You are here > Home > Primate Studies > Discussion
 
Primate Studies — Where BTF's Members Investigate the Grand Old Game Monday, August 19, 2002Win Values: A New Method to Evaluate Starting Pitchers  Part 4 Part
1: Introduction FormulasIn this section I will present the formulas underlying the Win Values system.? There are actually only a few formulas.? Each formula is conceptually simple but may look complex.? We seek the formula for a starting pitcher?s Win Value for a game in which he pitched Z innings,[1] the score at the conclusion of the Zth inning was RS to RA,[2] and the game was played in a ballpark with a park factor of PF.? Let me write this as WinVal(RS,RA,Z,PF). From the conceptual framework described above, we know that Win Value is the difference between the team?s expected probability of winning a game given that the score is RS to RA at the conclusion of the Zth inning, in a PF ballpark, and the team?s expected probability of winning a game given that the team has scored RS runs at the conclusion of the Zth inning, with league average pitching, in a PF ballpark. Let me write the first win probability as WinProb1(RS,RA,Z,PF) and the second win probability as WinProb2(RS,Z,PF).? WinProb1 uses the pitcher?s actual RA information whereas WinProb2 assumes average pitching.? So we have: [Eq.1]?? WinVal(RS,RA,Z,PF) = WinProb1(RS,RA,Z,PF) ? WinProb2(RS,Z,PF).? Let?s first turn to how park effects are handled.? It turns out to require a couple of simplifying assumptions since there are not enough games played in any leagueseason with the same park factor to estimate WinProb1 and WinProb2 for different PF?s. It won?t be immediately clear why we would want to do so, but let?s rewrite Equation 1 as follows: [Eq.2]?? WinVal(RS,RA,Z,PF) = WinProb1(RS,RA,Z,PN) ? WinProb2(RS,Z,PN) ??????????????????????????????????????????????? ??? + [WinProb1(RS,RA,Z,PF) ? WinProb1(RS,RA,Z,PN)] ??????????????????????????????????????????????? ??? + [WinProb2(RS,Z,PN) ? WinProb2(RS,Z,PF)] where PN denotes a park neutral setting.? The reason why we write the equation this way is so that we can consider each of the terms in brackets.? What do each of these terms represent?? The first term reflects how the win probability of a team that has scored RS runs and allowed RA runs at the conclusion of the Zth inning is affected by the park factor.? The second term reflects how the win probability of a team that has scored RS runs at the conclusion of the Zth inning with average pitching is affected by the park factor. Empirically, I have found that the first term is typically small and can safely be ignored.? Getting slightly ahead of ourselves, I have found that WinProb1 is reflected in the game?s deficit (RARS) rather than having to consider RS and RA separately.? By analyzing the inningbyinning runs scored distributions at various parks, I have found that the probability that a team overcomes a deficit of a given size does not depend significantly upon the park factor.[3]? On the other hand, I have found that the second term is potentially significant, and merits special treatment.? Dropping the first term in brackets in Equation 2, and defining a new variable, we have: [Eq.3]?? WinVal(RS,RA,Z,PF) = WinProb1(RS,RA,Z) ? WinProb2(RS,Z) ??????????????????????????????????????????????? ??? + ParkAdder(RS,Z,PF) where, for convenience, I have dropped the PN labels in WinProb1 and WinProb2, and where ParkAdder(RS,Z,PF) = WinProb2(RS,Z,PN) ? WinProb2(RS,Z,PF). I will now describe the formulas for each of these three terms.? WinProb1 involves two concepts.? The first is how RS and RA interact in the formula, and the second is how we ?smear? the run support probabilities. As described above, WinProb1 will be reflected in the deficit that the team faces, RARS, rather than a different formula for every (RS,RA) pair.? There are simply not enough games that have the same score in any leagueseason, and fortunately, I have found empirically that the win probability of overcoming a given deficit does not significantly depend upon the actual score.? For example, the probability that a team trailing 63 comes back to win the game is similar to the probability that a team trailing 85, say, will come back to win the game.[4] We have previously motivated the notion of ?could have been? runs scored possibilities for a pitcher?s run support.? The idea is that a pitcher should be evaluated based not only on how many runs he allowed but also on how many runs he received that game in run support.? However, we do not want to be fanatical about fixing his run support.? This would lead to evaluations that I am not in favor of.[5]? Thus, we seek a middle ground, and this is where the ?could have been? run support smeared probabilities come in. Putting these two ideas together, then, we have: [Eq.4]?? WinProb1(RS,RA,Z) =? å_{m}Smear(m;RS,Z) * DWin(RAm;Z) where the summation is over m, the runs support possibilities (m goes from 0 to 25, say), Smear(•) is the ?could have been? run support probabilities and DWin(?) is the probability that a team trailing by a specific number of runs will come back to win the game.? The DWin probabilities can be derived empirically for each inning Z (from 1 to 9) using all the inningbyinning scoring data from all games in the leagueseason under study.[6]? Note that I smooth the DWin probabilities to remove any effect of small samples or weird games.? That is, I ensure that the probability of winning increases if the lead increases (holding the inning constant), and I ensure the probability of overcoming a deficit decreases as the inning increases (holding the deficit constant). For the derivation of the Smear probabilities, we will need two new variables.? Let R(x;Z) denote the probability that a team scores x runs at the conclusion of the Zth inning.? Let S(w,y;Z,C) denote the probability that a team that has scored y runs at the conclusion of the Cth inning will score w runs at the conclusion of the Zth inning.? For our purposes, C will be less than or equal to Z.? For example, we may be interested in knowing the ?propogation? probabilities of a team that has scored 4 runs at the conclusion of the 6th inning.? That is, how likely is such a team to wind up with 7 runs at the conclusion of the 8th inning, say. I will not repeat the description of the derivation of these smearing probabilities via backwards Bayesian bootstrapping.? The interested reader should revisit the derivation described above in the section on the system?s conceptual framework.? Suffice it to say here that the backwards Bayesian bootstrapping method requires me to designate how far back in a game to go for the ?could have been? phenomenon to kick in. Remember that Wolverton?s system essentially says to ignore the game?s actual run support; this is equivalent to pushing the ?could have been? experiment all the way back to the beginning of the game.? I have described why I do not want to follow that approach.[7]? The further back towards the beginning of the game I choose for my ?could have been? cutoff, the more my method looks like Wolverton?s supportneutral method.? The closer to the end of the game I choose for my ?could have been? cutoff, the more my method reflects the pitcher?s WL record.? I have experimented with different cutoffs and have reviewed the distribution of runs scored by inning in great detail.? I have settled upon the following ?could have been? inning cutoffs: 6th inning for a 9inning outing, 5th inning for a 7 or 8inning outing, 4th inning for a 5 or 6inning outing, 3rd inning for a 3 or 4inning outing, 2nd inning for a 2inning outing, and 1st inning for a 1inning outing. To save on some notation, let C denote these innings to which we allow the ?could have been? smearing to begin.? C will depend upon Z, but I will suppress that in the formula below. By backwards Bayesian bootstrapping, we have: [Eq.5] Smear(m;RS,Z) = å_{n}{[R(n;C) * S(RS,n;Z,C)] / å_{j }[R(j;C) * S(RS,j;Z,C)]} * S(m,n;Z,C) where the two summations, over n and j respectively, go from 0 to 25, say.? We can derive all the required R and S probability distributions empirically for each inning Z (1 to 9) using all the inningbyinning scoring data from all games in the entire leagueseason under study.? Therefore, we are able to derive the required WinProb1 probabilities. Let?s return to the second term of Equation 3, WinProb2(RS,Z).? Remember this is the probability that a team that scores RS runs in Z innings will win the game with average pitching.? There are two elements to consider.? First, for any run scored we estimate the probability that a team scoring that many runs at the conclusion of the Zth inning will win the game with league average pitching.? Second, as above, we smear the runs scored probabilities based upon RS and the ?could have been? smearing probabilities.? Thus, we have: [Eq.6]?? WinProb2(RS,Z) = å_{m} Smear(m;RS,Z) * AWin(m;Z) where the summation is over m, the run support possibilities (m goes from 0 to 25, say), and AWin is the probability that a team that scores a specific number of runs in Z innings will win the game with league average pitching.? Equation 5 previously gave the formula for Smear(•), so all we need is AWin(•). AWin is estimated empirically for each inning Z (1 to 9) using all the inningbyinning scoring data from all the games played in the leagueseason under study.[8]? Note that I smooth the AWin probabilities to remove any effects of small samples or weird games.? That is, I ensure that the probability of winning increases if more runs are scored (holding the inning constant).? I also ensure that the win probability when scoring any number of runs at the conclusion of Z innings decreases with Z (holding the number of runs constant). The remaining term of Equation 3 is the Park Adder.? There is insufficient data to estimate a separate park adder for every possible park factor and every possible run scored.? I therefore pool the park data of each leagueseason into three groups: hitters parks, neutral parks, and pitchers parks.? This allows me to estimate the park adder as a percentage of the park factor of the home park for each of the possible run scored figures (say from 0 to 25).? I then smooth these park adders to remove any effect of small samples or weird games. An example will help here.? Suppose that I find that the home park affects the probability of winning a game when scoring exactly 5 runs by .009 per percentage point of the home park?s park factor.? For concreteness, let?s use the Oakland Coliseum in 2000 which had a Total Baseball Park Factor of 97.? This implies that runs were 6% less prevalent in games at the Oakland Coliseum compared to a league neutral park.? Multiplying the 6 by the .009 yields an estimate that a team scoring 5 runs at Oakland Coliseum had a .054 higher chance of winning the game than a team that scored 5 runs in a league neutral park. The last step to deriving the Park Adder is to prorate the change in win probability by the number of innings the pitcher pitched in the game.? Algebraically, then, we have: [Eq.7]?? ParkAdder(RS,Z,PF) = PAddPct(RS) * (2*(100PF)) * (Z/9)^{+} where PAddPct(RS) is the per percentage change in the probability of winning with RS runs (in the above example this was the .009).? The middle term reflects the effect of the ballpark on runs scored and the fact that only half a team?s games are played at home (in the example above PF was 97, so that the middle term is 6).? (Z/9)^{+} is Z/9 capped at 1 to properly handle partial games and pitchers who pitch more than 9 innings. Finally, we have completed our formulaic journey.? Equations 37 provide the formulas underlying the Win Values system.? Since formulas can be rather dry and imposing, I next turn to giving numerical examples of all of the terms appearing in these formulas. [1]? Since my Win Values system relies upon the inningbyinning runs scored distributions, the system treats partial innings pitched, including facing one or more batters in an inning without recording an out, as a complete inning.? In addition, no distinction is made between earned runs and unearned runs. [2]? RS is his team?s runs scored and RA is how many runs he allowed.? The evaluation is based solely upon the score at the conclusion of the last inning the pitcher appeared in.? Accordingly, the performance of the pitchers who relieve the starter is not considered germane to the evaluation of the starter?s performance. [3]? Mathematically, this would be considered a secondorder term. [4]? Let me give a representative example.? Using the AL 2000 season, when leading after 7 innings by one run, the probabilities of winning the game based upon the score are as follows.? 73% if 10; 82% if 21; 80% if 32; 84% if 43; 71% if 54; and 72% for combined 65, 76, 87, etc., where this last grouping is necessary to achieve a sufficient number of games to make the comparison valid. [5]? Remember, if we ?fix? run support at RS, then a pitcher who wins a game 62 would be deemed to have the same win probability as a pitcher who wins a game 65. [6]? Interleague games are included where appropriate as follows.? For distributions of runs scored, runs scored by a team in the league under study are included.? For distributions of runs allowed, runs scored by the opponent of a team in the league under study are included. [7]? See footnote 4. [8]? See footnote 12

BookmarksYou must be logged in to view your Bookmarks. Hot TopicsWhat do you do with Deacon White?
(17  1:12pm, Dec 23) Last: Alex King Loser Scores (15  12:05am, Oct 18) Last: mkt42 Nine (Year) Men Out: Free El Duque! (67  10:46am, May 09) Last: DanG Who is Shyam Das? (4  8:52pm, Feb 23) Last: RoyalsRetro (AG#1F) Greg Spira, RIP (45  10:22pm, Jan 09) Last: Jonathan Spira Northern California Symposium on Statistics and Operations Research in Sports, October 16, 2010 (5  12:50am, Sep 18) Last: balamar Mike Morgan, the Nexus of the Baseball Universe? (37  12:33pm, Jun 23) Last: The Keith Law Blog Blah Blah (battlekow) Sabermetrics, Scouting, and the Science of Baseball – May 21 and 22, 2011 (2  8:03pm, May 16) Last: Diamond Research Retrosheet SemiAnnual Site Update! (4  4:07pm, Nov 18) Last: Sweatpants What Might Work in the World Series, 2010 Edition (5  3:27pm, Nov 12) Last: fra paolo Predicting the 2010 Playoffs (11  5:21pm, Oct 20) Last: TomH SABR 40: Impressions of a FirstTime Attendee (5  11:12pm, Aug 19) Last: Joe Bivens, Minor Genius St. Louis Cardinals Midseason Report (12  12:42am, Aug 10) Last: bjhanke Napoleon Lajoie: Definition of Grace (9  12:38am, Jul 01) Last: Hang down your head, Tom Foley Youth Baseball Hitting Drills: Shine the Light (5  6:47am, Mar 11) Last: Pat Rapper's Delight 



Page rendered in 0.3011 seconds 
Reader Comments and Retorts
Go to end of page
Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.
You must be Registered and Logged In to post comments.
<< Back to main