Announcement

Collapse
No announcement yet.

Progressing Toward Better Stats Thread

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

  • Ok here's the guys I've done so far. Haven't transferred to chart mode yet.

    This is sorted by DNRA++ but that's before IP is taken into account, so maybe it shouldn't be? DNRA++ is a rate stat right?

    Check out who comes in at #34 just behind Spahn. Mr. Bambino himself. Yes, just 1221 innings, so the numbers other than DNRA++ are affected, and rightfully so. But I was curious if he put up the same rates over 3500 IP, which isn't a lot to ask imo, assuming he stayed pitching full-time for '18 and '19. He was still developing. Still honing a changeup actually and would have become better individually AND relatively imo. Anyway, it's only 3500 IP, but WAA comes to 28.680 and NRS 223.12, right behind Marichal for #32 if sorted by NRS.

    Ok enough of that. Actually wanted to ask you something, because this is going to come up again. The last guy on here, Darling. Not sure how to calculate a negative number. Don't think it's working out right. I'm getting -6.031 for WAA and then 125.096 for NRS. Is that right?

    DNRA++chartpic.jpg

    Comment


    • yea it looks like Darling was slightly below average using the league adjusted ERA as the baseline. If you put an = sign before the numbers it should kick out an answer I think.

      I think that if we have Hershizer, Saberhagen, Guidry, Hudson and maybe Appier would be similar. Really puts Schilling in rare company too.
      Last edited by brett; 06-25-2014, 04:48 PM.

      Comment


      • That's a great list you have there, Sultan! I am especially happy with #8.
        "It ain't braggin' if you can do it." Dizzy Dean

        Comment


        • Ha...yeah dude was great. Just didn't have the IP of some others and you can see how that affects his NRSWins.

          I suspect he's that high, in part, because of the run environment he pitched in. And he'll probably get slightly dinged when road numbers are factored in. This is just the first step toward a great pitching stat I hope.

          From what I've seen in our previous work, I'm seriously considering Seaver as a top 5-7 pitcher all time. Not sure about Dizzy until I see more. I do value peak over longevity so we'll see.
          Last edited by Sultan_1895-1948; 06-26-2014, 10:16 PM.

          Comment


          • Ok, need some quick feedback.

            The charts in the first post were all lined up. Had to restore my entire comp due to virus bs. Now everything is smaller and they don't line up.

            I want them to lineup on zoom 100% and font size medium for google chrome. Do they lineup for anyone? And if so, what settings are you on?

            Comment


            • Still working on the league SLGeff numbers but for now, these guys are done. Multiplied by 100 at the end to get an OPS+ type number.

              Incredible how well Cobb holds up to the boppers despite 52% of his total bases coming by way of singles.

              To put that into perspective, Foxx is at 31%, Gehrig is at 30%.

              So what else can we do here Brett? Does Rbat finally sneak into our work?

              Code:
                                           
                            .TB / (AB - H)   ..HOME     AWAY    (AWAY * 1.04)             (EXP SLGeff / LG SLGeff) * 100 
                              SLGeff      SLGeff    ..SLGeff   EXP OVR SLGeff   LG SLGeff      ADJ REL SLGeff
              Code:
              Ruth            1.048     1.068    1.029      1.070       .5245          204.0
              
              TWilliams       .9667     1.020    .9143      .9508       .5174          183.7     
              
              Gehrig          .9583     .9236    .9921      1.031       .5620          183.4
              
              Cobb            .8080     .7821    .8221      .8549       .4776          178.9
              
              JDiMaggio       .8570     .7983    .9154      .9520       .5354          177.8
              
              Hornsby         .8987     .9193    .8790      .9146       .5268          173.6
              
              BaBonds         .8645     .8833    .8471      .8809       .5459          161.3

              Comment


              • Originally posted by Sultan_1895-1948 View Post
                Ok, need some quick feedback.

                The charts in the first post were all lined up. Had to restore my entire comp due to virus bs. Now everything is smaller and they don't line up.

                I want them to lineup on zoom 100% and font size medium for google chrome. Do they lineup for anyone? And if so, what settings are you on?
                Everything is looking great, all lined up, for me. The second table, with the stats you have posted above here, gets a little messy after Bonds. Just a little, with some of the columns not falling in line, but that is the only one. The pitchers are all dressed up and still looking great, too.
                "It ain't braggin' if you can do it." Dizzy Dean

                Comment


                • Originally posted by Herr28 View Post
                  Everything is looking great, all lined up, for me. The second table, with the stats you have posted above here, gets a little messy after Bonds. Just a little, with some of the columns not falling in line, but that is the only one. The pitchers are all dressed up and still looking great, too.
                  Thanks for the feedback Herr.

                  Yeah, in the first post, the alignment won't be correct unless their full line is complete. Only thing missing is the UNDER CONTRUCTION highlighted flash but it's work in progress

                  Recently re-aligned with new settings, so it's good to know they're lining up.

                  Before I had to restore my computer, I think I was on 110% letter zoom or something. Everything is smaller now but probably that's how it should be with the high screen res.

                  A little somethin' for ya. Check him goin' upper tank at the 21 sec mark

                  Last edited by Sultan_1895-1948; 07-07-2014, 06:32 PM.

                  Comment


                  • SORTED BY DNRA++

                    DNRA++sort.jpg

                    This seems kind of similar to what Matt used to to with the GI Method, breaking things down by mastery and longevity.

                    The list changes dramatically when IP are factored in, as it should. But when ranking pure greatness, the scope should narrow.

                    I'm wondering about a chart that shows where they were at certain IP. Maybe 500-1000, 1001-2000, 2001-3000, 3001, 4000, and after that, just 4001+ (40 guys have done that).

                    I don't know just throwing out stuff. Would be sorta like my 5/3 chart, showing how great they were for a short time before falling off, holding on too long to decline drastically, or who was just pretty good for a long, long time.
                    Last edited by Sultan_1895-1948; 07-08-2014, 08:29 PM.

                    Comment


                    • So we're using 10,000 PA to add some weight. Anything over 10,000 and the player gets "extra" credit.

                      What I'm wondering, is that too high? Only 80 players in history reached 10,000 PA.

                      If you look at the top 500 in PA, the average is 8454 PA.

                      If you look at the top 400 in PA, the average is 8869 PA.

                      If you look at the top 300 in PA, the average is 9357 PA.

                      Would one of those numbers make more sense?

                      Comment


                      • Originally posted by Sultan_1895-1948 View Post
                        So we're using 10,000 PA to add some weight. Anything over 10,000 and the player gets "extra" credit.

                        What I'm wondering, is that too high? Only 80 players in history reached 10,000 PA.

                        If you look at the top 500 in PA, the average is 8454 PA.

                        If you look at the top 400 in PA, the average is 8869 PA.

                        If you look at the top 300 in PA, the average is 9357 PA.

                        Would one of those numbers make more sense?

                        I picked it because it is almost exactly 15 full seasons, and I would think that 10 is at the short extreme for great players. I think it is right Sultan, or a good choice given that 80 players have reached that level because that would probably be 80 of the 160 best position players of all time, about half of them. But it doesn't really matter where you set it. If you set it at 8000 you just bring everyone up. Also, 10,000 PAS (about 15 full seasons) is half way between a minimal 10 full seasons and 20 full seasons which is basically the top extreme. If you wanted to rate everyone though, including the average joe players, you might want to use an equivalent of just 10 full seasons, or maybe 6500-7000 PAs.

                        Comment


                        • -----------------------------
                          Last edited by Sultan_1895-1948; 07-19-2014, 01:24 PM.

                          Comment


                          • That makes sense.

                            Did about a dozen players in the first post and just seemed extreme. Do you really think a 6500 PA guy would drop 15 OPS points had he reached 10,000? Or for example Mize was 2630 PA shy of 10,000 and his OPS drops 14 points. Is that reasonable?

                            Then there's the war issue. For some guys like Ted and Joe I thought about posting the raw result...but then calculate them with assumed 600 PA per missed year in parenthesis.

                            Then again, that assumes the OPS wouldn't change.

                            Either way, the last column is there to view and people can use their own judgement.

                            Here's a thought. Going over 10k is certainly impressive and credit should be given. Is there a way to keep THAT bonus the same, while at the same time lowering the hit a player takes for being under 10k?

                            Example.

                            Ott's had 11,348 PA which is 1348 more than 10k. It gets him from 151 to 158. I think that's reasonable.

                            Let's take 10k - 1348 to end up with 8652. If he had that number of PA (same difference), that would take his 151 DOWN to 131.

                            So it would ding him 20 points for not getting to 10k, and the bonus after 10k is only 7 points. I'm not suggesting the bonus be lower, just that the ding not be as harsh.
                            Last edited by Sultan_1895-1948; 07-19-2014, 01:51 PM.

                            Comment


                            • Based on stats from a few players I've looked at, I wouldn't penalize OPS+ numbers very much. If the theory is a historically great hitter's OPS+ would drop a significant amount over several more years & PA's, the few stats I've seen don't back that up.

                              Hornsby- First 12 years: 178 OPS+
                              23 year career: 175

                              Heilmann- First 12 years: 147 OPS+
                              17 year career: 148

                              Ott- First 12 years: 155 OPS+
                              22 year career: 155

                              Mantle- First 12 years: 176 OPS+
                              18 year career: 172

                              Mays- First 12 years: 160 OPS+
                              22 year career: 156

                              Aaron- First 12 years: 158 OPS+
                              23 year career: 155

                              Schmidt- First 12 years: 150 OPS+
                              18 year career: 147



                              Of course this is a very small sample. Others, like Lajoie & Foxx & Musial had pretty big dropoffs after 12 years. It seems that "predicting" OPS+ declines for great hitters with shortish careers over x amount of additional PA's is a crap shoot. Some would probably drop rather quickly, some would hardly drop at all.

                              Rewarding longevity is a good thing IMO. Penalizing for an unknown decline seems kind of haphazard though.
                              Last edited by layson27; 07-19-2014, 04:40 PM.

                              Comment


                              • Originally posted by Sultan_1895-1948 View Post
                                That makes sense.

                                Did about a dozen players in the first post and just seemed extreme. Do you really think a 6500 PA guy would drop 15 OPS points had he reached 10,000? Or for example Mize was 2630 PA shy of 10,000 and his OPS drops 14 points. Is that reasonable?

                                Then there's the war issue. For some guys like Ted and Joe I thought about posting the raw result...but then calculate them with assumed 600 PA per missed year in parenthesis.

                                Then again, that assumes the OPS wouldn't change.

                                Either way, the last column is there to view and people can use their own judgement.

                                Here's a thought. Going over 10k is certainly impressive and credit should be given. Is there a way to keep THAT bonus the same, while at the same time lowering the hit a player takes for being under 10k?

                                Example.

                                Ott's had 11,348 PA which is 1348 more than 10k. It gets him from 151 to 158. I think that's reasonable.

                                Let's take 10k - 1348 to end up with 8652. If he had that number of PA (same difference), that would take his 151 DOWN to 131.

                                So it would ding him 20 points for not getting to 10k, and the bonus after 10k is only 7 points. I'm not suggesting the bonus be lower, just that the ding not be as harsh.
                                For the last point first, the 8652 would bring his 51 down to 44, then you add 100 and it becomes 144. You did the 158 correctly. Its basically the same effect.

                                The thing about this method is that it creates a single number and you should not again have to look at playing time. It is actually career value above average produced by the player, but condensed into a 15 year period.

                                A 170 USING THIS SYSTEM produced 70 wins above average batting for every 60 wins that a 160 guy produced above average, for every 30 that a 130 guy did, not matter whether they achieved the marks in 8000, 10000 or 12,000 PAs.

                                And consider an option which would be to simply weigh OPS+ by plate appearances.

                                Say we have three guys with 150 OPS+. One has 8000 PAS, one 10K and one 12K.

                                If we just multiplied rate of production (OPS+) by PAs and divided by 10K we would get scores of:

                                120 for the 8K player
                                150 for the 10K player
                                180 for the 12 K player

                                That is their raw career batting win value.

                                But since I am only multiplying the margin above 100 by PAs, a player gets no added credit for just being average (100 OPS+) for extra years.

                                The 8K guy comes to 140
                                The 10K guy comes ot 150
                                The 12 K guy comes to 160

                                So it is really weighing peak level very highly as you only get value for producing above 100 OPS+.

                                But Mize and Williams need to probably have PAs adjusted for lost time.

                                I like it a lot. If anything, I might lower the baseline slightly to 85 or 90 to give a smidgen of value for being near average, but I am happy enough setting a high standard and just multiplying the value above 100 by PAs/10K

                                Comment

                                Ad Widget

                                Collapse
                                Working...
                                X