RA versus ERA 1871-2010: Baseball Databank Data Dump 2.3.1

by Matt Klaassen on Mar 1, 2011 4:30 PM EST in Sabermetric Primers

I hate Earned Run Average. I doesn't add anything truly useful over Runs Allowed, unless you think that errors are a Really Good Stat. I mean, I guess they're better than Pitcher Wins or something, but... I don't even have the energy to get into it right now. I'm not even talking about ERA vs. FIP, or DIPS or tRA or SIERA or anything like that. I just mean the "E" part of it. There are many, many good posts on why the "earned" rules is silly, for example this one by Jack Moore or this classic by Michael Wolverton, and I'm not going to get into it, here. Moreover, if you're, say, calculating pitcher WAR, for example, the ERA scale (with whatever statistics) will undervalue pitchers when compared to position players since you aren't comparing it to a "real" runs scale, the runs-to-win conversion is thrown off, and so forth. Sigh. However, I grant that it is still the most common pitching statistic, and most of the newer pitching stats are scaled to it (not because the creators "believe" in ERA, but because they understandably want to use a scale people are familiar with). But until That Great Day comes when ERA is banished and we're all on an RA scale, having a quick "scale" to convert from one to the other might be helpful.

That's right, it's time for another thrilling Baseball Databank Data Dump.

I decided to do this rather simple "dump" because I am thinking of doing a few slightly more involved pitching-related examples down the road. In the past, I've posted both RA and ERA scaled-stuff, but if a chart has more than a few columns, it really gets messed up. On one hand, I'm tempted just to go cold turkey and just post everything scaled to RA. On the other hand, despite my irritation with ERA and my thought that the best way to "cure" people of ERA is for all the stats sites to simply switch to using all RA-scale, all the time, I understand the difficulty of working against such a high coefficient of static friction involved. I haven't decided what I'm going to do in my future posts (and I do want to make them user-friendly for all 18 non-BtBS readers who persuse them).

If I do decide to go "all RA" or if you want the conversion or comparison for other reasons, here is the a table with the yearly list of major league-wide RA versus ERA from 1871 through 2010. As you can see., in recent seasons the "scale" of ERA divided RA has been about 0.92. So if you want to use an ERA-scaled stat in an RA fashion, a quick workaround is to divide it by 0.92. If you want to be more precise, you can use the specific conversion. The "Diff" column is, as you might expect, lgRA minus lgERA for each season. As you might surmise from the graph above as well as the table below, there's a pretty big "improvement" in the early part of the century. My assumption (I'm far from a historian of baseball) is that it's partly an improvement in fielding skills and positioning, but probably mostly improvements in glove design and playing-surface upkeep. It would be fun to have a discussion on those topics.

Hopefully this is useful, or at least interesting, to someone out there. As always, the source data is from the freely available Baseball Databank, another great resource proviced by Sean Forman.

RA and ERA, 1871-2010
*Season*	RA	*ERA*	*Scale*	*Diff*
2010	4.4284	4.0794	0.9212033	0.3489
2009	4.6629	4.3218	0.92684776	0.3411
2008	4.6881	4.3238	0.92229356	0.3643
2007	4.8335	4.4731	0.92543521	0.3604
2006	4.9099	4.5268	0.92198822	0.383
2005	4.6476	4.291	0.92326988	0.3566
2004	4.8482	4.4647	0.92090178	0.3835
2003	4.7722	4.4	0.92201236	0.3722
2002	4.6609	4.2788	0.91802035	0.3821
2001	4.8234	4.4175	0.91585844	0.4058
2000	5.197	4.7697	0.91778463	0.4273
1999	5.1426	4.7083	0.91555628	0.4343
1998	4.8213	4.4316	0.91915936	0.3898
1997	4.8064	4.3895	0.9132568	0.4169
1996	5.0661	4.6109	0.910166	0.4551
1995	4.8842	4.4515	0.91142477	0.4326
1994	4.9593	4.5116	0.90972575	0.4477
1993	4.6356	4.1906	0.90399732	0.445
1992	4.1256	3.7456	0.90790612	0.3799
1991	4.3194	3.9112	0.90550008	0.4082
1990	4.2933	3.8613	0.89938055	0.432
1989	4.1534	3.711	0.89347889	0.4424
1988	4.1526	3.7335	0.8990794	0.4191
1987	4.7624	4.2935	0.90152391	0.469
1986	4.4302	3.9727	0.89673767	0.4575
1985	4.3534	3.8936	0.89437857	0.4598
1984	4.2777	3.811	0.89091011	0.4667
1983	4.3328	3.8728	0.89383599	0.46
1982	4.303	3.8629	0.89773606	0.44
1981	3.9977	3.5835	0.89638468	0.4142
1980	4.2914	3.8419	0.89525287	0.4495
1979	4.4937	4.0029	0.89077112	0.4908
1978	4.1406	3.687	0.89044113	0.4536
1977	4.4841	3.9986	0.89171941	0.4855
1976	3.9921	3.5141	0.88026078	0.478
1975	4.2234	3.7105	0.8785517	0.5129
1974	4.1329	3.6273	0.87766422	0.5056
1973	4.2256	3.7459	0.88648021	0.4797
1972	3.6877	3.2639	0.88508682	0.4238
1971	3.9019	3.4688	0.88900683	0.4331
1970	4.3582	3.8859	0.89164692	0.4722
1969	4.0885	3.6131	0.8837224	0.4754
1968	3.42	2.9822	0.87199568	0.4378
1967	3.7654	3.3031	0.87723178	0.4623
1966	3.9974	3.5208	0.88077519	0.4766
1965	3.9926	3.4989	0.87633246	0.4938
1964	4.0536	3.5804	0.8832673	0.4732
1963	3.9519	3.459	0.87527387	0.4929
1962	4.4871	3.9553	0.88147431	0.5318
1961	4.5761	4.0274	0.88008036	0.5488
1960	4.3226	3.8171	0.88306452	0.5055
1959	4.4197	3.9041	0.88335023	0.5156
1958	4.3176	3.8605	0.89411987	0.4571
1957	4.284	3.8329	0.89469725	0.4511
1956	4.4933	3.9642	0.88224096	0.5291
1955	4.526	3.9993	0.88363899	0.5266
1954	4.4039	3.8971	0.88491734	0.5068
1953	4.6675	4.1377	0.88648696	0.5298
1952	4.1944	3.7028	0.88279061	0.4916
1951	4.5664	4.0392	0.88454029	0.5272
1950	4.9205	4.3586	0.88579039	0.562
1949	4.6644	4.1184	0.88294519	0.546
1948	4.642	4.1188	0.88728784	0.5232
1947	4.4217	3.8853	0.87867036	0.5365
1946	4.0455	3.4575	0.85464591	0.588
1945	4.2057	3.5823	0.85177616	0.6234
1944	4.1788	3.5205	0.84246575	0.6583
1943	3.8844	3.3332	0.85810322	0.5512
1942	4.0982	3.4845	0.8502401	0.6137
1941	4.5277	3.89	0.85915115	0.6377
1940	4.7221	4.1123	0.87087295	0.6097
1939	4.8946	4.2666	0.87169334	0.628
1938	4.9724	4.2826	0.86127361	0.6898
1937	4.9582	4.2651	0.8602204	0.6931
1936	5.2509	4.5229	0.86136381	0.728
1935	4.9479	4.2363	0.8561809	0.7116
1934	4.9715	4.2805	0.86100225	0.691
1933	4.5127	3.8067	0.84354795	0.706
1932	4.9238	4.1774	0.84840975	0.7464
1931	4.8674	4.1225	0.84696881	0.7449
1930	5.6354	4.81	0.85353203	0.8254
1929	5.2526	4.4754	0.85204642	0.7771
1928	4.7485	4.0087	0.84421288	0.7398
1927	4.8024	4.0234	0.83780332	0.7789
1926	4.7034	3.9202	0.83347887	0.7832
1925	5.1894	4.3266	0.83374374	0.8628
1924	4.7937	4.0463	0.84408694	0.7474
1923	4.8389	3.9882	0.82418323	0.8508
1922	4.9042	4.0647	0.82883481	0.8394
1921	4.8766	4.0311	0.82662305	0.8455
1920	4.3535	3.4572	0.7941313	0.8962
1919	3.8842	3.063	0.78859447	0.8211
1918	3.6111	2.7656	0.76585366	0.8455
1917	3.577	2.682	0.74977639	0.8951
1916	3.5522	2.7186	0.76533483	0.8336
1915	3.839	2.9013	0.75575097	0.9377
1914	3.9001	2.9069	0.74533434	0.9932
1913	4.0691	3.0642	0.75303197	1.0049
1912	4.5802	3.3664	0.73497979	1.2138
1911	4.5523	3.365	0.73919283	1.1873
1910	3.8585	2.7681	0.7173981	1.0904
1909	3.5682	2.5298	0.70898149	1.0384
1908	3.3991	2.3661	0.69609588	1.033
1907	3.5945	2.5025	0.69619091	1.092
1906	3.6832	2.6574	0.72149049	1.0258
1905	3.9483	2.8199	0.71419683	1.1284
1904	3.8072	2.6613	0.69901139	1.1459
1903	4.5396	3.1119	0.68550666	1.4277
1902	4.5168	3.1695	0.70171001	1.3473
1901	5.0976	3.4876	0.68416298	1.61
1900	5.3851	3.6966	0.68644639	1.6885
1899	5.4791	3.8485	0.70239453	1.6306
1898	5.1497	3.6029	0.69963851	1.5468
1897	6.1339	4.3079	0.70231092	1.826
1896	6.2487	4.3581	0.6974359	1.8906
1895	6.8593	4.7764	0.69633658	2.0829
1894	7.6555	5.313	0.69400685	2.3425
1893	6.696	4.6609	0.69607368	2.0351
1892	5.2458	3.2845	0.62611845	1.9613
1891	5.8506	3.5315	0.60361569	2.3191
1890	6.1883	3.8818	0.6272826	2.3065
1889	6.1458	3.9319	0.63977211	2.2139
1888	4.9631	2.9455	0.59348522	2.0176
1887	6.5734	4.1727	0.63479168	2.4007
1886	5.6707	3.3702	0.59432242	2.3005
1885	5.3197	3.032	0.56995265	2.2877
1884	5.568	3.1073	0.55806839	2.4607
1883	5.857	3.2171	0.54928018	2.6399
1882	5.4041	2.8036	0.51878589	2.6005
1881	5.1499	2.7749	0.53882078	2.375
1880	4.7616	2.3741	0.49858978	2.3875
1879	5.2926	2.4965	0.47169258	2.7961
1878	5.1552	2.3041	0.44695378	2.8511
1877	5.6649	2.8075	0.49558824	2.8575
1876	5.8147	2.3073	0.39679948	3.5075
1875	6.1537	2.4592	0.39962211	3.6946
1874	7.4844	3.0175	0.40317003	4.4669
1873	8.9883	3.2513	0.36173184	5.7369
1872	9.284	3.7174	0.4004131	5.5666
1871	10.636	4.22	0.3967657	6.416

Tweet Comment 7 comments | 1 recs |

Do you like this story?

Comments

This is really cool stuff, Matt

Any theories on why the scale has changed so much over the years? Changes in rules? Changes in scoring mores?

Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.

by J-Doug on Mar 1, 2011 4:45 PM EST reply actions

Yeah, as I said in the post

my guess for the early years is the quality/nature of gloves, and groundskeeping. I’m sure scoring mores have something to do with it. We’d also want to parse how much it has to do with runs scoring after errors and stuff.

Making watching baseball as fun as doing your taxes.
My Twitter feed.

by Matt Klaassen on Mar 1, 2011 6:36 PM EST up reply actions

This is interesting stuff

my first instinct is that the changes in defensive equipment has a lot to do with this. I imagine dropped popups were fairly common in the late 1800’s, when baseball gloves were still often fingerless.

Is there any way to try and parse the factors? I would be curious how much of this is physical (equipment, higher likelihood of an error off of a spitball) and how much is scoring changes and the like. It would give us a way to assess the real changes in the game itself over the years.

by djg2111 on Mar 1, 2011 6:25 PM EST reply actions

Why not scale RA to ERA?

Scaling RA to ERA would eliminate two problems:
1) We wouldn’t have to use ERA over RA
2) We wouldn’t have the issue of recalibration

This would probably be confusing at first, but it would eliminate the need to use ERA.

by yankee82093 on Mar 1, 2011 7:22 PM EST reply actions

I'm not sure what you mean

One of the primary motivations behind this post is that it’s the ERA scale itself that is distorting. We already have RA. It’s even easier to track than ERA (no worries about arbitrary distinctions). We simply have to start using it.

Making watching baseball as fun as doing your taxes.
My Twitter feed.

by Matt Klaassen on Mar 2, 2011 10:21 AM EST up reply actions

I agree, it's not a tough scale.

We already use it to measure team runs scored and runs allowed per game. If you like the RC/27 metric, you’re already using it. It’ll be a bit annoying at first, but won’t take long to get used to. Not like OPS, where you’re completely blind.

@Sky_Kalkman

by Sky Kalkman on Mar 2, 2011 11:36 AM EST up reply actions