Announcement

Collapse
No announcement yet.

Progressing Toward Better Stats Thread

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

  • #31
    Here's Double X

    Foxx-sOPS+.jpg

    League OPS+ file.....

    Brett, you weren't kiddin' about the league OPS+ not changin' much. Just seems like it's not fully accounting for run environment. Heck, the '68 NL is the same as '28 NL.

    Can we use something like innings pitched per earned run instead? Looking at that, 1968 would show to have 3.01 IP/ER and 1928 NL would be 2.25 IP/ER. Just a thought.
    Attached Files

    Comment


    • #32
      Originally posted by Sultan_1895-1948 View Post
      Here's Double X

      [ATTACH]138351[/ATTACH]

      League OPS+ file.....

      Brett, you weren't kiddin' about the league OPS+ not changin' much. Just seems like it's not fully accounting for run environment. Heck, the '68 NL is the same as '28 NL.

      Can we use something like innings pitched per earned run instead? Looking at that, 1968 would show to have 3.01 IP/ER and 1928 NL would be 2.25 IP/ER. Just a thought.

      OPS+ is set at 100 each season. the only problem is that sOPS+ keeps pitchers in the league numbers, but traditionally we take pitchers out. The .94 just takes them back out. The .94 should not change because of run environment, it just says that removing pitchers raises the league production by about 6% per plate appearance. It is only a measure of pitchers batting effect on hitting, not run setting.

      Comment


      • #33
        Originally posted by brett View Post
        OPS+ is set at 100 each season. the only problem is that sOPS+ keeps pitchers in the league numbers, but traditionally we take pitchers out. The .94 just takes them back out. The .94 should not change because of run environment, it just says that removing pitchers raises the league production by about 6% per plate appearance. It is only a measure of pitchers batting effect on hitting, not run setting.
        Just like there's different levels of position hitters, there's different levels of hitting pitchers. Some get bunts down more often, move runners along, put the ball in play more often, etc...but you know that.

        So..its' fine to remove pitchers. Shouldn't there be a step AFTER that though? Let's remove pitchers, and THEN adjust to run environment. There's no reason why the 1927 AL should have the same league OPS+ as the '59 and '60 AL.

        Comment


        • #34
          Originally posted by Sultan_1895-1948 View Post
          Just like there's different levels of position hitters, there's different levels of hitting pitchers. Some get bunts down more often, move runners along, put the ball in play more often, etc...but you know that.

          So..its' fine to remove pitchers. Shouldn't there be a step AFTER that though? Let's remove pitchers, and THEN adjust to run environment. There's no reason why the 1927 AL should have the same league OPS+ as the '59 and '60 AL.
          Sultan, OPS+ has already been adjusted to the league run environment. Ruth's road sOPS+ scores, and overall OPS+ has already been done relative to run environment. Same with someone in 1968. The 93, 94, 95 is the effect of pitchers, and yes it is a little different from year to year.

          Comment


          • #35
            Originally posted by brett View Post
            Sultan, OPS+ has already been adjusted to the league run environment. Ruth's road sOPS+ scores, and overall OPS+ has already been done relative to run environment. Same with someone in 1968. The 93, 94, 95 is the effect of pitchers, and yes it is a little different from year to year.
            If it's already been fully adjusted, then why do we constantly see people still adjusting for era? They adjust even after seeing these numbers that are already adjusted. I'm trying to get to a place where we've adjusted all the way, and then some.

            Comment


            • #36
              Originally posted by Bothrops Atrox View Post
              Yes. Any issues somebody has with how WAR deals with position, defense, ANYTHING will be the same issues with WAA. The ONLY difference is the comparison baseline. ONLY. . You can take the same formula and compare to any baseline you want to. Wins Above Good. Wins Above All Star. Wins Above Cobb. Whatever you want.
              Damn, that would be ugly. There would be a whole mess of negative numbers all over the place!
              "It ain't braggin' if you can do it." Dizzy Dean

              Comment


              • #37
                Originally posted by Herr28 View Post
                Damn, that would be ugly. There would be a whole mess of negative numbers all over the place!
                And I am sure the subtraction would be doubled if the player was a first baseman.
                Your Second Base Coach
                Garvey, Lopes, Russell, and Cey started 833 times and the Dodgers went 498-335, for a .598 winning percentage. That’s equal to a team going 97-65 over a season. On those occasions when at least one of them missed his start, the Dodgers were 306-267-1, which is a .534 clip. That works out to a team going 87-75. So having all four of them added 10 wins to the Dodgers per year.
                http://www.youtube.com/watch?v=p5hCIvMule0

                Comment


                • #38
                  So, I have the PA, HR and sOPS+ H/A split data for most of the RH batters for 1936-51 for the Yankees. (I only tabulated for position players with 100+ PA in a season.) A few pitcher seasons (Chandler and Ruffing) were left off plus all the RH seasons where a batter had less than 100 PA. There was one player in 1950 who split time with the Yanks and someone else, and with just ~160 PA, I just left him off.

                  Code:
                  Home                    
                               PA   HR   sOPS+  
                  DiMaggio    3,790 148  141.6       
                  All other  16,047 234   97.4


                  Code:
                  Away                    
                               PA    HR   sOPS+  
                  DiMaggio    3,883  213  171.5       
                  All other  17,322  307  104.1

                  What I noticed were two things:

                  1) YS did not hurt "all other" RH hitters for HR as much as Joe D. While "all other" hit 31% more total HR on the road, Joe D was at 44%. However, batters on winning teams, usually bat more on the road, so totals are misleading. In fact, once you adjust for the H/A splits in PA, "All other" averaged 21.5% more HR per PA on the road (68.6 vs 56.4), but Joe D averaged 40.5% more (25.6 vs 18.2)

                  2) YS did not have anywhere near the OPS+ impact on "all other" as on DiMaggio. "All other" was 6.8% better on the road, but Joe D was 21.1% better.


                  This leads to more questions of course. Why was DiMaggio in particular hurt while others were hurt far less? Were others making adjustments? Was YS only hurting one type of hitter? (Joe Gordon, a legitimate RH power hitter, had a road HR rate per PA that was only 8.6% higher than his YS rate----28.8 vs 26.5----so he was affected even less than the remainder of the "All Other" group.)


                  As a note, I examined the 1940-42 splits and it looks more like a .93 adjustment for pitchers. Also, I have a correction of the 1947 road PAs.
                  Last edited by drstrangelove; 05-04-2014, 02:39 AM.
                  "It's better to look good, than be good."

                  Comment


                  • #39
                    Originally posted by drstrangelove View Post
                    So, I have the PA, HR and sOPS+ H/A split data for most of the RH batters for 1936-51 for the Yankees. (I only tabulated for position players with 100+ PA in a season.) A few pitcher seasons (Chandler and Ruffing) were left off plus all the RH seasons where a batter had less than 100 PA. There was one player in 1950 who split time with the Yanks and someone else, and with just ~160 PA, I just left him off.

                    Code:
                    Home                    
                                 PA   HR   sOPS+  
                    DiMaggio    3,790 148  141.6       
                    All other  16,047 234   97.4


                    Code:
                    Away                    
                                 PA    HR   sOPS+  
                    DiMaggio    3,883  213  171.5       
                    All other  17,322  307  104.1

                    What I noticed were two things:

                    1) YS did not hurt "all other" RH hitters for HR as much as Joe D. While "all other" hit 31% more total HR on the road, Joe D was at 44%. However, batters on winning teams, usually bat more on the road, so totals are misleading. In fact, once you adjust for the H/A splits in PA, "All other" averaged 21.5% more HR per PA on the road (68.6 vs 56.4), but Joe D averaged 40.5% more (25.6 vs 18.2)

                    2) YS did not have anywhere near the OPS+ impact on "all other" as on DiMaggio. "All other" was 6.8% better on the road, but Joe D was 21.1% better.


                    This leads to more questions of course. Why was DiMaggio in particular hurt while others were hurt far less? Were others making adjustments? Was YS only hurting one type of hitter? (Joe Gordon, a legitimate RH power hitter, had a road HR rate per PA that was only 8.6% higher than his YS rate----28.8 vs 26.5----so he was affected even less than the remainder of the "All Other" group.)


                    As a note, I examined the 1940-42 splits and it looks more like a .93 adjustment for pitchers. Also, I have a correction of the 1947 road PAs.

                    That is interesting, but also keep in mind that since home runs make up a larger percentage of Dimaggio's OPS+, it can effect his OPS+ even more. If a 30 home run guy is reduced to 24 its a greater affect than if a 10 home run guy is reduced to 8. Also, Dimaggio hit a LOT more triples at home, so it looks to me like he was having a margin of his home runs removed based on distance breakdowns and not necessarily on approach (at least as much). If we had the average guys as a group hitting a few IPHRs over the outfielders head, while perhaps they just played deeper on Dimaggio, it could matter.

                    Do you have any idea how many of the home runs hit in YS were IPHRs also?

                    Comment


                    • #40
                      Originally posted by brett View Post
                      That is interesting, but also keep in mind that since home runs make up a larger percentage of Dimaggio's OPS+, it can effect his OPS+ even more. If a 30 home run guy is reduced to 24 its a greater affect than if a 10 home run guy is reduced to 8. Also, Dimaggio hit a LOT more triples at home, so it looks to me like he was having a margin of his home runs removed based on distance breakdowns and not necessarily on approach (at least as much). If we had the average guys as a group hitting a few IPHRs over the outfielders head, while perhaps they just played deeper on Dimaggio, it could matter.

                      Do you have any idea how many of the home runs hit in YS were IPHRs also?
                      There is no breakdown on BBRef for IPHR by season, and a lot of the players split time with other teams. But if you allocate their career IPHR as a proportion based upon the % of HR hit at home in 1936-51 in my list to total HR hit in their career, it adds up to 13 YS IPHR by the "all other" group.

                      There's another way to do it of course. What is the most possible if YS created an insane amount more than any other stadium?

                      a) Lazzeri hit 8 IPHR, but he hit 98% of his HR as a Yankee. Allocating 1 to him makes sense since 1936-37 was about 1/8th of all his Yankee HR.
                      b) Bauer hit 7 IPHR, but he hit 96% of his HR as a Yankee. Allocating no more than 2 to him makes sense since 1949-51 was about 1/5th of all his Yankee HR.
                      c) Chapman had 15 IPHR, but he only hit 1 HR as a Yankee in YS, so the most he could have is 1.
                      d) Let's make all of Crosetti's career IPHR at YS, so that's 4.
                      e) Let's do the same for Powell (1), Glenn (1), Dahlgren (3), Gordon (1), Steinweiss (2), Hemsley (1). Total is 9.

                      So the most it could probably be is 17. So the total IPHR by RH at YS in 1936-51 (excluding DiMaggio) is probably in the 13-17 range. So that is 7% of all the HR not hit by DiMaggio in 1936-51.

                      If you take Dimaggio's 3 IPHR and assume all were at YS, and add that to half of his "extra" triples, you get 3+8, or 11. If all those "would" have been IPHR, you get 11 IPHR, and then 148+8 HR at YS. In that case, 11/156 is 7%. So, although perhaps coincidental, it's possible that either a faster runner, or a less deep outfielder could have matched the same IPHR rate.

                      There's a speed factor as well, since depending on actual raw speed, one player might get farther than another player, even if the ball went just as far over an outfielder's head.

                      Until we get location data for players in that period, there's no way to know if players changed their approach.
                      Last edited by drstrangelove; 05-04-2014, 03:30 PM.
                      "It's better to look good, than be good."

                      Comment


                      • #41
                        I was curious to see how this sOPS+ thing would handle Piazza, who has a 91 to 108 home/road tOPS.

                        The home/road edge is significantly reduced. Guess you were right Brett, when you said tOPS has serious flaws.

                        If the expected home advantage here is 2% (or whatever it is), then how do we handle a 3.71% road edge?
                        Attached Files
                        Last edited by Sultan_1895-1948; 05-04-2014, 05:53 PM.

                        Comment


                        • #42
                          Originally posted by Brett
                          In other words, when road OPS+ is higher than home OPS+ it strongly suggests that the player SPECIFICALLY benefitted in relative production from his home park.
                          Did you mean, benefitted from the road parks?

                          Comment


                          • #43
                            Bench(sOPS+).jpg
                            -----------------------------------

                            Comment


                            • #44
                              Originally posted by Sultan_1895-1948 View Post
                              I was curious to see how this sOPS+ thing would handle Piazza, who has a 91 to 108 home/road tOPS.

                              The home/road edge is significantly reduced. Guess you were right Brett, when you said tOPS has serious flaws.

                              If the expected home advantage here is 2% (or whatever it is), then how do we handle a 3.71% road edge?
                              Yea. He had a 91 home tOPS+ because his home park was a low offensive park, but he was only hurt a little more than the average player by that park.

                              Comment


                              • #45
                                Originally posted by Sultan_1895-1948 View Post
                                Did you mean, benefitted from the road parks?
                                Sort of. I meant to say HURT by his HOME park.

                                Comment

                                Ad Widget

                                Collapse
                                Working...
                                X