BuzzerBeater Forums

He has a tough team we will see.

Rankings Reposted, with certain visual modifications

First Period: http://dl.dropbox.com/u/1703053/BuzzerBeater/Season_13_To...

Second Period: http://dl.dropbox.com/u/1703053/BuzzerBeater/Season_13_To...

Third Period: http://dl.dropbox.com/u/1703053/BuzzerBeater/Season_13_To...


I've been developing a new ranking system that does requires less work (thankfully) and actually does creates a more accurate representation of measuring player performance over the course of the season. It's not completely functional yet but I hope to have post these numbers at the end of the season. Everything I have seen from this overhaul of the ranking system actually demonstrates better balance and a realistic distribution of points.

To give a general idea, I've figured out a method that takes into consideration a player's Z-Score when compared against the population and awards points based on their Z-Score. I am still valuing the 14 Conference matchups at a max of 4 points per game that a player can earn while valuing the 8 Non-Conference matchups at a max of 7 points per game.

The Z-Score is a representation of where a player's HGS falls on the normal bell curve. When studying a Normal Bell Curve, 99.96% of a statistical population falls between both ends of the bell curve. This is representing by a max value of -3.00 for all players that fell below the mean (average) and +3.00 for all players that fall above the mean (average). The middle of the bell curve is represented by 0.00. Basically, a Z-Score can fall anywhere between -3.00 and +3.00. Players that play extradorinary when compared to the rest of the league can score higher than +3.00 whereas players that play downright horrible can achieve a Z-Score below -3.00. Keep in mind that only 0.02% of a population will score above or below the minimum/maximum. A good example for this season was the first week when Clay Chow (prior to being sold) recorded a Z-Score 4.40, a whopping 1.4 greater than what your greatest players tend to score.

Unfortunately, Gavno sold Chow and we've been subject to the likes of Sean Ryan and Mario Mannu obliterating us.

Last edited by kaygdanimal at 9/1/2010 10:49:31 PM

What I don't get is how, even with this volume of data, you've only managed to beat the Magic once in a competitive match. Zing!

What I don't get is how, even with this volume of data, you've only managed to beat the Magic once in a competitive match. Zing!

Such a great "ziiiiinnnnnnng!" that you had to rub it in for good measure?

Don't you dare sit there and think for a second that I haven't thought of that myself, lol. I use it myself as a way to measure my players' performance compared to the rest of the league to make sure I am playing within my limits ... or determine if I need to upgrade.

Francisco Candia was a result of these rankings. Cantrell is sitting on the fringe of possibly moving, either to SG or to another team ... as was Lowell Webber. However, they locked down right now thanks to discovering I completely overpaid for Galvao Veiga Paz as he does not suit this quartet. Right now, I need players that compliment Candia and Partridge or otherwise I'm going to have to move both players and start over.

all i know if you need to do something because we got rocked in interleague play

... and it showed in my current player rankings model as well, lol. We had key players throughout the set of rankings but as a whole, the Big 8 rocked the Great 8. Their player performance rankings are spread evenly while ours are grouped in bunches it seems.

Final Rankings are uploaded. These are the last of this version. I have developed a brand new version that I will explain the process later, but to give a bit of a nugget, it utilizes a bell curve to measure performances. At the same time, it does away with normalizing scores based on time due to a correlation study that I performed over several seasons worth of scores that proved minutes played has very little correlation to Hollinger Game Scores being higher when longer minutes are played. Instead, I have turned the focus towards the computed HGS and limited the qualifiers, for this first set of ratings, to players that played a minimum of 20 minutes for a game. What that means is that a player must play 20 minutes in a game for their HGS to be considered in the ratings. Next season, I will be limiting minutes to 16 minutes. I have found, however, that a higher level of minutes will help keep a player near the top of the ratings.

With that said, the first file is the final rankings using my previous method (stays standard with the set of rankings I have used all season). The second file is a taste of what next seasons rankings will tend to look like, including utilizing the use of giving a BuzzerBeater rating from Legendary to Atrocious. I'm hoping to give it more of a "Buzzerbeater"-like feel.

http://dl.dropbox.com/u/1703053/BuzzerBeater/Season_13_To...

http://dl.dropbox.com/u/1703053/BuzzerBeater/League_III.4...

awesome work, and i love seeing my boys way at the top. thanks for doing this, sorry i couldnt continue my match day recaps

I am speechless! great work!