in "big man" categories, and vice versa. This includes shooting percentages, where players were given bonus points for shooting above their positional average, or docked points for shooting below the average. I took each season separately, and gave players points based on how far they performed above the 90th percentile in each season. These points were added up and totaled to rank each player in history.

The result is that players are awarded for performance, winning, and longevity, and players who peaked for a short time (like Bill Walton) don't quite do as well. I plan to start revealing the results of this project countdown-style in July, so keep an eye out.

Predictions - 18 Jun

Most Likely MLB Blowouts

Diamondbacks over White Sox

Red Sox over Brewers

Dodgers over Astros

Reds over Blue Jays

Twins over Padres

Most Likely MLB Upsets

Athletics over Giants

Angels over Mets

Prediction Results - 16 Jun

MLB Blowouts: 1-2 (.333) 16-12 overall (.571)

MLB Upsets: 2-0 (1.000) 11-9 overall (.550)