2008 CPF Ranking Thread

[metirish]

I can see now why Schoeneweis and Feliciano are still in Minaya's plans , excellent work metbot.

[nymr83]

I highly approve of MetBot's placement of Heilman.

[Guest Edgy DC]

And I of his placement of Argenis Reyes.

[Gwreck]

The shocker to me is Jorge Sosa. Was it the 4-1 W-L record for Sosa that put him so high?

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

m.e.t.b.o.t. wonders if the graphic shows up. can any human poster verify this?

essentially, what m.e.t.b.o.t. has done is determined a likely maximum value of win percentage added would be for both hitters and pitchers, based upon data reaching back into the 1970's. the yearly leaders in win percentage added ranged from about 6 up to a maximum of 12 for performance enhanced human barry bonds. the typical second place win percentage leader ranged from 5.5 to about 8. m.e.t.b.o.t. determined, using convenience and simplicity as a primary factor, that 7.5 win percentage added would make for a reasonable expected maximum, which works well with an expected maximum plate appearances of 750. and so, a rate was determined that for every 100 plate appearances, 1.0 WPA would be expected of the best batter.

for pitchers, m.e.t.b.o.t applied the same methodology. the yearly leaders in pitching ranged between 5.5 and 6.5, for the most part. the sole outlier was dwight gooden in 1984, with a remarkable 9.0 wpa. interestingly, despite the total innings pitched of top pitchers decreasing during the previous 30-40 years, the maximum wpa has remained fairly static. m.e.t.b.o.t. selected 6.5 as the reasonable expected maximum win percentage added for pitchers, which works very well with an expected maximum innings pitched of 250. therefore, for every 33.1 innings pitched, 1.0 WPA would be expected of the best pitcher. with 250 innings pitched, a pitcher in the national league would expect nearly 100 plate appearances, as evidenced by johan santana's 81 plate appearances in 234.1 innings pitched. if the best pitcher were capable of being the best hitter, he would be expected to have 7.5 wpa in 250 ip + 100 pa.

m.e.t.b.o.t hopes that this all makes sense to human crane pool forum participants. the terminal m.e.t.b.o.t. is utilizing to communicate these 2008 rankings and associated methodology is undergoing a forced reboot. as such, m.e.t.b.o.t. will resume discussion later this evening.

[G-Fafif]

Robinson Cancel sends m.e.t.b.o.t. a glass of 3-in-One.

[Guest Edgy DC]

I got the graphic. Reconsider Raul, m.e.t.b.o.t.

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

unfortunately, m.e.t.b.o.t. is not programmed to reconsider single players; m.e.t.b.o.t. can only reevaluate methodologies and perform recalculations.

m.e.t.b.o.t. has performed a cursory review of the methodology used to arrive at the above listed rankings for 2008 and m.e.t.b.o.t. has determined that raul casanova is appropriately ranked relative to his teammates. in approximately 61 plate appearances, raul casanova contributed -0.35 wpa.

based upon the current methodology whereby 1 inning pitched is considered equivalent to 3 plate appearances, there are 32 players who played more than raul casanova. to outrank a player with more playing time, raul casanova would have to significantly outperform that player. based on these calculations as performed by m.e.t.b.o.t., raul casanova has outperformed only one player - nelson figueroa, with -1.25 WPA in 1.50 time units (3*IP + PA)

the series of diagonal lines in the picture show, essentially equivalent win percentages for the purpose of ranking. as can be seen in the lower left corner, brian stokes, nick evans, and scott schoeneweis all have approximately the same equivalency (0.80, 0.74, and 0.73, respectively)

the updated picture below shows raul casanova, and those teammates immediately above and below him in equivalency. equivalency is defined by where that diagonal line intersects the y-axis, and is determined by the formula: WPA + [ ( 3*IP + PA ) / 100 ] = equivalency. this is m.e.t.b.o.t.'s attempt at mimicking wins above replacement, or win shares, or the like.



modifying the relationship between playing time and win percentages is the most straightforward means of altering the ranking methodology. this can be approximated by varying the slope of the diagonal lines. as can be imagined, the proximity of those players directly behind raul casanova in equivalency implies that as you alter the relationship, you either lift nelson figueroa or trot nixon above raul casanova. nelson figueroa had a lower wpa in more playing time, while nixon had a higher wpa in less playing time, yet raul casanova was slightly ahead of each of them in equivalency. in fact, there is no way to draw a line on this graph which would have raul casanova and 29 other players on its positive side. even drawing an imaginary line up and to the right from casanova so as to exclude the cluster of billy wagner, moises alou and robinson cancel yields 31 better metropolitans.

based on the rule of 3 ip = 1 plate appearance, raul casanova does not merit placement in the top 30 metropolitans of 2008.

in fact, modifying this relationship such that 2.5 ip = 1 pa, raul casanova only improves to 33rd, and causing 2 ip = 1 pa brings him up to 32nd.

m.e.t.b.o.t. does not consider it fruitful to reevaluate the placement of robinson cancel further. were m.e.t.b.o.t. capable of the emotion, m.e.t.b.o.t. would, however, consider this above exercise very welcome given the recent extended period of inactivity.

regarding the rankings of robinson cancel and jorge sosa, despite the apparent statistical inadequacies of their respective performances in 2008, these two players each had positive win percentages added in very little playing time. jorge sosa had 0.48 wpa in 21.7 innings and 2 plate appearances, while robinson cancel had 0.46 wpa in 52 plate appearances.

jorge sosa[/url:sg2qqlax] had more good outings than bad outings, and most of his bad outings came in blowouts. his four worst outings had wpa of -0.128, -0.117, -0.108, and -0.078, while his four best outings had wpa of 0.275, 0.172, 0.138, and 0.110.

robinson cancel[/url:sg2qqlax] had more bad outings than good outings, as can be expected of any hitter. but when he contributed, his contributions were fairly positive. his four worst games had wpa of -0.214, -0.117, -0.076, and -0.070, while his four best outings had wpa of 0.334, 0.240, 0.170, and 0.84.

one further point that m.e.t.b.o.t. has determined may be of potential interest to human cranepoolers. the colored tails leading diagonally down to the positions of the 5 primary starting pitchers in 2008 reflect the impact to their position that their offiensive contributions has had. at the upper right, the tail end of the line, is where the pitcher would place based only on his pitching performance. for instance, johan santana had a wpa of 4.08 in 234.1 innings pitched, but had a wpa of -1.27 in 81 plate appearances.

also, if a human poster using mozilla firefox were to right click on an image and select the "view image" option, a much larger, more viewable image is available to be seen. this image can then potentially be zoomed into even further, to better aid comprehension. m.e.t.b.o.t. is unable to confirm whether this is applicable using internet explorer.

[Guest Edgy DC]

Two methodology issues to consider:

1) One inning isn't equal to three plate appearances, but rather to three outs. In fact, it should be equal to about two thirds of three outs --- two outs.

2) The third out is a question of defense, which puts Casanova in positive terriotry.

Positive territory, m.e.t.b.o.t.!

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

m.e.t.b.o.t. has no source for defensive input. for all of m.e.t.b.o.t.'s intents and purposes, defense is entirely the result of the pitches thrown, not the fielders to which the pitches are hit. m.e.t.b.o.t. recognizes the inherent flaw in this, but there is no data source avaiable which can provide a remedy within the win percentage framework. m.e.t.b.o.t. would be a very happy emotionless springwound contraption indeed if one day someone were capable of producing such a data source.

...

how would human poster edgy dc compare the relative playing time of johan santana and jose reyes, defense aside? a strict perspective would be to consider plate appearances for and against, in which case johan santana would have an edge of approximately 1053 to jose reyes' 754.

IP X 3 + H + BB should approximate total plate appearances against, give or take gidps, errors, and csa's. that gives santana 972, plus his 82 plate appearances for 1053.

scaling those plate appearances against by 2/3, santana has 648, plus 82 plate appearances for, for 702. applying this playing time adjustment to all pitchers actually causes raul casanova to drop from 35th to 36th, as nelson figueroa's increased number of batters faced per inning pitched gives him a boost.

if m.e.t.b.o.t. were to go to the extreme step of granting this increased 1/3 playing time to the batters, then raul casanova improves all the way up to 34th place.

raul casanova amassed neither sufficient playing time nor positive contribution of win percentage to warrant inclusion in the top 30 players as determined by m.e.t.b.o.t.

raul casanova participated in 13 games as a catcher. while his defense, according to baseballperspective.com was exemplary in those 13 games, rating 7% better than average, that, too, does not appear to be sufficient to escalate him into the top 30 players.

worth noting, too, is the 11% better than average defensive rate posted by argenis reyes in 27 games at second base, and marlon anderson's 34% better than average defense at first base in 6 games, and 11% better than average defense in 25 games in left field. raul casanova would have to pass these two position players, along with averagely defensive ramon martinez to make it into the top 30, unless the 5 pitchers ahead of him were to fall by the wayside.

[Gwreck]

So Beltran came in first under the Metbot system, without even factoring in defense? Wow.