Announcement

Collapse
No announcement yet.

Progressing Toward Better Stats Thread

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

  • Oh man I was way off on that. Thanks.

    Here's the correct one.

    WalterOPS.jpg

    So could we take that road number and apply their Adjusted Relative ERA+?


    ADJ-rel-ERA+.jpg
    Last edited by Sultan_1895-1948; 06-09-2014, 06:38 PM.

    Comment


    • Originally posted by Sultan_1895-1948 View Post
      Oh man I was way off on that. Thanks.

      Here's the correct one.

      [ATTACH]139525[/ATTACH]

      So could we take that road number and apply their Adjusted Relative ERA+?


      [ATTACH]139527[/ATTACH]
      So again, their road ERA it looks like from looking at the data that you should divide it by 1.03 not 1.04. I looked through many seasons to estimate home road splits for ERA and it is virtually right on 1.03. For OPS+ you take the road OPS+ divided by 1.04. I can look at more seasons but from what I've seen, those numbers are right to a couple thousandths.

      Comment


      • Ok, I'll change 'em.

        By the way, did you see who fell right between Pete and Walter? Crazy huh.

        Comment


        • Originally posted by Sultan_1895-1948 View Post
          Ok, I'll change 'em.

          By the way, did you see who fell right between Pete and Walter? Crazy huh.

          Interesting. What made you think to run him. Still just 1880 innings.

          Comment


          • Originally posted by brett View Post
            Interesting. What made you think to run him. Still just 1880 innings.
            Was just thinkin about cookie cutter stadiums of the 80's and 90's. Those Reds beat my beloved steroid riddled A's.

            That carpet was brutal for fielders and pitchers. Ball scooted like on concrete. Actually, I remember Will Clark or someone talkin about how short they cut Jack Murphy back in the day, and it behaved like an artificial field. To aid Gwynn I think he said.

            You make a good point about IP though. Either there needs to be minimum to make the list, or ERA needs to be weighted by IP.

            Is there an average expected level of ERA decline after a certain number of IP?
            Last edited by Sultan_1895-1948; 06-09-2014, 09:58 PM.

            Comment


            • Sultan,

              That is crazy about Jose Rijo! I remember him as a very fine pitcher with the Reds, especially in that unfortunate sweep in 1990 (I loved those over-sized A's), but I had no idea he pitched so well for that long. From 1988-1994 he had one helluva run, though most of that does coincide with the huge drop in offense from 1988-1993 or thereabouts. Good catch on Rijo. I do remember when he made that comeback in 2001 at age 36, I had no idea he was so young when he had that awesome 1986 season in Oakland (he was 21 that year). Rijo's comeback in '01 was my first summer back in the States after being stationed overseas for the previous 4 baseball seasons.
              "It ain't braggin' if you can do it." Dizzy Dean

              Comment


              • Originally posted by Herr28 View Post
                Sultan,

                That is crazy about Jose Rijo! I remember him as a very fine pitcher with the Reds, especially in that unfortunate sweep in 1990 (I loved those over-sized A's), but I had no idea he pitched so well for that long. From 1988-1994 he had one helluva run, though most of that does coincide with the huge drop in offense from 1988-1993 or thereabouts. Good catch on Rijo. I do remember when he made that comeback in 2001 at age 36, I had no idea he was so young when he had that awesome 1986 season in Oakland (he was 21 that year). Rijo's comeback in '01 was my first summer back in the States after being stationed overseas for the previous 4 baseball seasons.
                Rijo actually got drafted on our 5 year peak draft a few years ago and higher than Saberhagen, Hudson, Hershizer. It was 5 consecutive year peak.

                Comment


                • Probably even higher if they knew everyones road production. You wanna run adjusted on just those five years?

                  By the way...the gap between Maddux and Pedro...can longevity close a distance that large?
                  Last edited by Sultan_1895-1948; 06-10-2014, 10:08 AM.

                  Comment


                  • Originally posted by brett View Post
                    Rijo actually got drafted on our 5 year peak draft a few years ago and higher than Saberhagen, Hudson, Hershizer. It was 5 consecutive year peak.
                    I know he had some fine seasons, but looking at it now I had no idea they were that good. Again, that may be because there were a lot of pitchers that had a very fine season or two during that post-1987 to pre-expansion mini-era. Rijo may have been brushed aside back then because his win totals were so low. I would think more noise would have been made about him due to his strikeouts and ERAs just about every year then. Perhaps if he pitched somewhere other than Cincinnati I guess, which is strange given that he was the ace of a World Series sweeping underdog staff.

                    Poor Orel and Bret never seemed to stay healthy long enough to run up a stretch like that. Good stuff! I love discovering, and in this case re-discovering, how good some guys were!
                    "It ain't braggin' if you can do it." Dizzy Dean

                    Comment


                    • Originally posted by Sultan_1895-1948 View Post
                      Probably even higher if they knew everyones road production. You wanna run adjusted on just those five years?

                      By the way...the gap between Maddux and Pedro...can longevity close a distance that large?
                      A simple way to get a pitchers value (above average) is, well let me show you.

                      If a guy has a 140 ERA+, you make it 1.4. Then you take 1/1.4 to get 0.714. Then you subtract that from 1, or 1-.714 to get .286. That basically says that a pitcher saved .286 games worth of runs per 9 innings.
                      Then you take the .286 x innings divided by 9.

                      So let's say someone has a 130 ERA+ and 5000 innings. Someone else has a 160 ERA+ and 3000 innings.

                      So 130 becomes 1.30.
                      Take the reciprocol (1/1.30) to get .769
                      Take 1 minus .769 to get .231
                      Multiply by innings/9 or .231 x 5000 / 9 = 128 games worth of runs above average.

                      Player 2
                      Take 160 and make it 1.6
                      Reciprocol is .625
                      1-.625= .375
                      .375 x 3000 / 9 = 125

                      So that would be close.

                      by the way, this number should be about double the number of WAA that a player scores. (you have to save about 2 games worth of runs to net one win). And notice that Pedro is about 61 WAA (doubled is 122) and Maddux is about 65 WAA (doubled is 130). WAA already has them virtually the same.

                      And there is another issue and that is that when you save more runs in the same game it has an interaction effect on winning.

                      So we could also do a pythagorian estimate.

                      My first player pitched 555.6 games worth of innings. It his team scored average, his ERA+ would represent the proportion that they outscored their opponents by. So they would have outscored their opponents 1.3 to 1. A quick pythagorian estimate says that they would win 1.3 x 1.3/ (1.3 x 1.3 + 1) ratio of their games. That would be a 62.8 winning percentage on 555.6 games, or 349 wins and about 207 losses. That is 142 wins above losses, or 71 wins above what they would have done if they had won half of those 556 games (349 versus 278). This is close to the 65 WAA but a little higher because the runs saved are somewhat concentrated.

                      Now the guy who was at 160 for 3000. He pitched about 333 games worth of innings. His team would have a 1.6 to 1 scoring edge. That gives about 1.6 x 1.6 / (1.6 x 1.6 + 1) or .719 winning rate (71.9 winning percentage). That would give his team 239 wins and 94 losses. That is 145 wins above losses, or about 73 wins above 50%.

                      War has Pedro again at about 61 WAA and Maddux at about 65, but WAR looks like it just converts runs to wins. IF it used a pythagorian approach which factors in that pitchers save runs in a concentrated fashion, it looks like Pedro might already flip flop with Maddux and really have been worth 3 or 4 MORE WAA rather than 3-4 fewer.

                      The problem though is that while Pedro can be worth more WAA, he wont be worth more in total, because teams have at least some below average pitchers. When you add a #1 starter he doesn't take the place of an average starter, he takes the place of a 5th starter with maybe a 70 ERA+, or a 30% winning percentage.

                      Comment


                      • I like that first method. If we added that final column to the pitching chart, would we then have adjusted AND weighted road expected ERA+?

                        I was surprised how high Seaver ended up. Him vs Pedro and Koufax vs Spahn or Ford would be interesting. Just to see how much their IP hurts or helps.

                        And for relievers too. Riveras already sky high raw ERA+ will prob go up a handful of points since his road ERA is better.
                        Last edited by Sultan_1895-1948; 06-10-2014, 12:02 PM.

                        Comment


                        • Originally posted by Sultan_1895-1948 View Post
                          I like that first method. If we added that final column to the pitching chart, would we then have adjusted AND weighted road expected ERA+?

                          I was surprised how high Seaver ended up. Him vs Pedro and Koufax vs Spahn or Ford would be interesting. Just to see how much their IP hurts or helps.

                          And for relievers too. Riveras already sky high raw ERA+ will prob go up a handful of points since his road ERA is better.
                          Yes the last one would be weighted by innings, and if you use the road predicted, it is adjusted or neutralized.

                          Basically it is "games worth of runs SAVED above average." When I came up with my first system for rating players, I used games worth of offense or pitching above average. Again, a game worth might be 4.8 runs per game, but a win might be twice that many. I effectively counted "half" games in the standings, or games above .500.

                          Remember regarding Rivera that a 200 ERA+ is not twice as good as a 150 in terms of saving runs. A 200 ERA+ means that the player allowed 50% as many runs as average. A 150 means that he allowed 1/1.5 or .67 as many runs as average. I kind of wish that instead of ERA+, they used relative ERA, where a 200 ERA+ would be 0.50 and a 150 would be 0.67.

                          Or then you could take that and put it on a 4.5 ERA scale, or a 4.0 era scale. So on a 4.0 ERA scale, a 160 ERA+ would be about 2.40 relative ERA, or on a 4.50 scale it would be about 2.70.

                          Comment


                          • Ok here's Koufax

                            127.9 Adjusted Rel ERA becomes 1.279

                            1 / 1.279 = .7818

                            1 - .7818 = .2182

                            .2182 x 2324.1 IP (should .1 be turned into a 3, and .2 be turned into a 6?) = 507.11 / 9 = 56.34

                            -----------------------------------------------

                            Now let's do Ryan. Much lower AdjRelERA but way more innings

                            102.7 Adjusted Rel ERA becomes 1.027

                            1 / 1.027 = .9737

                            1 - .9737 = .0263

                            .0263 x 5386.0 = 141.65 / 9 = 15.73

                            ------------------------------------------------

                            Wow, so the clear message here, is that simply having a ton more innings, isn't going to bring you much closer; if there's a large enough starting gap in AdjRelERA

                            -------------------------------------------------

                            Well...just did Rivera. His insane 205 ERA+ goes up to 238.2 adjusted (better on road and high league ERA), but because of his low 1283 IP, his number is 82.72. Haven't done any others, so this is probably very good, but are the low IP being fully factored in? Of course during the vast majority of his innings, the batters are facing him just one time. Perhaps there's a further adjustment needed for closers. Any way to get the league ERA for closers, or maybe for relievers in general, with a max IP ceiling.
                            Last edited by Sultan_1895-1948; 06-10-2014, 07:32 PM.

                            Comment


                            • Well that is in wins above a .500 level again, about a 100 ERA+ baseline, so if Ryan is adjusted to just 102.7 that is almost nothing. In effect he is getting 2.7% above average while Koufax is getting 28% above average. Ryan's is pretty stunningly low.

                              In reality, even if we like WAA for batters, a pitcher is getting us wins above a #5 starter. That's not really a replacment, its a real player. And a #5 starter is probably around an 80 ERA+. 1/.8 is 1.25 so if you want to use a lower baseline, you would put 1.25 instead of 1 that you subtract the .9737 and .7818 from. Also, starters tend to have a little worse ERA+ than average, somewhere around 92, but it goes down the more relievers get used. This is kind of important. In the method we are using, a season below a 100 ERA+ will hurt a players final score. Greg Maddux for example gained no WAA in his last 3 seasons with a 100 ERA+ for around 600 innings, but the next best pitcher clearly would have cost about a win per season on average for those 3 years. WAR uses a baseline of about a .77 ERA+. Using a .90 ERA+ would give you about a 40% winning percentage pitcher for a baseline.



                              Anyway, innings do help a lot. If a pitcher had a 133 ERA+ (0.75 relative ERA) for 4000 innings it would about match someone with a 200 ERA+ (.50 relative ERA) for 2000 innings. The amount below 1.00 is proportional to the innings needed. The first guy has a margin of .25 and the second one .50. .25 x 4000 = .50 x 2000.

                              Round .1 to .3, and .2 to .7

                              If Ryan had had a 112 predicted ERA+ to match his career total, he would have scored about 64 instead of 15.73.

                              Ryan's total RA+ (including unearned runs) was also even better at 114, and defense adjusted was 116. And if you ever run his total OPS+ allowed (not sure about road) well, he allowed .204/.307/.298 overall (though .219/.326/.323 on the road). It looks like he allowed only about a 70 OPS+ for his career, but about 80 on the road. Even the road 80 OPS+ would correspond to a 115 ERA+ on the road, so he obviously got hurt more by wild pitches, baserunning, and not pitching as well with men on base (sequencing/big innings).

                              BUT the most interesting thing about the Ryan/Koufax comparison is that Ryan was relatively much worse on the road than at home, that is relative to other pitchers. It looks like it was largely in the slugging component so I wonder if his park supressed home runs.
                              Last edited by brett; 06-11-2014, 06:22 AM.

                              Comment


                              • Checking, he gave up 10.3 home runs per 650 PAs on the road, and 9.7 on the road.

                                Neutralized, that is about .2 home runs per 650 BF, or about .3 per season. That is only about .45 runs a season which is insignificant.

                                His road problems are almost all explained by batting average (he got hit almost 10% worse on the road than at home relative to normal splits).

                                So basically he gave up about a .190 home average, and a .220 road average, and if we adjust them to neutralize them both it would be like allowing .215 on the road and .195 at home. Seems trivial but its 10% more hits.

                                And what's more, it comes down largely to one factor, He allowed a .286 BABIP on the road, and only .260 at home. Normally a .260 at home would correspond to about .270 on the road. That explains about 13-14 points of his higher road average.

                                He also averaged about 1 less K on the road per 9 innings. Well that explains the other 6-7 points.

                                He gave up more hits on the road by about 10% than effected, about 7% due to BABIP and 3% due to lower K rates.

                                There is a problem though. I am using raw home numbers, but he had lower offensive parks. In fact his parks explain virtually all of the home/road OPS+ difference and half of the difference in batting average.

                                He actually got charged with a higher percentage of earned runs on the road.

                                So we are left with, fully neutralized, that he was about 16% worse on the road than at home. About 3% is due to BABIP, about 3% to having 1 less K per 9 innings, about 2% is due to more runs being scored as earned on the road. And with him, it must just mean that because he was hurt more by having men on base than others, he suffered more from the raw higher road rates than would be expected because pretty much everything else is relatiely comparable home and road.

                                Comment

                                Ad Widget

                                Collapse
                                Working...
                                X