A look at offensive efficiency

Much of the available baseball related data and statistical analysis is very useful for evaluating individual performance. But trying to evaluate teams is an inherently more complex endeavor. Nevertheless, understanding how team performance on offense and defense contributes to winning is critical. On defense a simple measure of efficiency (Defensive Efficiency) was developed that tells us the rate at which a team converts balls put in play into outs. A defensive efficiency of 71% is around average, over 73% is considered excellent, and below 69% is terrible. While this efficiency metric does not necessarily capture the nuance of defense, it can be used to broadly assess a team's defense and has a relationship* with winning. What about a measure of offensive efficiency?

The basic idea of an offensive efficiency measure would be to examine the rate at which teams turn baserunners into runs and there have been some strong efforts made to pin down a measure of offensive efficiency. The basic idea is to examine if a team has scored as many runs as would be expected given their number of singles, doubles, triples, home runs, etc. We know that getting on-base is the most critical aspect of a plate appearance and having more baserunners should lead to more runs. But perhaps there are teams that are doing more with less.

One way to approach to this question is to compare a team's actual runs scored with their weighted runs created (wRC). wRC gives us a measure of how many runs a team should have scored given their offensive output (i.e., singles, doubles, etc). The ratio of these numbers will give us an indication of the extent to which a team is outperforming their expectation, which can be taken as an index of efficiency. Below are the ‘efficiency' rankings according to this analysis for the 2013 season:

Rank	Team	Actual Runs	wRC	Difference	Efficiency
1	STL	783	715	68	109.5
2	NYY	650	603	47	107.8
3	HOU	610	583	27	104.6
4	NYM	619	594	25	104.2
5	MIA	513	493	20	104.1
6	KCR	648	624	24	103.8
7	BAL	745	719	26	103.6
8	OAK	767	743	24	103.2
9	CLE	745	725	20	102.8
10	WAS	656	646	10	101.5
11	TOR	712	704	8	101.1
12	TEX	730	722	8	101.1
13	PHI	610	604	6	101.0
14.	SDP	618	612	6	101.0
15	CIN	698	692	6	100.9
16	CHW	598	597	1	100.2
17	ATL	688	689	-1	99.9
18	ARI	685	690	-5	99.3
19	LAA	733	741	-8	98.9
20	MIL	640	647	-7	98.9
21	BOS	853	863	-10	98.8
22	COL	706	719	-13	98.2
23	SFG	629	645	-16	97.5
24	CHC	602	618	-16	97.4
25	SEA	624	642	-18	97.2
26	PIT	634	658	-24	96.4
27	TBR	700	733	-33	95.5
28	MIN	614	645	-31	95.2
29	DET	796	838	-42	95.0
30	LAD	649	685	-36	94.7

Cardinals and Yankees on top? This looks like an interesting measure of efficiency. Well not really if we consider how it relates to winning games. There are only two playoff teams in the top 10, and look at some of those teams in the top 10: Astros, Mets, Marlins, Orioles. At the bottom, on the ‘inefficient' side of things, we have a bunch of playoff teams: Red Sox, Dodgers, Rays, Pirates and Tigers. The correlation between this efficiency and the team's winning percentage is -0.002. Thus, while this measure is interesting it does not seem to be capturing something about winning, but perhaps is indexing a more random component of performance like luck.

Let's return to the initial idea of developing a defensive efficiency-like measure and look at rates for converting baserunners to runs as a way to define a measure of offensive efficiency. For any team we can approximate their number of baserunners with PA*OBP. We can then look at the rate at which baserunners score by taking the ratio of runs and baserunners. Here is how this looked in 2013:

Rank	Team	PA	OBP	Baserunners	Runs	Efficiency
1	BAL	6144	0.313	1923	745	38.7
2	BOS	6382	0.349	2227	853	38.3
3	STL	6202	0.332	2059	783	38.0
4	OAK	6209	0.327	2030	767	37.8
5	CLE	6165	0.327	2016	745	37.0
6	TEX	6196	0.323	2001	730	36.5
7	TOR	6152	0.318	1956	712	36.4
8	DET	6388	0.346	2210	796	36.0
9	LAA	6260	0.329	2060	733	35.6
10	COL	6152	0.323	1987	706	35.5
11	NYY	6044	0.307	1856	650	35.0
12	ATL	6133	0.321	1969	688	34.9
13	WAS	6047	0.313	1893	656	34.7
14	TBR	6242	0.329	2054	700	34.1
15	CIN	6293	0.327	2058	698	33.9
16	HOU	6020	0.299	1800	610	33.9
17	MIL	6064	0.311	1886	640	33.9
18	KCR	6093	0.315	1919	648	33.8
19	ARI	6334	0.323	2046	685	33.5
20	PHI	6014	0.306	1840	610	33.2
21	CHC	6079	0.300	1824	602	33.0
22	PIT	6135	0.313	1920	634	33.0
23	SEA	6172	0.306	1889	624	33.0
24	SDP	6122	0.308	1886	618	32.8
25	CHW	6077	0.302	1835	598	32.6
26	NYM	6207	0.306	1899	619	32.6
27	LAD	6145	0.326	2003	649	32.4
28	SFG	6168	0.320	1974	629	31.9
29	MIN	6212	0.312	1938	614	31.7
30	MIA	6021	0.293	1764	513	29.1

This looks a little better than our previous measure. Five of the top 10 teams are playoff teams, while the bottom 10 only has two playoff teams (Dodgers are down there again). The correlation between this measure and the team's winning percentage is 0.586, which is a pretty solid relationship and shows that efficiency is accounting for a reasonable proportion of the variance in winning. Thus it seems as though there are efficient teams and being efficient relates to winning.

We can examine if this simple measure of efficiency's relation to winning is evident in a larger sample. To do so I accumulated team data back to 1974. With this larger sample the correlation drops slightly to 0.463, which is again a pretty solid relationship. Out of interest here are the 10 most efficient teams from this sample:

Rank	Team	Year	R	BaseRunners	Efficiency (%)	Win%
1	CLE	1994	679	1577	43.1	0.584
2	CHW	2000	978	2282	42.9	0.586
3	COL	1996	961	2248	42.7	0.512
4	MIL	1982	891	2123	42.0	0.583
5	CHW	2004	865	2064	41.9	0.512
6	BAL	1996	949	2273	41.8	0.540
7	TEX	2004	860	2058	41.8	0.549
8	TEX	2005	865	2073	41.7	0.488
9	SEA	1996	993	2386	41.6	0.528
10	COL	2000	968	2336	41.4	0.506

These teams were bringing home over 40% of their baserunners. That is very impressive. One thing that the teams on this list likely make you think of is power. Most of these teams played during the ‘steroid era'. These 10 teams hit a lot of home runs. The average HR total per season for this group is 221.2, which is well above the per season average for the entire sample (146.9).

Here are the 10 least efficient teams:

Rank	Team	Year	R	BaseRunners	Efficiency (%)	Win%
1	SEA	2010	513	1785	28.7	0.377
2	CIN	1982	545	1891	28.8	0.377
3	SDP	1975	552	1908	28.9	0.438
4	LAD	1992	548	1890	29.0	0.389
5	NYM	1981	348	1197	29.1	0.390
6	SDP	1981	382	1311	29.1	0.373
7	MIA	2013	513	1764	29.1	0.383
8	SDP	1980	591	2027	29.2	0.448
9	SDP	1974	541	1848	29.3	0.370
10	CAL	1976	550	1868	29.4	0.469

These teams were converting fewer than 30% of their baserunners to runs and losing a lot of games. They did not hit HRs with an average season total for this group coming in at 74.6. Note that almost all of these inefficient teams played during the late 70s - early 80s; an era that did not have the same run-scoring as the late 90s and 2000s, which means some sort of era-adjustment would be useful going forward.

The best correlates of this offensive efficiency measure are SLG (r = 0.887), ISO (0.836), wOBA (0.812), HR (0.767), wRC (0.740), OBP (0.666), RE24 (0.584), and GUILLEN# (0.581). These relationships should not be particularly surprising, as the metrics involve getting runners on base and/or moving them around. The contribution of this offensive efficiency measure is the demonstration that there is important variance in a team's ability to take advantage of baserunners in order to score runs and win games. In fact, this offensive efficiency correlates better with winning than does runs scored (RS) and all of the above metrics with the exception of OBP, wOBA, and RE24. Teams that combine high rates of OBP with high rates of efficiency will likely find success, but some of that is also contingent on their ability to prevent runs as well.

Conclusion

While the measure of efficiency given here is certainly not perfect, the idea of offensive efficiency is an interesting aspect of team performance that could be fruitful for continued investigation. Examining park-adjusted, or era-specific efficiencies could add to the options for assessing team performance.

* r = 0.341 for the sample (1974-2013) discussed here.

. . .

All statistics courtesy of FanGraphs, Baseball Prospectus and Baseball-Reference.

Chris Teeter is a contributor to Beyond the Box Score. You can follow him on Twitter at @c_mcgeets.