tag:blogger.com,1999:blog-12133335.post6394889591655338976..comments2024-02-13T06:56:14.486-05:00Comments on Walk Like a Sabermetrician: Early NL Series: Batter Evaluationphttp://www.blogger.com/profile/18057215403741682609noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-12133335.post-58873246372684188152008-06-06T13:43:00.000-04:002008-06-06T13:43:00.000-04:00Thanks for that, it is helpful. I did quickly che...Thanks for that, it is helpful. <BR/><BR/>I did quickly check out how a moving 10-year average winning percentage of last place teams varied over time, and it does decline from 0.380 or so in the 80's and 90's to 0.330 is the 20's, and down to 0.300 or so in the early 1900's... So, there would seem to be some basis for dropping the cutoff. <BR/><BR/>One approach that I am considering might be to set the modern cutoff to 0.350, and then use the decline in average last place team winning percentage between today and whatever historical date I'm looking at to adjust the replacement level decline. That approach, however, would drop it to 0.300 by 1945 or so, and that might be too much of an adjustment based on your work. I guess it could reflect that modern day replacement level is set too low, of course...<BR/><BR/>Like you, I'm not entirely convinced that the modern day replacement level is built upon a particularly solid foundation. I view it as more of a convenient, reasonably low baseline that seems to do a good job of balancing the value of playing time vs. performance. So I can see that there might not be anything wrong with using modern day replacement level going back quite a ways into the past as well. <BR/><BR/>I dunno, I'll keep thinking on it. :)<BR/>-Justinjinazhttps://www.blogger.com/profile/07697776280178146413noreply@blogger.comtag:blogger.com,1999:blog-12133335.post-8628131882892586012008-06-03T23:14:00.000-04:002008-06-03T23:14:00.000-04:00As I've tried to express, I'm not particularly enc...As I've tried to express, I'm not particularly enchanted with any of the means of estimating replacement level. Anyway, I have been sketching out some stuff for post-1883, and I am going to switch to using .350 like I do today for those seasons. <BR/><BR/>The NL, after expelling PHI and NY in 1876, went through a period of several years in which they had to desperately look for teams from places like Indianapolis, Milwaukee (both much less impressive then than now), Worcester, Syracuse, Troy,...These teams often were run on shoestring budgets and had some pretty lousy players playing for them. It wasn't until 1879 that historian David Nemec says that playing in the NL really became seen as a status symbol, any more prestigious than playing for a strong independent club, of which there were many. The Buffalo Bisons in 1878 went 10-7 (IIRC) against their NL exhibition opponents, then joined the NL in 1879 and finished a solid third.<BR/><BR/>My point in all of that is that often the teams at the bottom of the standings were pretty weak, and one could argue that they both represented replacement level talent--relative to the world of high-level profession baseball as a whole. This is something we don't have to deal with today, since obviously the best talent all the way up the ladder is controlled by the major league clubs.<BR/><BR/>I think that treating the NL's weaklings, who moved through the league like it was a revolving door , as the baseline of talent that Chicago or Boston could reasonably acquire in an emergency is probably a mistake.<BR/><BR/>But of course I could be way off base about all of that.<BR/><BR/>But for your study, starting with the Reds in '82, I would suggest sticking with defining repl level however you do it now. And of course I would include a measure versus an average baseline as well so you have something on firmer ground. (Incidentally, I have already run my numbers and scrounged up some anecdotes for the 1882-83 AA if you ever want to compare notes--you'll probably get around to writing your series before I get around to posting my write-ups for those seasons here. Also, please be sure to explain to the Reds fans that the modern day Reds started with the 1882 team and not with the 1869 team. :-)phttps://www.blogger.com/profile/18057215403741682609noreply@blogger.comtag:blogger.com,1999:blog-12133335.post-89860486392241900432008-06-03T15:41:00.000-04:002008-06-03T15:41:00.000-04:00Hi,I somehow completely overlooked this series. ...Hi,<BR/><BR/>I somehow completely overlooked this series. But it's very interesting to me because (once i get some career stuff taken care of) I'm keen to do some statistical reviews of past winning Reds teams, including the 1882 team. The issue of a replacement baseline is one that I've been thinking about, as one would certainly think that replacement level has increased over time simply because the pool of talent from which teams can draw is so much bigger now than it used to be.<BR/><BR/>I'm obviously pretty new to these discussions, but I'm curious as to why the average winning percentage of the bottom-two teams would be a good indicator of replacement level. If anything, I'd think the average worst team would be your best indicator...and even then, I'd expect that few teams are actually as bad as an all replacement-level team unless incompetence is a more widespread problem than I think it is. I do see that those winning percentages match up reasonably well to the bench vs. starters comparison studies, but as you mention here, selection bias (among other problems) is a major issue with those studies...<BR/><BR/>Anyway, I'd just be interested in hearing your views on this.<BR/><BR/>Thanks,<BR/>Justinjinazhttps://www.blogger.com/profile/07697776280178146413noreply@blogger.com