Baseball statistic question

Boopotts

Would a stat like IP(1/ERA) do a better job of capturing a pitcher's actual value to a team than any of the other statistics (or combinations of stats) that are currently recorded for pitchers? Assume the pitcher's ERA is >0.

perkdog

Skinpinch? Paging MR.Skinpinch.....

Skinpinch

Lets see.

Greg Maddux in 2000(NO BALLPARK ADJUSTMENT) using that formula would give him....

83 pitcher runs saved.

But it does not take into account the legue ERA at all, so all the pitchers who pitched in the DH AL, or the live ball era will get severely penalized.

Here is the pitcher run formula usually used.......

League ERA - Player ERA X IP by pitcher/9.

In 2000 for Maddux it would be 4.63-3.00 X 249/9

1.63 X 27.66 = 45 pitcher runs.

So we have one figure crediting him for 83 runs and one for 45. THe question is which is doing a better job of capturing his contribution. Lets look at it on a TEAM level to get a better idea. Lets use the 2000 Milwaukee Brewers.

Using the formula you presented Boo, it gives the Milwaukee Brewers staff a number of 315.
Using the second formula usually used, it gives the Milwaukee Brewers staff a number of -1.62.

Now those are vastly different numbers! If this was done for all teams(IN THE SAME LEAGUE), each measurement would rank them pretty much in the same order. If the measurements were done comparing pitchers from different league, then it wouldn't take into account the ease/difficulty of the pitching environment, and would creat invalid lists.


Back to the 2000 Brewers. They had a team ERA of 4.64, and the legue had an ERA of 4.63.

Looking at their ERA and the leauge ERA, which number is more indicitive of their contribution 315, or -1.62?


How about the 2000 Cubs. Boo, your formual gives them 276 pitcher runs. The stanard one gives them -98. Their ERA was 5.58.


If you did this for every team in the SAME LEAGUE, both lists would give the same rank and order(I believe), BUT the standard list puts it into a more clear context AND enables you to compare an average pitchers worth to that of an average hitters worth, because they are being measured against how many runs they save/create over an average player. A run saved on defense is just as valuable as a run created on offense.


But I repeat, the biggest problem is that the proposed measurement takes no consideration in the LEAGUE ERA, and thus renders it very unfair and inaccurate when comparing different leagues, and different era's.


Some people might just say, then look at the ERA's for the idividual, but IP are a big contributing factor in a pitchers value. An ERA of 3.00 over 150 innings is not as valuable to a team as a guy with a 3.00 ERA over 250 innings. Both formulas take innings and earned runs into account, one just doesn't take the league environment into account.

jaxxr

Before expansion in 1961, most formula worked better. A dilution factor is really an unknown or estimate. The more teams added, the lesser the constant or equality of opposition, due to divisions teams playing more games against one division rather than other divisions.

Before intra-league play, at least the pitchers ( hitters too ) faced the same league, however only in somewhat different proportions. Now an AL hurler may face ONLY several NL teams and not even the same ones as a pitcher from a different AL team faces. Very tough to rate overall performance by a pitcher (or hitter) with so many variables/differences in the opposition played against.

A good rate of a pitcher's worth would have to take into consideration, ERA rank, Adj ERA+ rank, Wins rank, IP rank, Win% rank, WHIP rank, a few others, along with Quality of opposition, very subjective/difficult to evaluate..

It has been said a few times, Whitey Ford was not that great a pitcher, he never had to face the likes of Mantle, Maris, Berra, Etc. His Adj ERA + was quite good nevertheless, and he pitched as well as most pitchers of his time did, against Boston, Detroit, and so on. It would be at least nearly possible to compare his performance with , say Bob Lemon, remove Lemon's stats versus the Yanks, and remove Ford's stats versus the Indians, then evalute the numbers. I mention this to show how difficult it might be today. A Yankee hurler faces diferent teams in different proportion than an Indian pitcher within the AL itself, then in addition, they also face different NL teams. If only similar opponents are used the sample will probably be too small for a decent comparision.