Announcement

Collapse
No announcement yet.

Request

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • White Knight
    replied
    Originally posted by GiambiJuice View Post
    Great work. Thank you.

    Stats like these seem to undermine WAR just a bit. Then again, a "replacement player" on the Yankees will generally be better than a replacement player on an average team.
    It seems to favor Jeter, although I will give you that the sample sizes are a lot different. Many more and growing for A-Rod, a lot less for Captain Healthy.

    Leave a comment:


  • Ubiquitous
    replied
    You're basically talking about WOWY (with or without you) which has been around for a very long time. Tango likes to use it a lot and I've used it often to try and measure a lot of the stuff that people think cannot be measured but somehow has a large effect on games.

    Leave a comment:


  • drstrangelove
    replied
    I thought about this a bit more. WAR is a non perfect tool that attempts to quantify what we know that affects team wins in general and covert it into a numeric measure. It has limitations and it's base assumptions are not that it's perfect but that it quantifies what is objectively quantifiable towards team wins without regard to timing. It can't and doesn't try to quantify non statistical performances or performance that relates to the timing of the performance (e.g., men on base, score, etc.)

    This question looks at it differently and asks what does each player do to affect the outcome in toto. Thus, everything. Even things like: the player is a good leader, or the runner 'causes' problems when he is on base, or clutch hitting, the catcher 'handles the pithcers well', all gets accounted for in this question.

    It just requires a very advanced sort of metrics that I don't think I've seen. It would be an amazing feat to develop this.
    Last edited by drstrangelove; 08-01-2012, 11:00 AM.

    Leave a comment:


  • Brooklyn
    replied
    Originally posted by Ubiquitous View Post
    In each player's split page is a section for what they did in wins and losses. You can do it by single season or career and you can get what they did each and every year on a single page.
    looks like I did it the hard way but glad to see the numbers matched

    Leave a comment:


  • Ubiquitous
    replied
    You can also go to BRef's PI and do game finder queries based on wins and losses going all the way back to 1918.

    Leave a comment:


  • Ubiquitous
    replied
    In each player's split page is a section for what they did in wins and losses. You can do it by single season or career and you can get what they did each and every year on a single page.

    Leave a comment:


  • Brooklyn
    replied
    Originally posted by Jackaroo Dave View Post
    In theory, if this query could be mechanized, it would be possible to repeat it for all players of interest, no? You could do team splits for the Yankees and every regular over a given era, for example?
    I do think it would be interesting, but my method was very manual - not practical to do on a large scale. I simply dumped the game logs from each year into excel, and added the team result up. See example for Jeter 2012, the team result is in column 8. The data obviously exists, at least for current seasons, so it can be done - but I don't have the data to do it

    Leave a comment:


  • drstrangelove
    replied
    What Shoeless said is correct. It's obvious that Reyes did not cause his team to win 14 extra games (a .092% gain), which would match the best year of any player in baseball history. It's also obvious that ARod and Jeter weren't replacement level players.

    It's not accurate as well to imagine that the games they played reflected the same 'skill' as the games they didn't play. Players who are on 10 game hitting streaks are never benched; players who have nagging injuries, are ultra tired or are in hitting slumps are. So it's natural that it looks like the team doesn't do much worse, because of course they wouldn't if you take a slumping tired player out and replace them with someone that has 80% of their base skill level.

    Moreover, consider Jeter. He has let's say 75 WAR in 2800 games. That's about 2.7%. In the 300 games he did not play that equals 8 wins. 8 wins in 300 games is well within the margin of error (i.e., luck.) Add in all the factors Shoeless mentioned plus selective 'sitting' of slumping players, and the luck factor alone is huge.

    I admit it would be interesting to see if we could get this parcelled out to give 'adjusted' wins, but this is a fairly raw.

    Leave a comment:


  • SHOELESSJOE3
    replied
    Originally posted by Brooklyn View Post
    Excluding 1995, since he only played 15 games in 1995 (all 2012 data through yesterday):

    Yankees with Jeter: 1508-1001. W%: 60.1%
    Yankees without Jeter: 109-72. W%: 60.2%

    Almost exactly the same

    If you want to include 2005:

    Yankees with Jeter: 1513-1011. W%: 59.9%
    Yankees without Jeter: 183-127. W%: 59.0%

    Again, almost exactly the same




    Agree this isn't conclusive, but over the course of his year he has missed more then a full season of games. The sample size is pretty good
    Agreed Brooklyn, big enough sample, more revealing, more accurate.
    I was speaking more of these broadcasters, talking about how a team did with a hitter out of the line up for a couple of weeks or so, that proves nothing. We need lots more info other than a regular out of the line up.

    Leave a comment:


  • Jackaroo Dave
    replied
    In theory, if this query could be mechanized, it would be possible to repeat it for all players of interest, no? You could do team splits for the Yankees and every regular over a given era, for example?

    This approach to player wins has always interested me, but I had no idea it could be actually be done. It's already proved very revealing, but of what, it would take a lot more studies to determine.

    BBREF already has a stat measuring the theoretical impact of an individual player on a .500 team; this could be an empirical check.

    I can see it's fraught with problems. With a player like Aaron, well, he's just not out of the lineup much, until he becomes a marginal player himself. And when a player is out of the lineup, the very adjustments the team may make will make the comparisons difficult. It's often not just popping a bench player in the star's slot. But maybe broad strokes viewed at a distance might provide some patterns.

    Leave a comment:


  • ipitch
    replied
    Originally posted by Brooklyn View Post
    Excluding 1995, since he only played 15 games in 1995 (all 2012 data through yesterday):
    Thanks much!

    Leave a comment:


  • Brooklyn
    replied
    Originally posted by ipitch View Post
    Good stuff. Can you do it for Jeter's entire career?
    Excluding 1995, since he only played 15 games in 1995 (all 2012 data through yesterday):

    Yankees with Jeter: 1508-1001. W%: 60.1%
    Yankees without Jeter: 109-72. W%: 60.2%

    Almost exactly the same

    If you want to include 2005:

    Yankees with Jeter: 1513-1011. W%: 59.9%
    Yankees without Jeter: 183-127. W%: 59.0%

    Again, almost exactly the same


    Originally posted by SHOELESSJOE3 View Post
    I wouldn't put too much into that if it's for a short time, a few weeks or more, how a team does with or without a certain player in the line up.

    You would have to know if there was any other player out for a short time, on that same team. What was the schedule like, was it a good home team playing more games at home during that period, were they mostly on the road at that time.

    What was the schedule during that time, playing the better teams, playing the lower level teams, playing a good team that was slumping at that time. Was the team your interested in, playing good, hitting good as a team.

    There are other factors to consider other than who was missing from the line up and how the team made out during that period, not that simple.
    Agree this isn't conclusive, but over the course of his year he has missed more then a full season of games. The sample size is pretty good

    Leave a comment:


  • Jackaroo Dave
    replied
    Originally posted by GiambiJuice View Post
    Great work. Thank you.

    Stats like these seem to undermine WAR just a bit. Then again, a "replacement player" on the Yankees will generally be better than a replacement player on an average team.
    Additionally, Brett pointed out in another thread that at the highest level, additional wins are harder to come by than at the league average. but if you gain less than average by adding a superstar at the top level, you'll lose less than average by taking him out of the lineup.

    Leave a comment:


  • SHOELESSJOE3
    replied
    I wouldn't put too much into that if it's for a short time, a few weeks or more, how a team does with or without a certain player in the line up.

    You would have to know if there was any other player out for a short time, on that same team. What was the schedule like, was it a good home team playing more games at home during that period, were they mostly on the road at that time.

    What was the schedule during that time, playing the better teams, playing the lower level teams, playing a good team that was slumping at that time. Was the team your interested in, playing good, hitting good as a team.

    There are other factors to consider other than who was missing from the line up and how the team made out during that period, not that simple.
    Last edited by SHOELESSJOE3; 07-31-2012, 04:23 AM.

    Leave a comment:


  • Brooklyn
    replied
    Originally posted by ipitch View Post
    Good stuff. Can you do it for Jeter's entire career?
    give me some time. I should have saved my original work, it would have been easier to build off of it

    Leave a comment:

Ad Widget

Collapse
Working...
X