The Numbers Game: Projected Meta-Rankings
Friday, October 24, 2014
This column provides ‘meta-rankings’ for the top-200 players in 9-cat and 8-cat leagues, based upon the combined league-wide statistical projections of three prominent and well-regarded NBA sites. If you’re playing fantasy basketball, you’re on at least one of these sites on a daily basis.
Rotoworld is one of them, and our full projections are available in the NBA Draft Guide. Weekly rankings and projections will be available imminently in the Season Pass! The other two sites I’m using for bedrock data are ESPN.com and the ever-impressive Basketball Monster. It’s expected that many owners will vehemently disagree with some of these projections-based rankings, in which case I encourage you to check out the source materials.
Editor’s Note: Rotoworld’s partner FanDuel is hosting a one-day $200,000 Fantasy Basketball league for October 29th’s games. It’s $25 to join and first prize is $20,000. Starts at 7pm ET on October 29th. Here’s the FanDuel link.
Methodology
1) Aggregate league-wide player projections.
2) Assign each player a ‘value’ in 9-cat and 8-cat leagues by using population means, standard deviations and z-scores (I weighted both FG% and FT% by attempts…email me if you want to know more about that).
3) Eliminate any players who fell below the top-200 for each format, then re-calculate means, standard deviations and z-scores to determine new top-200 values.
4) Weight these z-score-based values by each player’s projected number of games played (compared to the population mean, not an 82-game season).
5) For each individual set of projections, rank players based upon these new, weighted top-200 values for 9-cat and 8-cat leagues.
6) Merge these 9-cat and 8-cat rankings and average each player’s ranking to determine a ‘Meta-Rank.’
7) Talk a walk to remind myself what fresh air feels like.
8) Share the results! (Of the analysis, not the walk.)
In case it’s not perfectly clear, these aren’t my personal rankings. They’re not even the official rankings of the sites whose projections I’m analyzing, though any list of projected stats betrays the expectations and biases of the person/people compiling the numbers. By combining multiple such projections, my goal is to arrive at as objective a ranking as I can possibly come up with. *A spreadsheet quirk initially had “Wes Matthews” showing up as “Wes Johnson,” thanks to Richard Cafarelli for quickly pointing that out to me! It’s now updated with the correct name. (Wes Johnson came out ranked 143 in 9-cat and 157 in 8-cat.)
Two more notes: Anyone heading into a draft should also check out Aaron Bruski’s top-150 list, as well as Mike Gallagher’s columns about ‘Abusing the Default Rankings’ for ESPN or Yahoo!. They’re both terrific resources to have you at the top of your game on draft day.
*Some players appeared on one or two projected top-200 lists but not all three, which is why there are more than 200 players for both formats.
How you use this is completely up to you. Jeremy Lin looks like an early-mid round value according to these aggregated projections, but he’s typically available toward the latter half of the middle rounds. Should you reach for him in the 60-70 range, or hope that he falls to you even later? Such decisions can only be made on a case-by-case basis, after careful consideration of your league’s depth, your team’s personnel, your strategy, which weeks are included in your league’s playoffs, and much more. There’s also a bias toward established players, it seems, as many fading veterans are given lenient projections while many young guys are short-shrifted. This is a tool to help you draft the best team possible, not a 1-2-3 cheat-sheet that will tell you how to pick your team.
For a ton of other draft tips, strategies and insights, dig through the archive of my ‘Numbers Game’ columns. And follow me on Twitter for season-long player updates, stats and fantasy advice.
While spending the past few days coaxing Excel spreadsheets into submission, some statistical glitches came to my attention. For instance, the world-wide leader in sports hasn’t provided stat projections for Draymond Green, Rudy Gobert, Solomon Hill, P.J. Tucker, Bojan Bogdanovic, Nikola Mirotic or Doug McDermott. They also have a few guys like Al Harrington listed at 12.3 points, 1.4 triples, 5.4 rebounds, 1.3 assists and 0.8 steals in 24 minutes per game, even though he’s currently playing in China. I deleted him from the analysis, along with some other oddballs (Baron Davis, Jason Kidd, Lester Hudson, Delonte West, etc.). In other words, I did my best to smooth the rough edges without altering the fundamental picture that arose from the data.
Enjoy, and good luck this season. We’re only a few days away!
This column provides ‘meta-rankings’ for the top-200 players in 9-cat and 8-cat leagues, based upon the combined league-wide statistical projections of three prominent and well-regarded NBA sites. If you’re playing fantasy basketball, you’re on at least one of these sites on a daily basis.
Rotoworld is one of them, and our full projections are available in the NBA Draft Guide. Weekly rankings and projections will be available imminently in the Season Pass! The other two sites I’m using for bedrock data are ESPN.com and the ever-impressive Basketball Monster. It’s expected that many owners will vehemently disagree with some of these projections-based rankings, in which case I encourage you to check out the source materials.
Editor’s Note: Rotoworld’s partner FanDuel is hosting a one-day $200,000 Fantasy Basketball league for October 29th’s games. It’s $25 to join and first prize is $20,000. Starts at 7pm ET on October 29th. Here’s the FanDuel link.
Methodology
1) Aggregate league-wide player projections.
2) Assign each player a ‘value’ in 9-cat and 8-cat leagues by using population means, standard deviations and z-scores (I weighted both FG% and FT% by attempts…email me if you want to know more about that).
3) Eliminate any players who fell below the top-200 for each format, then re-calculate means, standard deviations and z-scores to determine new top-200 values.
4) Weight these z-score-based values by each player’s projected number of games played (compared to the population mean, not an 82-game season).
5) For each individual set of projections, rank players based upon these new, weighted top-200 values for 9-cat and 8-cat leagues.
6) Merge these 9-cat and 8-cat rankings and average each player’s ranking to determine a ‘Meta-Rank.’
7) Talk a walk to remind myself what fresh air feels like.
8) Share the results! (Of the analysis, not the walk.)
In case it’s not perfectly clear, these aren’t my personal rankings. They’re not even the official rankings of the sites whose projections I’m analyzing, though any list of projected stats betrays the expectations and biases of the person/people compiling the numbers. By combining multiple such projections, my goal is to arrive at as objective a ranking as I can possibly come up with. *A spreadsheet quirk initially had “Wes Matthews” showing up as “Wes Johnson,” thanks to Richard Cafarelli for quickly pointing that out to me! It’s now updated with the correct name. (Wes Johnson came out ranked 143 in 9-cat and 157 in 8-cat.)
Two more notes: Anyone heading into a draft should also check out Aaron Bruski’s top-150 list, as well as Mike Gallagher’s columns about ‘Abusing the Default Rankings’ for ESPN or Yahoo!. They’re both terrific resources to have you at the top of your game on draft day.
*Some players appeared on one or two projected top-200 lists but not all three, which is why there are more than 200 players for both formats.
How you use this is completely up to you. Jeremy Lin looks like an early-mid round value according to these aggregated projections, but he’s typically available toward the latter half of the middle rounds. Should you reach for him in the 60-70 range, or hope that he falls to you even later? Such decisions can only be made on a case-by-case basis, after careful consideration of your league’s depth, your team’s personnel, your strategy, which weeks are included in your league’s playoffs, and much more. There’s also a bias toward established players, it seems, as many fading veterans are given lenient projections while many young guys are short-shrifted. This is a tool to help you draft the best team possible, not a 1-2-3 cheat-sheet that will tell you how to pick your team.
For a ton of other draft tips, strategies and insights, dig through the archive of my ‘Numbers Game’ columns. And follow me on Twitter for season-long player updates, stats and fantasy advice.
While spending the past few days coaxing Excel spreadsheets into submission, some statistical glitches came to my attention. For instance, the world-wide leader in sports hasn’t provided stat projections for Draymond Green, Rudy Gobert, Solomon Hill, P.J. Tucker, Bojan Bogdanovic, Nikola Mirotic or Doug McDermott. They also have a few guys like Al Harrington listed at 12.3 points, 1.4 triples, 5.4 rebounds, 1.3 assists and 0.8 steals in 24 minutes per game, even though he’s currently playing in China. I deleted him from the analysis, along with some other oddballs (Baron Davis, Jason Kidd, Lester Hudson, Delonte West, etc.). In other words, I did my best to smooth the rough edges without altering the fundamental picture that arose from the data.
Enjoy, and good luck this season. We’re only a few days away!
This entry passed through the Full-Text RSS service – if this is your content and you’re reading it on someone else’s site, please read the FAQ at fivefilters.org/content-only/faq.php#publishers.
