Worst Defense in 50 years

Home The Cardinal Nation Forums Open Forum Worst Defense in 50 years

Viewing 25 posts - 26 through 50 (of 59 total)
  • Author
    Posts
  • #73243
    Cardinals2016
    Participant

    Free

    I will add: another benefit of using Unearned Runs is that they are objective data points, not subjective

    #73245
    stlcard25
    Participant

    Paid - Annual

    I will add: another benefit of using Unearned Runs is that they are objective data points, not subjective

    I’d say that’s a stretch. Have you seen some of the plays that aren’t called errors anymore? You’re still dependent on the subjective opinion of an official scorer making the decision.

    #73249
    Cardinals2016
    Participant

    Free

    Well, true. But they are much more objective than comparing to a mythical “average” or “replacement” player.

    But yes, I was attempting to rationalize how teams could have more negative Defensive Runs Saved than Unearned Runs, and I was thinking about all the extra base hits given up by Jose Martinez playing first and not stretching for the ball, all the balls that drop when Fowler flashes his 4.4 second 18 yard speed and players that reach because Carpenter does not have all that extra oomph on his throws from 3B.

    Still doesn’t explain how the Red Sox have -31 DRS and only 39 Unearned Runs. With 5 Gold Glove nominees

    #73251
    Cardinals2016
    Participant

    Free

    I mean, if you want to think about it, by DRS the Red Sox only had 8 run saving plays the entire year with 5 Gold Glove nominees on the field. Bwahahahaha

    Except we know that the key to all these fancy metrics is a mythical “average” or “replacement” player.

    But carry on….

    #73263
    CariocaCardinal
    Participant

    Paid - Monthly

    Show me where the sum of Red Sox player’s bwar is significantly different than team wins!

    I could care less about gold gloves regarding defense. After Palmeiro won his for playing 28 games at 1B and 128 at DH in 1999 I have never given it a smitten of credibility.

    Unearned runs are objective only if you think errors are objective which we know they are not.

    I would add that plays that result in unearned runs are much fewer datapoints than you get with DRS.

    Thanks for the links to the DRS definitions but I have read those. That doesn’t mean I understand it.

    #73265
    Cardinals2016
    Participant

    Free

    Show me where the sum of Red Sox player’s bwar is significantly different than team wins!

    Ok:

    So, by batting they have 27.3 WAR Batting WAR

    by pitching they have 29.4 WAR Pitching

    27.3 + 29.4 = 56.7

    Let’s say the “average” team wins 82 games. 82 + 56.7 = 139 wins

    The Red Sox are at 118 Wins right now, correct? A 20 win difference from 118 is 17%. That is statistically significant (in Engineering, you will take less than 1% as statistically significant, in business, you generally take 10%).

    Thanks for the links to the DRS definitions but I have read those. That doesn’t mean I understand it.

    That is what they are counting on – people mindlessly regurgitating statistics that are spoon fed to them without having an understanding about the actual meaning or how they are calculated.

    I am not saying you are mindless. I respect you and can observe your thought process in your writings. But the first person that taught me about WAR was a college dropout who had had never taken above a 200 level in college.

    I have forgotten more about statistics than most people have ever learned.

    The metrics are inherently flawed, but they are regurgitated as the gospel.

    #73266
    Cardinals2016
    Participant

    Free

    Also note: DRS tells you they are a terrible team. WAR tells you they are a great team – even greater than they are. But people just whitewash these discrepancies.

    #73267
    stlcard25
    Participant

    Paid - Annual

    Let’s say the “average” team wins 82 games. 82 + 56.7 = 139 wins

    WAR is wins above replacement, not average. I can’t recall the exact number but the replacement team would be expected to win something like 47-50 games.

    #73278
    CariocaCardinal
    Participant

    Paid - Monthly

    Replacement level team is 55 wins (am almost sure but I will check that). Add that to the 56.7 and that gives them almost 113.

    These stats are only regular season only so the comparison total is 108 not 118.

    #73395
    Cardinals2016
    Participant

    Free

    Yes, you guys are correct – its like 47-48 wins for a WAR team. I still don’t know who those replacement level players are.

    Can anybody tell me:
    1. What does a replacement level 3B look like – what is his BA/OBP/SLG? How many errors does he commit? How hard does he throw it across the diamond and what % of players can he throw out? What is his reaction time off the bat? What is his range? Can he make the throws off-balance and bare-hand ground balls? What %?
    2. For SS & 2B, I’ll add in, how quickly can they get it out of their glove and pivot to turn a DP?
    3. For a 1B – how good is he at stretching that extra inch or so while staying on the bag? How good is he at stopping errant throws?
    4. For outfielders, how fast are they (hopefully faster than Fowler’s 4.4 second 18 yard dash)? How efficient do they run their routes? How many home runs do they rob? How many foul balls in the stands do they grab? How many outfield assists do they get?

    Now, for those that think DRS and WAR are more objective than errors and unearned runs:
    – How does the +/- get awarded? Is it an objective measurement, that can be repeatedly measured over time, or is a human making a judgment call, just like with errors?

    If you want to whitewash the Cardinals having +40 DRS while the Red Sox had -31, how do you explain the Phillies having -146 DRS while only committing 123 errors and giving up 63 unearned runs on the season? Do they have an entire team of Jose Martinez’ playing in the field? How did they win 80 games? Do you think all the home and away official game scorers conspired to suppress the true futility of the Phillies defense? Or do you think whoever was assigning defensive runs saved may be biased against the Phillies?

    Basically, I liken these advanced stats to how Enron kept their financials. Allegedly, Enron followed GAAP principles and had their books audited by independent consultants, yet they were exposed as a fraud years later and their house of cards came tumbling down.

    This obviously isn’t that magnitude, but you get the point. Similar to Enron’s “off the books” transactions, baseball analytics have all these calculations that nobody can understand, but everybody recites their prowess with certainty, even in the overwhelming evidence that the statistics are flawed.

    We all just shrug.

    #73400
    CariocaCardinal
    Participant

    Paid - Monthly

    As for the Phillies, I think most baseball people readily acknowledge that less range results in fewer errors.

    Harrison Bader’s 2018 OF fielding % .989

    Jose Martinez’s 2018 OF fielding % 1.000

    #73404
    jj-cf-stl
    Participant

    Paid - Annual

    2016, you’ve already demonstrated you lack a basic understanding of sabre metrics. That makes it extremely hard to converse about the topic with you.

    Some fans could care less for advanced metrics and I respect that position. But plz, as someone who has spent hours and hours and hours trying to self teach myself these metrics, realize there is also a sector of fans who would throw stones at it rather than throw hours at it.

    I hope you’re not a stone thrower because i enjoy your replies.
    Nothing replaces the eyeball test, especially not the D metrics, which are the worst of the bunch.
    For players we rarely get to see, the batting and pitching sabre stats are very helpful.

    #73574
    Cardinals2016
    Participant

    Free

    OK – let’s explain it this way:

    The Arizona Diamondbacks saved 157 runs according to DRS. They gave up 644 on the season. So, their replacement level team they were measured against would have given up 801 runs.

    The Cardinals saved 40 runs and gave up 691 on the season, so their replacement level team would have given up 731 runs.

    The Red Sox gave up 31 Defensive Runs, gave up 641 runs on the season, so their replacement level team would have only given up 610 runs.

    The Philadelphia Phillies gave up 146 Defensive Runs, and gave up 728 runs, so their replacement level team would have only allowed 582 runs.

    So, the Phillies played a replacement level team that allowed 582 runs and the Diamondbacks played a replacement level team that gave up 801 runs. Got it.

    #73575
    Cardinals2016
    Participant

    Free

    2016, you’ve already demonstrated you lack a basic understanding of sabre metrics.

    Here’s the deal. I have a master’s degree in economics. I’ve studied linear regression and econometrics. I tutored statistics in college.

    What’s your statistical background?

    #73593
    jj-cf-stl
    Participant

    Paid - Annual

    When you evaluated war as average instead of replacement level. The basics.

    Not a personal attack on you, but clearly not hours invested in the topic either.

    #73597
    Cardinals2016
    Participant

    Free

    You understand that in statistics you compare things to the mean right? Do you know why? So one of your baselines isn’t 582 runs and the other isn’t 801 runs. Talk about a moving target.

    You understand that only one MLB team allowed fewer than 582 runs this year, right? I’ll take a team of “replacement” players that only gives up 582 runs any day of the week.

    From a statistical point, WAR is junk science.

    Let’s come up with lame excuses why the absurdity of the data is valid.

    I’ll ask you again, what is your knowledge of statistics?

    #73599
    CariocaCardinal
    Participant

    Paid - Monthly

    2016, doesn’t your analysis of run differentials fail to take in to account differences in pitching staffs? Fielders can’t save runs that aren’t saveable.

    #73600
    PugsleyAddams
    Participant

    Paid - Annual

    I would love to weigh in on this heated debate, but I only received a C- in my frosh Stats class. And to make things worse, I was as proud as a peacock at the time to garner that grade. Carry on , gentlemen….carry on….I may forego watching tv with the wifey tonight and just grab a bag of popcorn and a grape soda and just keep an eye on this thread.

    #73625
    jj-cf-stl
    Participant

    Paid - Annual

    “WAR is a junk science”

    O well, enjoy your stone throwing.

    #73631
    bicyclemike
    Moderator

    Paid - Annual

    To weight in a bit, I know there was some talk about errors and fielding percentage not being as important to defense as other factors. And that is true, however we cannot just throw errors out of the mix either.

    Errors do not happen in a silo. In other words, they impact other parts of the game. By making your defense get an extra out, you extend the stress on your pitcher. As this keeps happening, the cumulative effect can be significant.

    Range works in the opposite direction. Making a play that would normally be a hit reduces the stress on pitching, and will especially benefit cumulatively. Poor range is just like errors, a drag on the pitching.

    #73639
    CariocaCardinal
    Participant

    Paid - Monthly

    Errors is part of the DRS computation I believe.

    #73832
    Cardinals2016
    Participant

    Free

    So, when I was in college, my forecasting professor told a joke:

    The owner of a company was facing a downsizing, so he had to let two employees go. He called three employees into his office, a statistician, an accountant and an economist. He told them he had to let two of them go. He didn’t know who to let go and who to keep, so he was going to ask them a question, and whoever gave the best answer would keep his job.

    He sent them all out to the hallway, and then called the statistician into his office. He asked, “what is 2 + 2?”

    The statistician said, “It’s somewhere in the range of 3 to 5.”

    He told the statistician it was a good answer, sent him out to the hallway, then called the accountant into the office. Again, he asked, “what is 2 + 2?”

    The accountant replied, “Four.”

    He told the accountant that was a good answer, then sent him out into the hallway.

    He then called in the economist. “I’m going to ask you the same question I asked the other two guys, what’s 2 + 2?”

    The economist looked around, closed the door, went over to the windows and drew the blinds, then whispered in his ear, “What do you want it to be?”

    I have also heard the joke where the accountant delivered the punch line.

    If you think about it, “what do you want it to be” is:
    – Enron for accountants
    – Theranos for making up your own measurements
    – WAR for baseball fans

    If I know what I need the numbers to be, I can make them do anything to get there. See Enron and Theranos for examples of companies that said, “What do you want it to be.”

    #73833
    Cardinals2016
    Participant

    Free

    On your first day of your first statistics class, you are taught that to do statistical analysis, the data has to be:
    – Objective
    – Observable
    – Measurable
    – Repeatable
    A “replacement” level player is not Objective – it is based on somebody’s recollection of what they think a replacement player is capable of

    A “replacement” level player is not Observable – he is made up

    A “replacement” level player is not measurable – he is not even in the data set

    A “replacement” level player is not repeatable, because you can’t repeat what you cannot observe

    Now, there are ways to correct for bias in statistical models, but when your analysis something that is made up, well…you get Theranos or Enron

    #73834
    Cardinals2016
    Participant

    Free

    Junk Science – rather than using accepted statistical norms, we’ll use our own that we make up. This way, when we publish numbers that are truly outlandish, nobody will be able to question them.

    Junk Science – instead of using data from observations, we’ll use a fictitious reference point nobody understands, is not defined and nobody can reference. PLUS, we’ll make it a moving target.

    Junk Science – we’ll make our “statistics” so complicated that nobody can possibly understand our statistics, nor replicate them. When anybody asks what a replacement level player is, we’ll call them stupid, as will all the people who have never studied statistics but know our statistical definitions inside and out.

    Junk Science – Even though regression models of more than 7 independent variables lose predictive capability due to the multicollinearity, bias and multiple sources of error from the variables, we’ll throw 30+ variables into our model. We’ll hide how truly bad our model is by never publishing it and using statements like, “Given the nature of the calculation and potential measurement errors, WAR should be used as a guide for separating groups of players and not as a precise estimate,” even though we never publish the +/- and the vast majority of our audience has never studied statistics and thus does not say “Player X was worth 2.5 WAR +/- 2 WAR.”

    Junk Science – Because our data with all of our shenanigans still isn’t meaningful, we’ll transform the data in ways only we understand, then regress it upon itself to make it meaningful. We’ll tell you that our regression of the data upon itself produces fantastic results, but then again, it always does. If you question these data transformations, we’ll call you stupid, as will all the statistically illiterate people for whom our statistics were created.

    #73835
    Cardinals2016
    Participant

    Free

    If you want WAR to have any statistical validity whatsoever, start replacing “Replacement” with “Average”

Viewing 25 posts - 26 through 50 (of 59 total)
  • You must be logged in to reply to this topic.