Week
13
Performance
Proportional Ranking allows a view of the players according to the performance of the pack. If the pack has 20 players, then the ranks go from 1 to 20, but each player can have rank anywhere in between. Players that perform similarly according to the metric, with have closer ranks, ie. 2 players with rank of 7.1 and 7.2 from pack of 28 players means both players perform similarly and are 1/4 between top and last ranked player. Packs is this case are the top players for the position, enough to occupy starting and backup positions for a team in my league of 14 teams
Below, shows charting of 4 ranking methods: By 1) season long position winning pct 2) season long point average 3) later games have higher weighting position winning pct 4) later games have higher weighting points average, each compared how they fared the latest week.
<IMG Reading data…
[SeasonWins.png]
113951 bytes
Error : The remote name could not be resolved: 'beyondtoiletbowl.wordpress.com'
at System.Net.HttpWebRequest.GetRequestStream(TransportContext& context)
at System.Net.HttpWebRequest.GetRequestStream()
at CookComputing.XmlRpc.XmlRpcClientProtocol.Invoke(Object clientObj, MethodInfo mi, Object[] parameters) in C:\Users\Bob\Downloads\USB30\Projects\FantasyFootball\AddImageToWordpress\XmlRpcClientProtocol.cs:line 148
at CookComputing.XmlRpc.XmlRpcClientProtocol.Invoke(MethodInfo mi, Object[] Parameters) in C:\Users\Bob\Downloads\USB30\Projects\FantasyFootball\AddImageToWordpress\XmlRpcClientProtocol.cs:line 103
at XmlRpcProxy8b5cc608-1e68-4d1d-b7c0-ca8dadcef8be.newImage(String blogid, String username, String password, Image theImage, Boolean overwrite)
at AddImageToWordpress.Program.Main(String[] args) in C:\Users\Bob\Downloads\USB30\Projects\FantasyFootball\AddImageToWordpress\Program.cs:line 109
How to look at it: Closer to left means top ranked at position, closer to bottom is how they performed last week relative to top ranked last week. Players above the dashed line did worse than expected. Players below the dashed line did better than expected.
<IMG Reading data…
[SeasonPoints.png]
114564 bytes
Error : The remote name could not be resolved: 'beyondtoiletbowl.wordpress.com'
at System.Net.HttpWebRequest.GetRequestStream(TransportContext& context)
at System.Net.HttpWebRequest.GetRequestStream()
at CookComputing.XmlRpc.XmlRpcClientProtocol.Invoke(Object clientObj, MethodInfo mi, Object[] parameters) in C:\Users\Bob\Downloads\USB30\Projects\FantasyFootball\AddImageToWordpress\XmlRpcClientProtocol.cs:line 148
at CookComputing.XmlRpc.XmlRpcClientProtocol.Invoke(MethodInfo mi, Object[] Parameters) in C:\Users\Bob\Downloads\USB30\Projects\FantasyFootball\AddImageToWordpress\XmlRpcClientProtocol.cs:line 103
at XmlRpcProxy880f1c38-c0f7-4b17-92bc-3056714a3cb9.newImage(String blogid, String username, String password, Image theImage, Boolean overwrite)
at AddImageToWordpress.Program.Main(String[] args) in C:\Users\Bob\Downloads\USB30\Projects\FantasyFootball\AddImageToWordpress\Program.cs:line 109
How to look at it: (See above)
How to look at it: (See above)
How to look at it: (See above)
Below is a comparison of how the player’s season average(x-axis), against their last week’s performance(points).
How to look at it: X-axis is the season average points per game by the player, y-axis is points they scored last week. Opposite of above charts, players above the dashed line performed better than their average. Below the line means they performed worse.
Below is a comparison of how the player’s late game weighted, season average(x-axis), against their last week’s performance(points).
How to look at it: (See above)
Below, shows charting of how our 4 ranking methods performed, how many differences to their predicted rank (are they who we think they are, rewording Cardinal’s coach, Denny Green): By 1) season long position winning pct 2) season long point average 3) later games have higher weighting position winning pct 4) later games have higher weighting points average, putting them into groups of how they performed according to their rank. We don’t have negative bins. Every player that did better, has to result in a player that did worse using the ranking system. Plus there no going up from rank #1. So players that do slightly better, are in same group as players who do slightly worse, the 0..4 bin. Plus here is, if the horde all decides to be better at same time b/c the weather is nice that sunday, it reflects as everyone did better using points, but everyone did mostly the same using rankings.
How to look at it: (See above)
How did the defensive handicaps do? Below are 3 charts: 1) season average (x) against last game pts(y) 2) forecasted pts using defense handicap (x) against last game points(y) 3) The amount of point difference from average predicted by our defense handicap(x) against actual point difference from average and season average(y). What we are looking for is better correlation from the 1st chart, to the 2nd. And the 3rd chart, we hope to see more points closer to line y=x, but in the past, the effects of a defense aren’t that stable from play to play or game to game. There are large random effects observed, as seen from the large y-range.
Below is the distribution of our difference (or deviation or delta) from the average, and the Forecast with a defensive handicap applied. If our handicap was effective, there should be an observable number of predictions in the bins closer to zero for our handicap, compared to the average. Keep your expectations in check, however, and remember that 20 is only 2 points for a player averaging 10pts a game.
Look below for view of players that are worse or better. If the defense coefficient was major factor in predicting how the majority of players (market) will move from week to week, it’s predictions will be more balanced than using the average.
Look below to see the middle 40, where some sort of windfall or unstability didnt result in some huge, or dreadful performance.
How to look at it: More players in bins closer to zero means our predictions are more precise. But as long as it’s even on both sides of 0, we are in the ballpark
Do players feast on bad defenses? Or is it a famine? Role of Linear Regression in a handicap.
beyondtoiletbowl 