Difference between revisions of "SIBR:Sibrmetrics"

From Blaseball Wiki

Line 17: Line 17:
 
<br>
 
<br>
 
<big>Offensive WhAT</big><br>
 
<big>Offensive WhAT</big><br>
 
+
(BATR + BsR + Replacement Runs)/Runs per win<br>
 
A core component of WhAT is the use of linear weights to convert counts of performance (Home Runs, walks, stolen bases, etc.) into a run value. A linear weight is effectively the average of how many runs occur in an inning after a given game action is performed. Hitting value for WhAT is measured in Batting Runs (BATR). Batting Runs utilize linear weights to determine how many runs a players performance was worth relative to league average performance given the same amount of plate appearances. Further, Batting Runs are adjusted based on the parks and leagues the player played in. Baserunning (BsR) is similar, assigning runs relative to a league average performer given the situations the player was in on the base paths. The main components are value from stolen bases (wSB), avoiding double plays (wGDP), and advancing on balls in play while on base (UBR). The players share of replacement runs are credited per plate appearance.<br>
 
A core component of WhAT is the use of linear weights to convert counts of performance (Home Runs, walks, stolen bases, etc.) into a run value. A linear weight is effectively the average of how many runs occur in an inning after a given game action is performed. Hitting value for WhAT is measured in Batting Runs (BATR). Batting Runs utilize linear weights to determine how many runs a players performance was worth relative to league average performance given the same amount of plate appearances. Further, Batting Runs are adjusted based on the parks and leagues the player played in. Baserunning (BsR) is similar, assigning runs relative to a league average performer given the situations the player was in on the base paths. The main components are value from stolen bases (wSB), avoiding double plays (wGDP), and advancing on balls in play while on base (UBR). The players share of replacement runs are credited per plate appearance.<br>
  
Offensive WhAT = (BATR + BsR + Replacement Runs)/Runs per win
 
 
<br>
 
<br>
  
 
<big>Pitching WhAT</big><br>
 
<big>Pitching WhAT</big><br>
 +
((IP*(RA9opp-(9*DRiP/IP)-(RA9/park factor))/9)+(Outs*Replacement Runs/Out))/dynamic Runs per Win<br>
  
 
The primary aspect of pitching WhAT is simple, how did the pitcher's average runs allowed (RA9) compare to the average runs scored by the teams they faced(RA9opp)? Unfortunately for your favorite internet statistician, Defense Really impacts Performance(DRiP), and needs to be accounted for. The value of Defensive Runs implicitly Prevented (DRiP) is estimated and then subtracted from the pitcher's average opponent to effectively create the ERA of an average pitcher with an average fielding team against the opponents (RA9avg). Linear weights return to help us estimate the value of fielding that the pitcher received. First Baserunning defense is accounted for by measuring the value of stolen bases and extra advancement that occurred while the pitcher was on the mound (this is fBsR, double plays are not included due to pitcher shakespearianism being considered the driving factor). Next the run value of balls in play allowed (wBIPA) is taken into account. Because pitcher and fielders are both responsible for this statistic, a composite between the team's overall wBIPA and the wBIPA when the pitcher is on the mound is created. This either shrinks or expands the amount of runs credited for the fielding based on whether the fielding was better or worse than when the given pitcher was on the mound (ex: a pitcher that gets more ground/fly outs than team average will get relatively more run prevention attributed to themselves).  Similarly to overall run scoring, the of the average opponents ball in play value (wBIPAopp) is compared to the composite value. The fBsR and BIPruns saved are added to create DRiP, and along with park factor, adjust the run environment the pitcher faced accordingly. Finally because pitchers more directly influence their run environment through their performance, the runs per win calculation is dynamically calculated based on innings pitched per game and how many runs are allowed (this effectively helps good pitchers and hurts bad ones).<br>
 
The primary aspect of pitching WhAT is simple, how did the pitcher's average runs allowed (RA9) compare to the average runs scored by the teams they faced(RA9opp)? Unfortunately for your favorite internet statistician, Defense Really impacts Performance(DRiP), and needs to be accounted for. The value of Defensive Runs implicitly Prevented (DRiP) is estimated and then subtracted from the pitcher's average opponent to effectively create the ERA of an average pitcher with an average fielding team against the opponents (RA9avg). Linear weights return to help us estimate the value of fielding that the pitcher received. First Baserunning defense is accounted for by measuring the value of stolen bases and extra advancement that occurred while the pitcher was on the mound (this is fBsR, double plays are not included due to pitcher shakespearianism being considered the driving factor). Next the run value of balls in play allowed (wBIPA) is taken into account. Because pitcher and fielders are both responsible for this statistic, a composite between the team's overall wBIPA and the wBIPA when the pitcher is on the mound is created. This either shrinks or expands the amount of runs credited for the fielding based on whether the fielding was better or worse than when the given pitcher was on the mound (ex: a pitcher that gets more ground/fly outs than team average will get relatively more run prevention attributed to themselves).  Similarly to overall run scoring, the of the average opponents ball in play value (wBIPAopp) is compared to the composite value. The fBsR and BIPruns saved are added to create DRiP, and along with park factor, adjust the run environment the pitcher faced accordingly. Finally because pitchers more directly influence their run environment through their performance, the runs per win calculation is dynamically calculated based on innings pitched per game and how many runs are allowed (this effectively helps good pitchers and hurts bad ones).<br>
 
Pitching WhAT = ((IP*(RA9opp-(9*DRiP/IP)-(RA9/park factor))/9)+(Outs*Replacement Runs/Out))/dynamic Runs per Win
 

Revision as of 23:41, 22 January 2022

Under Construction
This page or section is being initially created, or is in the process of an expansion or major restructuring. You are welcome to assist in its construction by editing it as well.

Sibrmetrics are statistics that have been created to help evaluate player performance in Blaseball.

Below are a list of statistics and their definitions that are tabulated in the Compendium of Blaseball Advanced Statistics.

WhAT

Wins (historical) Above Tokyo Lift, a wins above replacement (WAR) analog for blaseball. This statistic estimates the cumulative win value of a player's performance on offense or from pitching above that of a replacement player. For offense, batting performance and baserunning performance relative to league average are added up along with the seasons' value of the average player above replacement, then the value is converted into wins based on a formula. For pitching, the pitcher's runs allowed are compared to the expected average runs allowed based on opponents and their defense. This run value is then added to the average above replacement value and then converted into wins.

To generate the value of the average above replacement for a given season of blaseball, we use the season 11 Tokyo Lift as a replacement team, and model how many wins fewer than average they would have won. A model (negative binomial) is fitted to the relationship between the strength of a team's roster in Stars and it's win total for each season, the historical season 11 roster of the Tokyo Lift is inputted into that model to predict their win total. The wins below average are then credited as the win value of the average player in that season above replacement. In this way the amount of replacement wins increased over the course of blaseball's history(starting at 0 when all teams were around replacement level) .


Offensive WhAT
(BATR + BsR + Replacement Runs)/Runs per win
A core component of WhAT is the use of linear weights to convert counts of performance (Home Runs, walks, stolen bases, etc.) into a run value. A linear weight is effectively the average of how many runs occur in an inning after a given game action is performed. Hitting value for WhAT is measured in Batting Runs (BATR). Batting Runs utilize linear weights to determine how many runs a players performance was worth relative to league average performance given the same amount of plate appearances. Further, Batting Runs are adjusted based on the parks and leagues the player played in. Baserunning (BsR) is similar, assigning runs relative to a league average performer given the situations the player was in on the base paths. The main components are value from stolen bases (wSB), avoiding double plays (wGDP), and advancing on balls in play while on base (UBR). The players share of replacement runs are credited per plate appearance.


Pitching WhAT
((IP*(RA9opp-(9*DRiP/IP)-(RA9/park factor))/9)+(Outs*Replacement Runs/Out))/dynamic Runs per Win

The primary aspect of pitching WhAT is simple, how did the pitcher's average runs allowed (RA9) compare to the average runs scored by the teams they faced(RA9opp)? Unfortunately for your favorite internet statistician, Defense Really impacts Performance(DRiP), and needs to be accounted for. The value of Defensive Runs implicitly Prevented (DRiP) is estimated and then subtracted from the pitcher's average opponent to effectively create the ERA of an average pitcher with an average fielding team against the opponents (RA9avg). Linear weights return to help us estimate the value of fielding that the pitcher received. First Baserunning defense is accounted for by measuring the value of stolen bases and extra advancement that occurred while the pitcher was on the mound (this is fBsR, double plays are not included due to pitcher shakespearianism being considered the driving factor). Next the run value of balls in play allowed (wBIPA) is taken into account. Because pitcher and fielders are both responsible for this statistic, a composite between the team's overall wBIPA and the wBIPA when the pitcher is on the mound is created. This either shrinks or expands the amount of runs credited for the fielding based on whether the fielding was better or worse than when the given pitcher was on the mound (ex: a pitcher that gets more ground/fly outs than team average will get relatively more run prevention attributed to themselves). Similarly to overall run scoring, the of the average opponents ball in play value (wBIPAopp) is compared to the composite value. The fBsR and BIPruns saved are added to create DRiP, and along with park factor, adjust the run environment the pitcher faced accordingly. Finally because pitchers more directly influence their run environment through their performance, the runs per win calculation is dynamically calculated based on innings pitched per game and how many runs are allowed (this effectively helps good pitchers and hurts bad ones).