View RSS Feed

jorgenswest

Time to Drop the Numbers and Focus on WAR

Rate this Entry
In the current era of baseball it doesn't make sense to quantify pitchers as #X starter. It used to be that #1 or #2 starters would get more starts because they would skip over guys at the back end of the rotation. I think Verlander and the Tigers are the only team that did that with any consistency.

Virtually all teams roll their starters so that all slots get about the same number of starts. Even when the opportunity at the all star break comes to skip some starts at the back end, those opportunities aren't often taken. Last year the Twins took the opportunity to give Diamond a longer rest at the all star break taking away a start rather than getting an extra.

It is more important to look at the staff's overall contribution in terms of WAR as a whole. Instead of a #4 starter, it would be better to quantify a starter as a 2 WAR starter or a 1 WAR starter. The #4 no longer has any value. The pitcher is not going to get their starts skipped.

Note: Top starters in the playoffs often do get an extra start and rotations are dropped to 4 pitchers.

There were 2268 starts last year in the AL. The ERA for those starts was 4.40. That included 95 pitches in 5.9 innings. Looking at AL pitchers who pitched the full season and had that profile you will find the league average pitcher contributes just less than 2 WAR over a full season.

If the Twins can get 5 guys to give them 2 WAR they will far exceed the total of 3 WAR the starters earned last year.

Let's look at the Twins recent acquisitions.

According to fangraphs, from 2008-2011 Mike Pelfrey had 8.1 WAR over 4 seasons. If he can do 2 WAR for the Twins he will be well worth the contract.

Vance Worley had 4.3 WAR over the last two seasons. It seems reasonable to project 2 WAR from him.

Kevin Correia was worth 0.9 WAR last year and that was his best season in the last three. At his age, it doesn't seem likely he will produce more than 1 WAR. If he does achieve 1 WAR, he will be worth about the
level of his contract.

Of the returning staff, only Scott Diamond had more than 2 WAR (2.6) as a starter.

Instead of debating pitcher number, the Twins should be striving towards putting together a league average pitching staff that contributes about 10 WAR. That alone is a 7 game improvement.

I used fangraphs WAR calculation for this article.



Updated 12-17-2012 at 10:16 PM by jorgenswest

Tags: None Add / Edit Tags
Categories
Uncategorized

Comments

  1. old nurse's Avatar
    I think it is a nice idea except for the fact that there isn't a correlation between WAR and actual wins.
  2. Todd G's Avatar
    Just to get the obligatory statement out of the way so we can continue with an informed discussion.

    "WAR, what is it good for?"

    Carry on.
  3. jay's Avatar
    Quote Originally Posted by old nurse
    I think it is a nice idea except for the fact that there isn't a correlation between WAR and actual wins.
    Excellent sarcasm.

    For the true non-believers, the total WAR accumulated by a team has had a correlation to actual wins between .83 and .89 over the last few seasons.

    It includes some variables that are not perfectly measured (ie - defense) and can't account for real-life variation caused by the context in which contributions are provided (ie - winning an abnormal amount of close games or "clutch" and "luck"). It may not be exactly perfect, but it is extremely useful.
  4. Shane Wahl's Avatar
    I think an average of 1 is the current situation if those five were involved (Correia and Pelfrey may combine for 1). Marcum could be worth 2.5 and maybe Hendriks/Gibson can produce 2. That way the 10 WAR might be reachable.
  5. Shane Wahl's Avatar
    Yes, thank you jay for pointing that out.
  6. Winston Smith's Avatar
    You used AL numbers for your base and all 3 of these new guys are NL guys. Is that apples to apples?
  7. Jim H's Avatar
    While I tend to agree with Old Nurse on WAR, I think even you believers in WAR are missing something here. It is very hard to predict what pitchers will do. WAR or all pitching stats really, are not predictive. The Twins staff in 2010 was good enough(I have no idea what their collective WAR was) to be major contributors to a 95 win team. Virtually the same staff was considerably poorer in 2011 and contributed almost nothing in 2012.

    Also consider, last year at this time, nobody was predicting(including WAR) what Diamond went on to produce in 2012. What Ryan is trying to do is put together a staff that by any measure(use WAR if you like) has some chance of being league average. He is well aware that some of those will fail, which is why there are so many options. Now, this stragedy has a fair chance of failing, I think Ryan has acknowledged this. Part of the reason for this is that there is little chance of the assembled staff to be better than league average. If that is your upside......

    I am supportive of this stragedy because I don't think that by buying the higher priced, longer termed options actually much increases the staff's chances of being better than league average. Whether that is true or not, Ryan has acquired some higher upside pitching options, when added to the potentially higher sided options already in the system, gives a great deal of hope for the future.

    So getting back to the topic presented. Ryan is assembling a group of starters for 2012 that we all hope can be league average. He is avoiding any long committment that might inhibit assembling an above average staff in the future. Given where things were, I believe that is the right path.
  8. Clyde's Avatar
    Quote Originally Posted by Jim H
    While I tend to agree with Old Nurse on WAR, I think even you believers in WAR are missing something here. It is very hard to predict what pitchers will do. WAR or all pitching stats really, are not predictive. The Twins staff in 2010 was good enough(I have no idea what their collective WAR was) to be major contributors to a 95 win team. Virtually the same staff was considerably poorer in 2011 and contributed almost nothing in 2012.


    .
    fwar 2010 was14.5
    2012 was 8.1
  9. Clyde's Avatar
    You still need to build a team. What the team should be is a collection that is better than the sum of it's parts. The 2010 Twins did not have a lot of flash but they made their way to 94 wins despite any metric that said they should not have. WAR for pitchers is based on fip. The fangraphs site says that fip will not predict play so therefore WAR will not either. WAR will say that a player is good, but not always how bad they are. See Blackburn's -0.8 WAR
  10. jorgenswest's Avatar
    As for the NL comment...

    I don't think any NL dependent numbers were quoted for the new players. WAR is context, league and year neutral. Players can be compared across teams, leagues and seasons.

    As for using WAR...

    It may not be te best measure. Debating whether someone is a #3 or #4 starter is not a measure at all. Choose another neutral measure to compare pitchers across leagues and seasons.
  11. Clyde's Avatar
    Debating if a pitcher is a 1, 2, 3, 4, or 5 is completely meaningless and pointless debate. People want to quantify the starters. A 5 is replacement level. a 4 is a war under two, a three is under three, a two is under 4 and a nuber one is anything above. Everybody can now be happy with meaningless numbers,. Wether you use a 1-5 designation or a number based on a non predictive statistic, it doesn't change how good the pitcher is. It does not measure wear and tear or changing abilities due to learning how to pitch. At best you can say what was.
  12. old nurse's Avatar
    To me, the concept of WAR is not bad, the method of the computation is the problem.

    Heyman had to say this on WAR

    Jamey Carroll has a better WAR than Derek Jeter, 2.6 to 2.5. That's absurd.


    And Denard Span has a better WAR than Josh Hamilton, 4.0 to 3.7. Even more ridiculous.
    And Heymen supports the concept
  13. jorgenswest's Avatar
    Quote Originally Posted by old nurse
    To me, the concept of WAR is not bad, the method of the computation is the problem.

    Heyman had to say this on WAR

    Jamey Carroll has a better WAR than Derek Jeter, 2.6 to 2.5. That's absurd.


    And Denard Span has a better WAR than Josh Hamilton, 4.0 to 3.7. Even more ridiculous.
    And Heymen supports the concept
    Do those numbers include both offense and defense?
  14. old nurse's Avatar
    I believe that Heyman used the numbers posted on fangraphs. Having a problem with the inaccuracies of WAR?
  15. jay's Avatar
    Quote Originally Posted by old nurse
    To me, the concept of WAR is not bad, the method of the computation is the problem.

    Heyman had to say this on WAR

    Jamey Carroll has a better WAR than Derek Jeter, 2.6 to 2.5. That's absurd.


    And Denard Span has a better WAR than Josh Hamilton, 4.0 to 3.7. Even more ridiculous.
    And Heymen supports the concept
    In both of those cases, it's the defensive valuations that are causing that. They both play defensively-demanding positions that can generate a lot of value (or negative value) in WAR calculations. WAR is hammering Jeter on his range and Hamilton on a bit of everything.

    It's hard/impossible to defend something like that, especially given the remaining inaccuracies in defensive measurements. However, I will say that it raises awareness to the fact that we commonly value run production over run prevention.
  16. old nurse's Avatar
    Jay, as bad as people want to quantify things, some things you can't without subjectivity. When there is subjectivity, there is error.
  17. old nurse's Avatar
    Since the thread was on pitching. Fangraphs defines a good to great pitcher as one with a WAR above 3. There was only 35 of them last year in the majors, 46 the year before. A solid starter as one with a WAR of 2-3. There was only 36 of those last year, 21 the year before. Given those numbers are from fangraphs they should meet your standard. Random clicking on the names on the solid range of players reveals that either these pitchers are really inconsistent as their WAR score fluctuates widely, or there is an issue with the metric.
©2014 TwinsCentric, LLC. All Rights Reserved.
Interested in advertising with Twins Daily? Click here.