sitemeter:

June 14, 2007
  
Win Probability Added-WPA Numbers (Up to date) Shocking Relevations!
by: Future on Jun 14, 2007 7:45 PM | Filed under: Journals

WPA(Win Probability Added) Numbers thus far: (With a little Batting Runs Above Average Tossed In for Kicks)

Numbers from FanGraphs-Mets Team Stats

The numbers will surprise you!

1. Delgado: .90 This is due to some big circumstances like Delgado’s walk-off home run. But rest assured, he has been terrible this season. It is tempered by how many opportunities he has had to hit in a big spot and failed. His BRAA: -5.48. That’s pretty fugly. Ok, very fugly.
2. Wright .89 :This is due to his remarkable hitting without RISP, with his ability to drive himself in via the longball. His BRAA:12.46, which is almost a 18 points difference of BRAA between the #1 and #2 spots. And the #2 spot is the one up. That’s crazy.
3. Reyes: .88 :Who is surprised Reyes is up there? I’m only surprised it isn’t over 1. His BRAA: 11.54, very nice. Even more impressive, he leads the NL in Win Shares with 14.
4. Easley: .83 :Easley is rapidly worming his way into my heart. His BRAA: 2.22, which isn’t that strong, but suggests he’s come up big in bigger spots.
5. Green: .41 This is for you Dep. His BRAA: 4.24. Not bad. .
6. Franco: .13 (Huge Surprise to me) How someone so bad can have a positive WPA is shocking to me. His BRAA: .60. Well, not negative that’s a positive.
7. Gotay: .04 Also worming his way into my heart. His BRAA: 2.67. If you’re going to DH someone, DH this guy or Easley. (He’s done this in 44 at-bats)
8. Ledee: .02 (So stop hatin’ on him) Yes, he’s not an attractive man , but he’s not as bad at his game as you might think.His BRAA: .40.
9. Valentin: .02 I expect this to rise. His BRAA: 2.93. I expect this to definitely rise because of his good 2.93
10. Beltran: -.01 (Huge) With how bad he has been this season, I feel he’s rapidly approaching the “what have you done for me lately territory and will hear boo-birds at Shea.” His BRAA: 6.54. This makes me wonder how unlucky he has been, he crushed some balls that were all outs in the Mets’ last game in LA.
11. Lo Duca: -.19 (Ridiculously bad, you guys want him in the two spot? Well, at least he can’t crush the win probability as bad from there. His batting average is further proof that batting average alone is one of the more worthless stats out there. With his defensive shortcomings, it makes it apparent the Mets will be going elsewhere next season.) His BRAA: -6.74, the worst of any regular. One person the fanbase has not gotten on is PLD, and his batting average aside he has been terrible.
12. Newhan: -.24 (Considering he had that one big HR, this is pretty impressive to be that bad in such a short time.) His BRAA: -.53.
13. Castro: -.29 (And someone says he is as good as Lo Duca? He did this in 50+ at bats people!) His BRAA: -3.09.
14. Chavez: -.41 (Remember, this doesn’t include his defensive abilities) His BRAA: -3.20.
15. Gomez: -.51 (Yikes, that’s not so good.) His BRAA: . His BRAA: -3.85.
16. Johnson: -.54 (Wow…so bad in such a short amount of time. Can we get a refund from the Padres?) His BRAA: -3.43.
17. ???: -.75 (Yes, this is the worst stat on the team. This player has been so bad with the bat he is has almost cost the team 3/4 of a win all by himself. Can you guess who he is? Since you fans out there are smarter than the average bear just by being Mets fans, you’re right if you chose Moises Alou. So maybe we shouldn’t be worried he’s out longer. His BRAA: -2.70, bad, but nowhere near as bad as PLD.

Starting Pitchers:

1. El Duque: 1.48 (What can I say? He’s done this in 9 starts.)
2. Sosa: 1.27 (He’s done this in 8 starts! Very impressive.)
3. Maine: 1.16 (In 13 starts Maine has been solid.)
4. Perez: .66 (I was surprised when I saw this myself. At times he is our ace.)
5. Glavine: .12 (Not surprised by this at all. Glavine will undoubtedly have a spot in our playoff rotation if the team pulls out of the tailspin and takes the division. But he’s our worst starter and has been unimpressive. While he might have beef about the bats not protecting him in his first season with the Mets, the fault of not being closer to 300 this season is on him alone this season. Not to mention he drives many of us crazy with his nibbling routine.)

Relievers:

1. Wagner: 1.40 (Would be higher if he hadn’t ran into Pat Burrell.)
2. Feliciano: .55 (He’s been very solid for us this year.)
3. Smith: .31 (Pleasant surprise, actually surprised it’s not higher.)
4. Sele: .08 (Wow…this…is a positive?)
5. Schoeneweis: -.01 (See, he’s not that bad, he’s terrible but doesn’t really affect the outcomes of games)
6. Mota: -.09 (In 8 innings! Now that’s impressive!)
7. Burgos: -.23 (Yeah, not so good)
8. Heilman: -.79 (Some failures in the big spots, but he’s turned it around of late. I blame this partially on Willie because Heilman isn’t getting many of the high-leverage situations that would boost his WPA.)

Kalyan

Thanks to fangraphs, with which this journal would not have been possible.

Win Probability Added is a statistic that measures how much an individual player has positively or negatively changed the team’s probability of winning the game. If a player grounds out with runners on base, the team’s chances of winning go down, and that change is reflected in the player’s WPA. If he gets a hit and drives in a run(or runs) that changes the team’s chances of winning in a positive way.

Game-ending plate appearances that end positively, like walk-off home runs, obviously are huge WPA-boosters. It’s why David Ortiz’s WPA was so high last year.

To put it all into perspective, let’s look at the final 2006 WPA numbers.

To put this into perspective

At the end of last season, these were the total WPA numbers for the 2006 Mets.

1. Beltran: 5.21 and BRAA 63.44 (Last year was an epic season and none but the most optimistic Mets fans expected him to replicate his season, but he has slightly less than 100 games to rescue his season, because right now he isn’t anywhere close. And we know he’ll need many days off between now and then.)
2. Wright: 4.64 and BRAA 36.62 (With how Wright is hitting, he could actually replicate last season’s performance)
3. Delgado: 3.11 and BRAA 37.38 (Him and his home-boy Ponce De Leon need to hang out, last year’s regular season by Delgado was a disappointment to me, if only I knew how terrible he would hit this year, I would take all the disappointment away, put it in a doggie bag and save it for now)
4. Reyes: 2.11 and BRAA 28.22 (Reyes can surpass last season’s performance)
5. Valentin: 1.47 and BRAA 9.07 (Valentin still has time to reach these levels)
6. Chavez: 1.09 and BRAA 9.64 (Most likely last season was a fluke, but he has started many more games this year than he is used to and when he comes back from injury I expect the rest of the outfield :prays to God: to be healthy to allow him to spell the regulars and possibly replicate last season)
7. PLD: 0.96 and BRAA 9.22 (He needs to really start producing and maybe it does have something to do with hitting in the #2 spot_
8. Floyd: -0.24 and BRAA -0.40 (See, for as bad as Floyd was last year, and how unlucky he was, his WPA was around what PLD’s is rapidly approaching. Chew on that.)
9. Nady: 0.10 and BRAA 1.88 (His WPA is so low because he was terrible with RISP, his WPA reflects the sentiments of Met fans who don’t believe he is the answer)
10. Woodward: -1.12 and BRAA -14.01 (Yes, Chris Woodward was so bad he was worth one negative team win by himself- Thank God he can bring Atlanta down now, the gift that keeps giving)
11. Milledge: 0.23 and BRAA -5.91 (Surprising that with such a terrible BRAA he has a positive WPA, but these things happen)
12. Franco: -0.72 and BRAA -7.14 (Having a better season than last, if that means anything)
13. Castro: -0.65 and BRAA -3.23 (Wasn’t good last year, isn’t good now)
14. Green: -0.50 and BRAA -2.41 (Besides that stupid at-bat against Penny, has realized his limitations and now plays to his strengths)
15. Hernandez: -0.42 and BRAA -10.20 (Who is surprised Anderson Hernandez has these stats)
16. Matsui: -1.94 and BRAA -16.30 (He did this in 38 games! One of the worst players in Mets history! How do you have approximately -2 wins in only 38 games! How are you that bad!)

In summation, besides Wright and Reyes, all the Mets starters are much worse than last year. Franco is somehow….better.

Also our hitters were epically bad, costing our team several wins during the season.


34 Responses to “Win Probability Added-WPA Numbers (Up to date) Shocking Relevations!”

  1. Comment posted by Jersey John on June 14, 2007 at 8:33 pm (#377549)

    …Hello Future

    Those stats and commentary are apparently quite accurate, although for me, at least somewhat mysterious.

    I’m a little embarrassed to say I’m not familiar with
    WPA and BRAA. Can you or anyone else give me a link to the website that originated that? I assume there are explanations there.

  2. Gravatar
  3. Comment posted by Jessica on June 14, 2007 at 8:45 pm (#377551)

    JJ, go to http://www.fangraphs.com to find out more about WPA.

    The way it basically works is that for every given game situation, the probability of a team winning is determined. Let’s say that Delgado is at the plate with the Mets down by a run in the top of the 6th, runners on 2nd and 3rd, 2 outs. When he grounds out to the 2nd baseman to end the inning, the difference in the Mets’ chances of winning before his AB and now is added to his cumulative WPA total, and the opposite value is added to the opposing pitcher’s WPA total.

  4. Comment posted by NjMetsFan on June 14, 2007 at 9:07 pm (#377555)

    You crack me up Future.

  5. Gravatar
  6. Comment posted by David Simons on June 14, 2007 at 9:56 pm (#377560)

    Well done, fascinating intel there. Does this mean that Willie is actually using Schoenweis to perfection? Surprises never end!

  7. Gravatar
  8. Comment posted by Future on June 14, 2007 at 10:17 pm (#377567)

    There were two ways I could go with it.

    1) Schoeneweis is so fucking terrible that he has broken the WPA graph. He is a black hole of suckiness and the computers that calculate WPA have been crushed to smithereens.
    2) Schoeneweis comes in when we are already losing, gets lit up, we still lose. He comes in when we are winning and is usually pulled before he can execute the coup de grace to our WP.

  9. Gravatar
  10. Comment posted by Jessica on June 14, 2007 at 11:09 pm (#377611)

    Looking at Show’s gamelog ( http://www.fangraphs.com/statsd.aspx?playerid=33&position=P&season= ), there are only a handful of appearances in which he significantly decreased the team’s chances of winning – basically his extra-inning meltdown against the Phillies last week, the Edgar Renteria game-tying 3 run homer, the Aramis Ramirez grand slam that turned a close game into a blowout, and the blowout he turned into a close game against the Yankees. On the flip side, he’s had about 5 or 6 good appearances in high-leverage situations. The rest were either in garbage time or too short to really impact the outcome of the game.

  11. Comment posted by Jersey John on June 15, 2007 at 4:37 am (#377641)

    …Thanks Future and Jessica

    How about this?: As a releif pitcher’s W-L record can be misleading (gives up lead; then gets win), aren’t there any holes in those measurements?

    I wonder how A-Rod’s actual performance last year compared with what those numbers supposedly showed what his did. Many Yankee fans thought he failed a lot. Maybe those numbers made it appear differently.

  12. Gravatar
  13. Comment posted by Jessica on June 15, 2007 at 9:33 am (#377707)

    A-Rod had a cumulative WPA of 1.18 last year, which is really low considering his overall performance (to put it in perspective, his WPA for this year so far is 3.18). At least in the case of last year, the old gripe that he hit his home runs when his team was up or down by a lot but grounded into DPs with the game on the line seems to have held true.

  14. Gravatar
  15. Comment posted by dptydwg420 on June 15, 2007 at 9:51 am (#377720)

    Great journal future. fascinating stuff.

    I like WPA for batters much more than for pitchers. I feel like they just add the opposite score to the pitcher at the time for like accounting purposes so everything balances.

    #’s that surprised me a little, PLD’s being as bad as it is and Show’s was surprising. i love your black hole explanation in #5 above.

    Someone talk to me about BRAA – is it runs driven in above average? what does the “batted runs” part speak to exactly?

    PS. thanks for the green love and shoutout future. I liked it :)

  16. Gravatar
  17. Comment posted by Jessica on June 15, 2007 at 10:21 am (#377736)

    I like WPA for batters much more than for pitchers. I feel like they just add the opposite score to the pitcher at the time for like accounting purposes so everything balances.

    Yeah, that’s pretty much what Fangraphs does, which works pretty well with relievers but not so well for starting pitchers. I downloaded a spreadsheet last year that lets you track WPA on your own, and whenever there was a fielding error I wasn’t sure what to do (credit the pitcher for the out he would have gotten and give the fielder blame for the difference in win probability between what should have happened and what actually happened?). Same thing with a great defensive play to take away what would have been a hit. At least outfield assists were straightforward (blame the pitcher for what he gave up, credit the outfielder for the out and the erasing of the runner).

  18. Gravatar
  19. Comment posted by Future on June 15, 2007 at 10:54 am (#377760)

    I made a revision, adding some of the 06 numbers for comparison.

  20. Comment posted by sheadenizen on June 15, 2007 at 11:02 am (#377768)

    Future, when do you have time to study?

  21. Gravatar
  22. Comment posted by Future on June 15, 2007 at 11:08 am (#377776)

    The semester is over. I have a job with Princeton Review starting tomorrow.

  23. Gravatar
  24. Comment posted by dptydwg420 on June 15, 2007 at 11:12 am (#377779)

    Exactly Jess.

    This situation also bothers me for WPA for pitchers.

    say the offense puts up a 6 spot in the 1st inning, then the pitcher goes out and throws a shutout.

    he’s probly not going to get a nice chunk of WPA because the game was decided by the mets bats in the 1st inning.

    I think something similar happened with one of el duque’s starts this year.

  25. Gravatar
  26. Comment posted by dptydwg420 on June 15, 2007 at 11:16 am (#377787)

    Wow. Beltran was a god last year.

    nice addition of last year’s #’s future.

  27. Comment posted by john (1st member of the corey coles fan club) on June 15, 2007 at 11:57 am (#377803)

    say the offense puts up a 6 spot in the 1st inning, then the pitcher goes out and throws a shutout.

    I get what your saying…..but I mean what really would be the fair way of separating it out in that situation? Im not sure the answer to that.

    The problem I have with the fangraphs one and im not really sure they could do anything about is those graphs dont take into account starting pitching…for example….the mets have a better chance winning a game started by a pitcher with a 5.00 ERA then one with a 3.00 ERA.

    There are other things as well but for the most part i like it.

  28. Comment posted by john (1st member of the corey coles fan club) on June 15, 2007 at 11:58 am (#377804)

    I meant to say the mets have a better chance facing a pitcher with a 5 ERA then 3 ERA lol

  29. Gravatar
  30. Comment posted by dptydwg420 on June 15, 2007 at 12:03 pm (#377806)

    yea john, i dont know the fair way of breaking it out. i just like to point out shortcomings without offering solutions. its my thing lol ;)

    and yup, 100% agree with your 3.0 ERA vs 5.0 ERA.

    The WPA should not be split even 50/50 when the game starts.

    even going beyond just starting pitchers, when the mets play the nationals or some other lowly team. the mets should have a better than 50% chance of winning.

    there should be some odds baked into it if it wanted to be truer to what the real probabilities are.

  31. Comment posted by john (1st member of the corey coles fan club) on June 15, 2007 at 12:48 pm (#377851)

    The WPA should not be split even 50/50 when the game starts.

    even going beyond just starting pitchers, when the mets play the nationals or some other lowly team. the mets should have a better than 50% chance of winning.

    there should be some odds baked into it if it wanted to be truer to what the real probabilities are.

    Exactly. I think the reason they do it 50/50 tho is so that it reconciles to the actual team wins at the end of the season. Like say the total WPA for a team is 10 at the end of the season….that team went 86-76.

    You could add a whole bunch of factors, starting pitching, home field advantage, the how good or bad the other team is…..thing is once u start doing all that…your gonna have a ton of WP tables lol. So while it would increase the accuracy, you wonder if its even worth it.

  32. Comment posted by john (1st member of the corey coles fan club) on June 15, 2007 at 12:50 pm (#377852)

    Another problem is this…….say pedro is pitching……he wins what 70% of the time…..well then if say the mets win……the mets only get .300 Win probability added……so say pedro pitches a 1-0 gem…..he’ll get most the .300 WPA however say a bad pitcher does the same……he’ll get more WPA added since he was less expected to do that….thats where not separating it 50/50 could cause issues.

  33. Comment posted by argonbunnies on June 15, 2007 at 1:00 pm (#377854)

    Idea that just occurred to me:

    WPA stat-keeping includes tracking the situation of every plate appearance for every player. If you save that data and compare it with a player’s total hitting stats (OPS, types of outs, etc.), you should be able to produce an “Expected WPA” stat.

    The difference between Expected WPA and actual WPA would be that player’s Clutch index.

    Hypothetical illustration:
    Beltran’s .980 2006 OPS in the situations he batted in gives Expected WPA of 5, so his 5.21 = +.21 Clutch, while Wright’s .910 OPS the situations he batted in gives Expected WPA of 4, so his 4.64 = +.64 Clutch, backing up contentions that Beltran was the better player, but Wright more elevated his game in key spots.

  34. Comment posted by john (1st member of the corey coles fan club) on June 15, 2007 at 1:04 pm (#377855)

    Actually with clutch they did something like this.

    OPS Wins which is basically your worth not factoring in score and men on base

    BRAA which accounts for men on base

    WPA which accounts for men on base and score.

    So before with clutch all they were doing is WPA-OPS Wins which made a whole lot of sense to me.

    Now they seem to have gotten rid of OPS wins and is using clutch as WPA/LI….im not sure how that colrelates to clutch tho.

  35. Comment posted by john (1st member of the corey coles fan club) on June 15, 2007 at 1:06 pm (#377856)

    I should have added

    OPS Wins is basically the 1.8OBP + SLG……..it correlates to runs scored very well.

  36. Gravatar
  37. Comment posted by Future on June 15, 2007 at 1:07 pm (#377857)

    I’m thrilled, this sparked more comments than a feature!

  38. Comment posted by john (1st member of the corey coles fan club) on June 15, 2007 at 1:12 pm (#377859)

    oh i see what they are doing

    They are taking WPA…..and WPA/LI and the difference is Clutch

    For instance….beltran had a 5.21 WPA in 06……his WPA/LI was 3.98 for a difference of 1.24.

    I guess WPA is what he did in all situations….and WPA/LI strips out the leverage and makes everything equal…..so that clutch makes sense as well.

    Wright had 4.64 and a WPA/LI of 2.56 for a clutch total of 2.08

    So beltran was clutch but wright was even moreso….which looking back to last year makes sense. Wright had a couple walkoffs and so did beltran.

  39. Comment posted by john (1st member of the corey coles fan club) on June 15, 2007 at 1:13 pm (#377860)

    I’m thrilled, this sparked more comments than a feature!

    I love WPA….I just thought everyone hated it so i didnt wanna bore ppl by talking about it lol

  40. Gravatar
  41. Comment posted by Future on June 15, 2007 at 1:22 pm (#377863)

    Hah. WPA is awesome. As are Win Shares. (Reyes still leads the NL in Win Shares.

  42. Comment posted by john (1st member of the corey coles fan club) on June 15, 2007 at 1:28 pm (#377865)

    How do they figure out BRAA?

  43. Gravatar
  44. Comment posted by Future on June 15, 2007 at 2:27 pm (#377894)

    BRAA: (From Hardball Times article by David Studeman)

    “Base Runs is similar to Runs Created in many ways, but it has a number of advantages for certain analyses. For instance, in this case I set the “run scoring multiplier” on a team-specific basis, and then calculated the team’s runs with and without each specific batter. In that way, my Base Runs totals equal the team’s total runs scored. That makes this analysis more consistent with last week’s article. By the way, it took ten Base Runs to “predict” one WPA, which is very consistent with other run/win ratios.”

  45. Comment posted by JamesSC on June 15, 2007 at 4:51 pm (#378007)

    These new fangled stats don’t do it for me, I am an old school stat fan.

    Interesting info though.

  46. Gravatar
  47. Comment posted by Future on June 15, 2007 at 5:01 pm (#378016)

    Thankfully the old school stat fans aren’t running the Mets. No offense, but on paper PLD looks decent, but he is completely stinking up the joint.

    Batting average is over-rated. The problem is today people look at .300 as some sort of sign that a hitter is good. That’s simply not true. Now that access to statistics are free, we don’t have the excuse that we don’t know better.

    By the way, case in point of a statistic that is still derided by some old school guys: WHIP. Shockingly, the statistic has existed since the late 70s and used to be in many newspapers as Hits+Walks per Nine.

    It’s crazy, but statistics don’t make the game more boring, but they do make it more interesting for how things work.

  48. Gravatar
  49. Comment posted by dptydwg420 on June 15, 2007 at 5:15 pm (#378026)

    Steve Trachsel and PLD are the ultimate battery!

    dem dere new-fangled stats be damned!!!

  50. Gravatar
  51. Comment posted by David Simons on June 15, 2007 at 5:59 pm (#378086)

    Trachsel has a winning record, he must be good. He’s a winner wherever he goes!

  52. Comment posted by Matt Gelb on June 18, 2007 at 1:45 am (#380450)

    Hey, I was the one who made this stat popular around these parts.

    But yes, the numbers tell a lot. Still, be careful: They don’t tell everything.

rss/syndication: