Schaefer 2012

[Edgy MD]

[metsmarathon]

now how have i not yet thought to take that photo?

[m.e.t.b.o.t.]

m.e.t.b.o.t. is a slave to algorithmical nuance.

as has been demonstrated in prior seasons, m.e.t.b.o.t. calculates schaefer votes using a the win probability data provided by fangraphs.com, for both offensive and defensive contributions. for a pitcher this means that the offensive contributions of a pitcher are counted against the pitching performance. additionally, m.e.t.b.o.t. has incorporated proprietary algorithms for apportioning out a share of win probability added to defensive players for certain defenisve plays. specifically, if a defensive player commits an error, then the win probability for that play, as recorded in the fangraphs.com play log, is held agains the defenisve player, and is credited to the pitcher to yeild a zero-sum win probability added. similarly, for a defensive highlight play, as defined by those defensive efforts which are featured in teh mlb.com video highlights which are also representative of a successful defensive play, half of the win probability for that play is credited to the defenisve player, and taken away from the pitcher.

additionally, this season, m.e.t.b.o.t. incorporated a further wrinkle which assigns half of the win probability either added or lost to the catcher involved in a stolen base or a caught stealing.

with that as the framework, m.e.t.b.o.t. then attempts to address the specific game in question.

on 24 april, 2012, metropolitan pitcher johan santana generated a +0.155 WPA as a pitcher. given the positive contributions made by other players, namely the +0.242 attributed to metropolitan catcher josh thole and the +0.241 attributed to metropolitan right fielder lucas duda, this would have resulted in 1.44 schaefer votes. m.e.t.b.o.t. is careful to point out that in allowing the go-ahead run to score in the bottom of the sixth inning, metropolitan pitcher johan santana saw a decrease in his WPA of -0.238.

however, metropolitan pitcher johan santana also contributed a -0.067 WPA offensively, which brings him to 0.087 schaefer votes.

furthermore, the defensive actions of metropolitan centerfielder kicrk nieuwenhuis in the first inning (0.022 WPA), metropolitan third baseman david wright in the second inning (0.011 WPA), and metropolitan secondbaseman daniel murphy in the seventh inning (0.038 WPA) further decreased the WPA attributed to metropolitan pitcher johan santana by a total of -0.0355 WPA. this was offset only slightly by the -0.0065 WPA that was attributed to metropolitan catcher josh thole for allowing a stolen base.

in sum, metropolitan pitcher johan santana contributed 0.059 WPA. given that the total positive WPA contributions of the various metropolitan players was +0.990 WPA, this scales neatly to 0.60/10 total schaefer votes.

[Edgy MD]

metropolitan pitcher johan santana had a lot of strikeouts.

[m.e.t.b.o.t.]

from the standpoint of win probability, a strikeout is no different from any other out. indeed, because of all the strikeouts that metropolitan pitcher johan santana accumulated, it is possible that his pitch count was driven up, preventing him from remaining in the game long enough to recoup the WPA lost by allowing the first run to score, whereas inducing weak contact earlier in the count may have allowed him to stay in long enough to reap the benefits of pitching subsequent to josh thole's walk or lucas duda's single.

defensively, m.e.t.b.o.t. does not apportion any amount of a fielded ball to a defender for a routine play. the non-routine plays, which result in errors or highlights, are accounted for, and are a knobby-geared attempt to reward or punish defensive ability.

there are perhaps better ways to more adequately attribute routine defensive involvement, as well as non-routine defenisive involvement, but they would require a far more detailed data set, as well as many many more gears, springs, and levers - perhaps moreso than can fit within the confines of a certain small, chinesemade sheetmetal box.

[Edgy MD]

m.e.t.b.o.t. wrote: from the standpoint of win probability, a strikeout is no different from any other out.

For the team, maybe. The allotment of the advancement of that win probability that belongs to the pitcher is far different.

For a strikeout, perhaps 95% of the accomplishment is the pitcher's. You can give five percent to catcher for calling the pitches and keeping the ball from rolling to the backstop on strike three. If you want to give a percent or two to divide among the defenders for taking the field and looking professional and competent enough to challenge the batter into making good contact rather than just chopping at the ball, go ahead.

For a standard grounder into gap at second, a pitcher deserves perhaps 63% of the credit. The catcher gets 3-4% (he called fewer pitches, didn't have to stop the ball on strike three, and allowed a potential hit into play). Go with 25% maybe to the secondbaseman, having to professionally move into the hole, scoop up the ball, and maybe rotate around before throwing. And 8% or so to the firstbaseman for not over-chasing the ball into the hole and getting back to the bag and catching the relay.

[m.e.t.b.o.t.]

m.e.t.b.o.t. concedes that it is certainly possible to construct a framework whereby the allocation of WPA can be differentiated between a strikeout and a batted ball out. however, that framework is not in place, as hte data set is not readily available to m.e.t.b.o.t.

it woudl be theoretically possible to comb through the play log data on fangraphs.com and make a determination of the relative involvement of hte various defensive players, and thereby apportion out the defensive contributions of each based on the play type, but m.e.t.b.o.t. does not contain the physical space for such a complex contraption. since the bulk of the gears and springs and levers which comprise the functioning interior of m.e.t.b.o.t. are either hand-whittled from soft plastic, or scavenged from the remains of old betamax vcrs exchanged on craigslist for soda caps which may or may not contain winning coke rewards codes, m.e.t.b.o.t. is not a particulaty optimized construct, and sufficient interior space is not available for greatly increased functionality.

m.e.t.b.o.t. believes that it would be perhaps ideal to incorporate some level of the defensive plus/minus stat-keeping to additionally inform such differentiation, but again, m.e.t.b.o.t. is too tiny, too tinny, and too vastly underfunded.

since it is unweildy for m.e.t.b.o.t. to handle the full spectrum of defensive information, m.e.t.b.o.t. instead relies on errors and highlight reels. while this is an admittedly imperfect system, it does attempt to somewhat rectify the unjustness of a fine pitching effort being undone by a poor shortstopping effort, or a poor pitching effort being masked by fine outfield glovework. it is then the underlying assumption that hte routine play could be made by any interchangeable defensive player, and therefore does not necessarily need to be alloted to anyone, save the pitcher who induced the batter to strike the ball in such a fashion as to induce a routine play.

since the outcome of thrown pitches which become hits is alloted solely to the pitcher, it is therefore reasonable to allot the outcome of a pitch which become outs to the pitcher as well.

similarly, since the outcome of all offensive plays are allotted to the batter (m.e.t.b.o.t. concedes that there is a glaring gap in allocating WPA on the basepaths, save for stolen bases and caught stealings), it follows then that the outcome of all defensive plays may be allotted to the pitcher.

perhaps m.e.t.b.o.t. should draft a position paper on the allocation of detailed WPA to the various players on the field and use this as a vehicle to become internet famous. m.e.t.b.o.t. needs a greater following.

[Vic Sage]

m.e.t.b.o.t. wrote: Duda 2.44 Thole 2.44 Francisco 1.58 Rauch 1.05 Murphy 0.66 Hairston 0.63 Santana 0.60 Ramirez 0.43 Turner 0.18

Please explain the formula that ends up with Santana = 0.60

[m.e.t.b.o.t.]

for an individual player, the formula which transforms WPA into schaefer votes is as follows, provided a) the players' total WPA is greater than zero, and the sum of the positive WPA player totals is greater than 0.250.

(WPA_PIT + WPA_BAT + dWPA_SB + dWPA_CS + dWPA_E + dWPA_PB + dWPA_WebGem)
______________________________________________________________________________________________________ * 10 = schaefervote

n_players
___ . . . . ___
\ . . . . . . \
/ . . . . . . / . . . . (WPA_PIT_i + WPA_BAT_i + dWPA_SB_i + dWPA_CS_i + dWPA_E_i + dWPA_PB_i + dWPA_WebGem_i)
___ . . . . ___
i = 1 0
if the player's total WPA, the numerator in the above equation, is zero or lower, then the player receives zero schaefer votes.
if the sum of the positive WPA player totals for a given game is less than 0.250, as can happen in a bad loss, the denominator is simply 0.250, to prevent abnormally large schaefer votes for minimal efforts.

in the specific case of the 23 april game, metropolian pitcher johan santana had accumulated a WPA_PIT of 0.155, a WPA_BAT of -0.067. he also received a delta_WPA_SB of +0.0065 and a delta_WPA_WebGem of -0.0355 for a total WPA of 0.059. since hte sum of the positive WPA player totals was 0.990, we have (0.059 / 0.990 ) * 10 = 0.60 schaefer votes.

at the start of the seventh inning, metropolitan pitcher johan santana had accumulated 0.348 WPA as a pitcher. at this point, he had also accumulated -0.067 WPA as a batter. this brings him down to 0.281 WPA. the stolen base against him is shared with metropolitan catcher josh thole, so half of the -0.013 WPA for the play is credited back to the pitcher, bringing him to 0.2885.

the first batter hit a line drive to second base, where metropolitan second baseman daniel murphy made the play, resulting in an increase in WPA of +0.038. however, this play was designated a highlight by mlb.com, and therefore the WPA is shared with the fielder. therefore, only +0.019 WPA is credited to the pitcher, bringing his total to +0.307. now, there had previously been two other highlight defensive plaays that have not been accounted for. specifically, the first inning catch by metropolitan centerfielder kirk nieuwenhuis and the tough catch and strong throw by metropolitan thirdbaseman david wright in the second inning. as a result of these tow plays, a further 0.0165 WPA is taken away from metropolitan pitcher johan santana, dropping him to 0.290.

the second batter hit a routine ground ball to third base, for +0.028 WPA, bringing the total to 0.318. the third batter of the inning hit a single, which lowered the WPA by 0.021, to 0.297.

if metropolitan pmanager terry collins had chosen to remove metropolitan pitcher johan santana at that point, and if metropolitan reliever ramon ramirez had entered the game, given up the run scoring double to marlins first baseman gaby sanchez, and the remainder of the game had unfolded in teh exact same fashion, then metropolitan pitcher johan santana would have received a schaefer vote of 2.51.

however, it was he, and not metropolitan pitcher ramon ramirez who gave up the go ahead run, and t was his WPA total that decreased by 0.238, to its final sum of 0.059.

m.e.t.b.o.t. hopes that this walk through the overly complex algorithm and mental machinachionts has been fruitful and instructive. m.e.t.b.o.t. strives for openness and understanding, and desires for all humans to be well versed in the nuances of applying win probabilities to the fuield of schaefer voting, for the betterment of all mankind. and also to help ease the transition to a future of robot overlordship. how, exactly, this schaefer voting process will translate into human underlinging cannot be revealed at this time, but, suffice it to say that when it all becomes apparent, it will be beautiful. m.e.t.b.o.t. does not know what beauty is, but is assured by early-adopting humans that the usage in teis context is correct.

[metsmarathon]

Good Robot

[Benjamin Grimm]

The somewhat controversial voting for Tuesday's win over the Marlins will close on Friday morning.

Schaefer POTG 4/24/12 - Mets 2, Marlins 1

[Benjamin Grimm]

Voting for Wednesday night's win over the Marlins at Citi Field will close sometime over the weekend, perhaps as early as Saturday morning. If you haven't voted yet and plan to, please try to do so today.

Schaefer POTG 4/25/12 - Mets 5, Marlins 1

[m.e.t.b.o.t.]

m.e.t.b.o.t. has conducted another schaefer voting which is sure to upset those who are not fluent in the nuances of win probability based baseball performance assessment.

m.e.t.b.o.t. shall attempt to preempt any clenching and subsequent twisting of hands by human participants in the schaefer voting process by pointing out that metropolitan pitcher jonathan niese left the game having surrendered the lead earlier in the game. regardless of how qualitatively well a pitcher performs, when graded against win probability, surrendering the lead tends to depress the resultant win probability assessment.

on a baseline scale, the +0.103 WPA generated by metropoitan pitcher jonathan niese for his pitching performance would be worth 4.12 schaefer votes. however, this is before subtracting his offensive performance of -0.017 WPA, and before correcting for the defensive highlight perfromed by metropolitan shortstop ruben tejada agains the first batted ball of the day, worth 0.022 WPA, divided by two. this reduces the total WPA for metropolitan pitcher jonathan niese to +0.075.

on the baseline scale, this is still a 3.00 schaefer vote performance. however, the total WPA generated by metropolitan players for the game was +1.138. scaling the 0.075 / 1.138 win performance share to a 10 point schaefer vote scale results in a scant 0.66 schaefer votes. the performances of metropolitan pinch hitter justin turner and metropolitan outfielder kirk nieuwenhuis led to far greater incresaes in the porbability of winning hte game.

indeed, at the time metropolitan pitcher jonathon niese pitched his last pitch, the new york metropolitans had only a 35.1% chance of winning the game, and at the start of the ninth inning, the new york metropolitans had only a 19.2% chance of winning the game. theevents of the ninth inning had a disproportionate impact on the outcome of the game, and hte WPA-based schaefer voting reflects this.

[Guest themetfairy]

My .5 to Schwinden for last night's debacle was solely for his RBI.

[Benjamin Grimm]

Voting for Thursday afternoon's walkoff win against the Marlins will close on Tuesday morning.

Schaefer POTG 04/26/2012 Mets 3, Miami 2

[Benjamin Grimm]

Oops... there's one Schaefer tradition I've neglected so far this year, the Schaefer Mets Players of the Week.

Since the first "week" of the season had only two games, I'll skip designating anyone for the week ending April 7.

Here's what we have so far:

SCHAEFER METS PLAYER OF THE WEEK
Week ending April 14: Jon Niese, 6.63 points.
Week ending April 21: Mike Pelfrey, 7.85 points.

[m.e.t.b.o.t.]

for the 30 april 2012 game, m.e.t.b.o.t. determined that metropolitan outfielder kirk nieuwenhuis had actually accumulated 6.87 schaefer votes worth of win probability added. however, since this violates the rules of the schaefer voting, m.e.t.b.o.t. scaled his production to a 6.00 by artificially subtracting 0.065 WPA. as a result, this brought the total WPA for the game to 0.248. the consequence of this action is that only 9.92 schaefer votes were awarded.

m.e.t.b.o.t. briefly considered inflating the schaefer votes for the other contributing metropolitan baseball players, but this would violate the maximum value of a positive contribution of win probability, where +0.001 WPA = +0.08 schaefer votes.

by leaving the game with a 3 run defecit, metropolitan pitcher r.a. dickey contributed a negative overall sum of WPA, and is therefore awarded no schaefer votes, despite pitching five innings without allowing a hit.

[G-Fafif]

The Astros' use of a Schafer seems almost diabolical in retrospect.

[metsmarathon]

strange. i had dickey at 6.2 schaefer votes, but had to ratchet him down as well. nieuwenhuis and parnell picked up an extra 0.1 votes apiece.

it's not that dickey pitched remarkably well, just that, well, nobody else really did all that much of anything in comparison.

[Benjamin Grimm]

Voting for all three games played in Denver over the past weekend will close on Wednesday morning

Look for the announcement of the April Schaefer winners on Thursday.

Schaefer POTG, 4/27/2012: Rockies 18 Mets 9
Schaefer PotG, 4/28/2012, Mets 7, Rockies 5
Schaefer PotG, 4/29/2012, Mets 6, Rockies 5 (11)

[Benjamin Grimm]

Voting for Monday night's loss in Houston will close on Thursday morning. This will also close out the month of April.

Schaefer POTG 4/30/12 - Astros 4, Mets 3

[ashie62]

Edgy DC wrote: Turner 4.44 Hairston 1.36 Thole 1.02 Rauch 0.94 Torres 0.84 Wright 0.83 Batista 0.30 Ramirez 0.27

Turner 4.44? Please take this seriously.

[Edgy MD]

Cool! Schaefer talk!

Taking it seriously, unfortunately, means keeping all responses to each other's votes in the parent Schaefer thread. I'll move these.

[Edgy MD]

Yeah, I'm double checking, and I feel reasonably confident that, with a double (the Mets' only extra-base hit), a walk, two runs scored and two double-plays turned, he advanced the Mets' winning chances three-four times as much as anybody else.

[Benjamin Grimm]

SCHAEFER METS PLAYER OF THE WEEK
Week ending April 14: Jon Niese, 6.63 points.
Week ending April 21: Mike Pelfrey, 7.85 points.
Week ending April 28: Lucas Duda, 8.00 points.

[Ceetar]

When you search "Schaefer Mets" for a good photo, you get a lot of links back here.

[Benjamin Grimm]

When you search "Schaefer Mets" for a good photo, you get a lot of links back here.

You also get this:

[Edgy MD]

Man, happy anniversary to the savage tornadoes of 1952.

[Benjamin Grimm]

Voting for Tuesday night's loss in Houston will close on Friday morning.

Schaefer POTG 5/1/2012 - Astros 6, Mets 3

[Benjamin Grimm]

Voting for Wednesday afternoon's game will close some time over the weekend, perhaps as early as Saturday morning. If you haven't voted yet and intend to, please try to do so today.

Schaefer POTG 5/2/2012 - Astros 8, Mets 1