Announcement

Collapse
No announcement yet.

Progressing Toward Better Stats Thread

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

  • I think longevity is of some importance because some guys played in a style that let them last longer, and some guys just put in the work, or had the makeup to last longer. But granted Mays was on the decline after age 34. But its still 1000 games with a 138 OPS+ after age 34. Simply stated, not every great player can have a great decline. But if we use a high enough baseline like near WAA (I use the average of WAR and WAA, or a 40% baseline), then added seasons at a reduced level don't put a lot of extra wins on the table. See Ripken. We can credit longevity for Ripken, but see that it just does not add up to that much in WAA, at least not compared to before that. Mays had 94 WAA through age 35 and only added 16 in another 800+ games. It doesn't put him on another level, but it still makes a difference in finding the fine lines between closely ranked players. And when a player has a freak injury by something outside of just his playing style I factor it differently too.

    Comment


    • Here's Dante Bichette, 1993-1999, Colorado years.

      Home sluggin - .642
      Road sluggin -- .432



      Bichette1993-1999.jpg

      Comment


      • Originally posted by brett View Post
        I think longevity is of some importance because some guys played in a style that let them last longer, and some guys just put in the work, or had the makeup to last longer. But granted Mays was on the decline after age 34. But its still 1000 games with a 138 OPS+ after age 34. Simply stated, not every great player can have a great decline. But if we use a high enough baseline like near WAA (I use the average of WAR and WAA, or a 40% baseline), then added seasons at a reduced level don't put a lot of extra wins on the table. See Ripken. We can credit longevity for Ripken, but see that it just does not add up to that much in WAA, at least not compared to before that. Mays had 94 WAA through age 35 and only added 16 in another 800+ games. It doesn't put him on another level, but it still makes a difference in finding the fine lines between closely ranked players. And when a player has a freak injury by something outside of just his playing style I factor it differently too.
        That makes sense. Don't think there is a right or wrong, just preference.

        I'd like to do top 10 or so from each position. Oh, another thing I was gonna ask ya....what would the process be for pitchers' expected ERA+?

        Comment


        • Originally posted by Sultan_1895-1948 View Post
          That makes sense. Don't think there is a right or wrong, just preference.

          I'd like to do top 10 or so from each position. Oh, another thing I was gonna ask ya....what would the process be for pitchers' expected ERA+?
          Wow I never thought about that before. Of course pitchers may get a special benefit from their park, especially if they favor their handedness.

          OK here's the working model. Take road ERA. Divide it by 1.04. Divide that by the league's real ERA for the season.

          By the way, it looks like Puckett should take a big cut like 10+ OPS+ points when we use road numbers. This is part of a discussion on the Garvey thread. It's important because he's already a fringe hall of famer for me, and would likely make the decision clear if he takes that kind of cut.

          Do whatever you want with a handedness adjustment or for catchers, but please leave the column with just the road predicted. Then add another column if you want. Then you are being totally honest and letting people see the one that is purely objective, and the one that has an objective component.

          Also, would you like me to give each player a +/- wins score based on the road numbers? That could go into WARP, WAR or WAA. For example Dimaggio would be something like +11 road predicted wins? Anyway that's where I'm headed.

          I'd rather use the WAR format and then adjust out the things I don't like, changing the baseline, using different defensive metrics, road numbers, post season adjustment. I think that is the best way to get the objective rating I want and then I can adjust one last time based on things that can't be put into numbers, or that are out of our reach.

          Comment


          • Originally posted by brett View Post
            Wow I never thought about that before. Of course pitchers may get a special benefit from their park, especially if they favor their handedness.

            OK here's the working model. Take road ERA. Divide it by 1.04. Divide that by the league's real ERA for the season.

            By the way, it looks like Puckett should take a big cut like 10+ OPS+ points when we use road numbers. This is part of a discussion on the Garvey thread. It's important because he's already a fringe hall of famer for me, and would likely make the decision clear if he takes that kind of cut.

            Do whatever you want with a handedness adjustment or for catchers, but please leave the column with just the road predicted. Then add another column if you want. Then you are being totally honest and letting people see the one that is purely objective, and the one that has an objective component.

            Also, would you like me to give each player a +/- wins score based on the road numbers? That could go into WARP, WAR or WAA. For example Dimaggio would be something like +11 road predicted wins? Anyway that's where I'm headed.

            I'd rather use the WAR format and then adjust out the things I don't like, changing the baseline, using different defensive metrics, road numbers, post season adjustment. I think that is the best way to get the objective rating I want and then I can adjust one last time based on things that can't be put into numbers, or that are out of our reach.
            Ok so just for an example. Maddux's away ERA is 3.36 (2.96 at home)....so 3.36/1.04 = 3.230

            From there, I need to get the NL league ERA from 1986 to 2008, and THAT'S the final number?

            No, no Brett...I'm doing nothing with handed-ness, catchers, or anything in the chart. Those are set-in-stone numbers and it's completely up to the viewer to do with them, what they may.

            I was only explaining to you, my thought process after seeing the final number. I definitely think catchers should get something. I also think righties should get something, although probably less than I thought before, since you made good points on the issue.

            The big break-through for me personally, was realizing that modern players need something. Many factors which we've discussed ad nauseam but the biggest thing for me, is that the true sluggers cannot separate like they should be able to. Some of the reasons why, also impact them favorably, but they don't gain anywhere near the advantage that the pretenders do.

            Anyway, I'm not going to F with the chart at all, so no worries.

            By the way, did a few more players. Check Kingman gaining 2 OPS points (no thanks to his OBP) and there's also a log jam at 147 with Schmidt, FRobinson, Heilmann, and FThomas, only .02 points separating them all.

            Also Kiner and Mcgriff tied at 138.4. Had to go to the next decimal (.002) to break that tie. If there's any doubt gaps are up near the elites, and it gets tighter and tighter as you get further down...look no further.

            Oh my..had never looked at Puckett's splits before. I'd say def an 8-12 point drop. Hitting on carpet, never any wind, great hitters backdrop, margin for error oppo with the baggie...that'll do it.
            Last edited by Sultan_1895-1948; 06-04-2014, 08:38 PM.

            Comment


            • Ok

              Walter

              Home - 4.254 IP/ER
              Away - 3.139 IP/ER

              That's more than inning. Pretty significant?

              Grove

              Home - 2.918 IP/ER
              Away - 2.970 IP/ER

              We need to talk more about this pitcher thing.

              Let's focus on the hitters for now, yes?

              Comment


              • Originally posted by Sultan_1895-1948 View Post
                Ok so just for an example. Maddux's away ERA is 3.36 (2.96 at home)....so 3.36/1.04 = 3.230

                From there, I need to get the NL league ERA from 1986 to 2008, and THAT'S the final number?
                Yes, that would be right. Now it is important to realize that modern pitching metrics do a lot to adjust pitching based on defense, and also based on even the opponents that a pitcher faces (teams).

                for example, if you look at the WAR data on BBRef, you see that Maddux has a 3.56 RA9 (total runs allowed per 9 innings, both earned and unearned). His opposition averaged 4.5 runs per game in those seasons. That would make his ERA+ 4.5/3.56 x 100 or 126. Then they show that he also had 0.11 saved by his defenses. That would boost his ERA to 3.67 versus 4.5 which would be a 123 ERA+. If we further adjust that by the road to overall ratio that you calculated we get a 120 effective ERA+ for Maddux.

                Mike Mussina meanwhile Allowed a 3.98, facied opponents who averaged 4.94 (124 RA+) had -8 defensive support which would push him to 4.94/3.90 or 127, and his road predicted ERA+ is 3.63 compared to his actual of 3.68 which would make him effectively 129 BUT we don't know if he faced the same offensive opposition on the road, and defensive support may also be better at home.


                But for starters yea, Road ERA/1.04 times the league ERA.

                Then there is the question of whether pitchers can pitch to their defensive strengths.

                Comment


                • Ok, did a handful of guys at the bottom of the first post, pre- league ERA numbers.

                  So dividing away ERA by 1.04 is giving us expected OVERALL?

                  Then do what with that? We can't multiply by league ERA or we get a huge number.
                  Last edited by Sultan_1895-1948; 06-05-2014, 06:56 PM.

                  Comment


                  • Originally posted by Sultan_1895-1948 View Post
                    Ok, did a handful of guys at the bottom of the first post, pre- league ERA numbers.

                    So dividing away ERA by 1.04 is giving us expected OVERALL?

                    Then do what with that? We can't multiply by league ERA or we get a huge number.
                    How about WHIP+? Instead of only relying on ERA+ to directly compare pitchers, WHIP+ could be used as well. OR use a pitcher's road WHIP to get expected/road predicted WHIP, similar to what was done for the hitters. In fact, that might be a better idea.

                    Comment


                    • Originally posted by layson27 View Post
                      How about WHIP+? Instead of only relying on ERA+ to directly compare pitchers, WHIP+ could be used as well. OR use a pitcher's road WHIP to get expected/road predicted WHIP, similar to what was done for the hitters. In fact, that might be a better idea.
                      Not a bad idea.

                      Are you suggesting that expected relative ERA get added to expected relative WHIP the way SA and OPB were done, to create a final pitching number?

                      Comment


                      • Originally posted by Sultan_1895-1948 View Post
                        Are you suggesting that expected relative ERA get added to expected relative WHIP the way SA and OPB were done, to create a final pitching number?
                        No, but that's not a bad idea either. Makes sense.
                        Last edited by layson27; 06-05-2014, 08:36 PM.

                        Comment


                        • Originally posted by Sultan_1895-1948 View Post
                          Ok, did a handful of guys at the bottom of the first post, pre- league ERA numbers.

                          So dividing away ERA by 1.04 is giving us expected OVERALL?

                          Then do what with that? We can't multiply by league ERA or we get a huge number.
                          Divide the league ERA by the road ERA/1.04

                          Comment


                          • WHIP is a how stat and ERA+ is a what stat. ERA+ already captures all of the things WHIP cares about...and a bunch of things it misses. Adding them would be like adding Runs Created and hits. I don't see how this is beneficial.
                            1885 1886 1926 1931 1934 1942 1944 1946 1964 1967 1982 2006 2011

                            1887 1888 1928 1930 1943 1968 1985 1987 2004 2013

                            1996 2000 2001 2002 2005 2009 2012 2014 2015


                            The Top 100 Pitchers In MLB History
                            The Top 100 Position Players In MLB History

                            Comment


                            • Thank you for the input Bo. We need others to chime in.

                              I've only done Walter and Grove so far. Even though Walter is only 1914-'27, I feel more confident than ever..my original thoughts confirmed, about moving Big Train down. A huge discrepancy in adjusted ERA+.

                              I think top ten, sorting by ERA+ might be more interesting than the hitters.

                              Comment


                              • Originally posted by brett View Post
                                I realized that I can compare the players' outs to the league rate of outs to figure out how many extra plate appearances a player like Ruth "earned for himself". Rbat already has a cost value of making outs. What I want to account for is that the low out producer is getting some of his extra plate appearances because he is saving outs.

                                So let's work through the math for Babe Ruth 1921, which I chose simply because it is the highest Rbat total in history at 116.

                                Rbat: 116 (above average)
                                PAs: 693
                                Batting outs: 336 (I will not look at Sacs or GIDPs at this point as they may be situational and also were not always official or kept stats. So I just take AB-H)

                                League runs: 6296
                                League PAs: 48698
                                League outs: 30280

                                The next part I am going to have to figure out how to put in into an algebraic equation later, but using plugging and checking I got that Ruth's 693 PAs were 103% higher because he made few outs. The average player would have gotten just 675.

                                (Basically and to within about 1 PA, The average player got 1.61 PAs per out. Multiplying by 336 outs that Ruth made, they would have gotten 540 PAs in the same number of outs that he got 693, however, only 1/9 of his 153 extra PAs would have gone to himself, so an average player in Ruth's spot would have only gotten 674-675 PAs while he got 693).

                                In 675 PAs the average player produced 87.3 Rbat.

                                Ruth had 116.

                                (116+87.3)/87.3= 2.33, or 233%, strikingly similar to his OPS+ of 238
                                Reading back through that post again, what you did was a lot of work but pretty telling.

                                Could you do that using only road numbers? Maybe for a Sandberg and a Mathews, so we can see two extremes.

                                Comment

                                Ad Widget

                                Collapse
                                Working...
                                X