Sacrifice Fly should be removed to normalize historical data.

Sacrifice Hit:

For most of baseball history the sacrifice has followed essentially the current rule: a sacrifice was awarded only on a bunt that was not an obvious attempt to bunt for a hit...

     1908. A sacrifice is now awarded if the batter succeeds in driving in a run with a fly out.

• 1909-1919. A sacrifice is now awarded if the batter drives in a run with a fly ball that results in an error but would have driven in the run if the error had not been committed. Although sacrifice flies were included in the sacrifice hit total, they were also counted separately.

• 1920-1925. The scoring distinction between sacrifice bunts and sacrifice flies is abolished.

• 1926-1930. A sacrifice is awarded on any fly out that advances a runner, or on a fly ball resulting in an error that would have advanced a runner had the out been recorded.

• 1931-1938. No sacrifice is awarded for a sacrifice fly.

• 1939. Sacrifice flies are again counted as sacrifice hits, but only those that result in a score.

• 1940-present. Sacrifice hits are awarded only on bunts that advance a runner and result in an out, or would have resulted in an out but for an error or unsuccessful fielder's choice.

Because the rules on exactly what constituted a sacrifice hit varied considerably during the early game, sacrifice totals from the early game can't necessarily be compared directly with modern ones.
______________________________________

Sacrifice Fly:

A sacrifice fly is a fly ball out that scores a baserunner. While typically a long fly ball to the outfield, a sacrifice fly can be recorded on any fly ball, whether fair or foul. The runner usually scores from third base, but there have been cases of runners scoring from second or even first base on a sacrifice fly; it is also possible for more than one runner to score on the play. The batter is charged with a plate appearance but not with an at bat, and is credited with an RBI for each run scored as a result. No sacrifice fly is credited if a baserunner advances one or more bases as a result of a fly ball that is caught but does not score.

The sacrifice fly is a recent statistic, dating back to 1954 in its present form, although it has always been possible for a runner to advance after a fly ball is caught for an out.
___________________________________

Sacrifice fly rule and the last two .400 hitters: Bill Terry, Ted Williams. Plus Joe DiMaggio. Sunday, June 12, 2011

Beginning with the 1926 season, a more liberal version of the sacrifice fly rule was instituted. The basic provision was that any players who hit fly balls advancing runners to second and third were credited with sacrifices; no time at bat was charged for a sacrifice. After the 1930 season, during which the collective batting average of the major leagues exceeded .290, the sacrifice fly rule was eliminated... The elimination of the sacrifice fly rule reduced the number of sacrifices from 1317 in 1930 to 789 in 1931 in the National League, and from 1283 to 650 for the same two seasons in the American League. In 1939 the scoring sacrifice fly, exempting a batter from a time at bat when a runner scored after the putout on a fly ball, was restored to the game. This lasted for a single season. after the 1953 season ... restored the sacrifice fly ... unchanged to the present day (no date on this SABR article) ... Beginning with the 1954 averages, sacrifice bunts and sacrifice flies were listed separately in the official averages (AL pitching statistics did not separate SF from SH yielded until 1955)... SABR member Pete Palmer has determined that the three sacrifices credited to Ted Williams in 1939 were all SF.
_________________________________________

Sacrifice fly is a dumb scoring rule. Dump it and normalize historical data. Tuesday, June 14, 2011

It's difficult enough to compare players from different eras without having the scoring rules and qualification criteria change. The data should be normalized for these two things. Eliminate the sacrifice fly scoring rule. It never made sense and it's erratic application increases confusion.
________________________________________________

I recently ran a query for seasons 1901 through 2014: on base percentage (OBP) >= .490 and qualified for league "batting title" sorted by OBP.

OBP: ((H+BB+HBP)/(AB+BB+HBP+SF)

OBP bugs me as it is not a percentage but an average but for purposes of this post I will use the standard abbreviation.

I was really looking for batters with OBP >= .500 but expanded it to see if there were some who were close.  That expanded the list from 14 to 23 seasons but added only one more player: Arky Vaughn.

Rk	Player	OBP	Year	Age	Tm	Lg	G	PA	AB	R	H	2B	3B	HR	RBI	BB	IBB	SO	HBP	SH	SF	GDP	SB	CS	BA	SLG	OPS	Pos
1	Barry Bonds	.609	2004	39	SFG	NL	147	617	373	129	135	27	3	45	101	232	120	41	9	0	3	5	6	1	.362	.812	1.422	*7/HD
2	Barry Bonds	.582	2002	37	SFG	NL	143	612	403	117	149	31	2	46	110	198	68	47	9	0	2	4	9	2	.370	.799	1.381	*7/DH
3	Ted Williams	.553	1941	22	BOS	AL	143	606	456	135	185	33	3	37	120	147		27	3	0		10	2	4	.406	.735	1.287	*7H/9
4	Babe Ruth	.545	1923	28	NYY	AL	152	697	522	151	205	45	13	41	130	170		93	4	3			17	21	.393	.764	1.309	97/83
5	Babe Ruth	.532	1920	25	NYY	AL	142	616	458	158	172	36	9	54	135	150		80	3	5			14	14	.376	.847	1.379	*978/3H1
6	Barry Bonds	.529	2003	38	SFG	NL	130	550	390	111	133	22	1	45	90	148	61	58	10	0	2	7	7	0	.341	.749	1.278	*7/DH
7	Ted Williams	.526	1957	38	BOS	AL	132	547	420	96	163	28	1	38	87	119	33	43	5	0	2	11	0	1	.388	.731	1.257	*7/H
8	Babe Ruth	.516	1926	31	NYY	AL	152	652	495	139	184	30	5	47	153	144		76	3	10			11	9	.372	.737	1.253	*79/H3
9	Barry Bonds	.515	2001	36	SFG	NL	153	664	476	129	156	32	2	73	137	177	35	93	9	0	2	5	13	3	.328	.863	1.379	*7/DH
10	Ted Williams	.513	1954	35	BOS	AL	117	526	386	93	133	23	1	29	89	136		32	1	0	3	10	0	0	.345	.635	1.148	*7/H
11	Babe Ruth	.513	1924	29	NYY	AL	153	681	529	143	200	39	7	46	124	142		81	4	6			9	13	.378	.739	1.252	*97/8H
12	Mickey Mantle	.512	1957	25	NYY	AL	144	623	474	121	173	28	6	34	94	146	23	75	0	0	3	5	16	3	.365	.665	1.177	*8/H
13	Babe Ruth	.512	1921	26	NYY	AL	152	693	540	177	204	44	16	59	168	145		81	4	4			17	13	.378	.846	1.359	*78/13
14	Rogers Hornsby	.507	1924	28	STL	NL	143	642	536	121	227	43	14	25	94	89		32	2	13			5	12	.424	.696	1.203	*4
15	Ted Williams	.499	1947	28	BOS	AL	156	692	528	125	181	40	9	32	114	162		47	2	1		10	0	1	.343	.634	1.133	*7
16	Ted Williams	.499	1942	23	BOS	AL	150	671	522	141	186	34	5	36	137	145		51	4	0		12	3	2	.356	.648	1.147	*7/9
17	Rogers Hornsby	.498	1928	32	BSN	NL	140	621	486	99	188	42	7	21	94	107		41	1	25			5		.387	.632	1.130	*4
18	Ted Williams	.497	1948	29	BOS	AL	137	638	509	124	188	44	3	25	127	126		41	3	0		10	4	0	.369	.615	1.112	*7/H
19	Ted Williams	.497	1946	27	BOS	AL	150	672	514	142	176	37	8	38	123	156		44	2	0		12	0	0	.342	.667	1.164	*7
20	Babe Ruth	.495	1931	36	NYY	AL	145	663	534	149	199	31	3	46	162	128		51	1	0			5	4	.373	.700	1.195	*97/H3
21	Babe Ruth	.493	1930	35	NYY	AL	145	676	518	150	186	28	9	49	153	136		61	1	21			10	10	.359	.732	1.225	*97/1
22	Arky Vaughan	.491	1935	23	PIT	NL	137	609	499	108	192	34	10	19	99	97		18	7	7		5	4		.385	.607	1.098	*6
23	Ted Williams	.490	1949	30	BOS	AL	155	730	566	150	194	39	3	43	159	162		48	2	0		22	1	1	.343	.650	1.141	*7

Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 8/12/2014.

At this point I intended to make insightful observations to relieve readers from making less insightful ones on their own.  But the null values in fields SH and SF again drew my attention.  I wondered about the results if I could somehow fill in the missing data.  That seemed like too much work and it only continued silly scoring rules that have varied over many decades.

Ted Williams back cover Baseball Digest May 1949
[Public domain] via Wikimedia Commons

Then I thought instead about my original idea of removing the offending data from all seasons and achieve these improvements:
1. make the data consistent
2. make the data make more sense.

The easy one to attack is the Sacrifice Fly (SF).  A batter should not get two credits for making an out: RBI and no at bat (AB).  The batter should certainly be charged with an AB.  That was done in many seasons as I will document below.

Sacrifice Hits is misnamed by today's rules as it means only sacrifice bunts.  But in some olden years it also meant sacrifice flies to advance a runner, some of which may have also scored a runner, like the current sacrifice fly.

It makes for a mess but it's also pretty interesting.  For instance, Ted Williams is the most recent batter with a batting average (BA) >+ .400: .406 in 1941.  Williams did that with zero SH and null SF.  I think several years ago someone did the research and determined that with the SF rule Williams would have had .414 BA.  Ironically in 1958 when Williams hit .388 and the SF was in effect, he had only two SF, so it made little difference: .386.

What I am suggesting would reduce averages as follows:

BA (H/AB): lower since SF would be added to AB and SH reduced and added to AB in years when it included advancing runners on a fly out

OBP (H+BB+HBP)/(AB+BB+HBP+SF): lower since SH reduced and added to AB in years when it included advancing runners on a fly out

SLG (TB/AB): lower since SF would be added to AB and SH reduced and added to AB in years when it included advancing runners on a fly out

OPS (OBP + SLG): lower

When I was considering a title for the original intent of this post I thought of simply writing about batters who reached base at least half the time but that would have required the stat reaching on an error, which is not part of OBP, so I dumped that idea.  Again, too much work.

Among the 23 seasons examined only Ted Williams (1954 and 1957), Mickey Mantle and Barry Bonds have non-null values in SF

In the 1920s Babe Ruth has many SH.  I'm guessing that none are bunts.  Same for Rogers Hornsby for 1924.

Oh, and while we're at it, use the same qualifying standard for averages for all seasons and base them on the denominator.  For BA, use AB, not PA.  Give Jimmie Foxx the highest BA (.364) in 1932 and another triple crown.  Dale Alexander in 1932 had 16 AB for Detroit, then 376 for Boston: .367 BA in 392 AB.  Ridiculous to consider Alexander the 1932 leader in BA.