Building A Farm: A Summary

We’ve spent the past few weeks taking a look at combined rankings for each organization, going division-by-division. I wasn’t really sure what I was going to find, but my goal was to take a look at the two main overall aspects of a prospect – his talent/reasonable ceiling and his risk of getting there – and see how farm systems graded out. The traditional 1, 2, 3 ranking system is fine because we’re ultimately looking at an educated subjective process, but a simple list doesn’t show the audience where the real gaps lie and where there’s negligible difference. My hope was to begin to approach a way to see these differences, and while there is certainly room for improvement, I believe it has led to some interesting results.

Averaged Publication Lists

Team	BA	BPro	Sic	Law	FG	BB	AVG	RANK
St Louis	1	1	1	1	1	6	1.83	1
Texas	3	2	4	9	2	2	3.67	2
Seattle	2	5	2	8	7	1	4.17	3
Tampa Bay	4	8	3	3	3	9	5.00	4
Minnesota	10	4	7	2	4	5	5.33	5
Pittsburgh	7	6	5	7	6	3	5.67	6
Houston	9	9	11	4	10	4	7.83	7
San Diego	14	3	6	6	8	11	8.00	8
Chicago (NL)	12	12	10	5	5	13	9.50	9
Miami	5	11	8	16	12	15	11.17	10
Boston	6	16	9	17	17	8	12.17	11
New York (NL)	16	10	12	14	14	7	12.17	12
New York (AL)	11	14	14	10	16	12	12.83	13
Arizona	8	17	13	15	13	16	13.67	14
Kansas City	18	7	21	11	11	14	13.67	15
Cincinnati	15	15	15	12	9	19	14.17	16
Baltimore	17	20	18	13	15	10	15.50	17
Los Angeles (NL)	19	21	19	18	18	22	19.50	18
Toronto	22	13	22	24	20	17	19.67	19
Cleveland	20	19	24	19	21	18	20.17	20
Washington	13	23	25	21	19	21	20.33	21
Colorado	21	22	16	23	26	20	21.33	22
Atlanta	26	18	27	20	22	24	22.83	23
Oakland	25	25	26	22	25	23	24.33	24
Philadelphia	24	24	20	27	24	28	24.50	25
San Francisco	28	26	17	26	27	25	24.83	26
Milwaukee	23	27	23	29	29	26	26.17	27
Chicago (AL)	29	28	28	28	23	30	27.67	28
Detroit	27	29	30	25	28	27	27.67	29
Los Angeles (AL)	30	30	29	30	30	29	29.67	30

The above chart is just an averaged ranking of the publications that I used for each prospect list. The “AVG” column is there average placement on these lists, and the “RANK” column is the list method of ranking.

There don’t seem to be any real surprises as you move down the list. St. Louis was ranked first by every publication but Bullpen Banter, and it has a pretty clear advantage over Texas and Seattle, who seem to have very fairly farm systems. Moving down the list, the next three are pretty close and can probably considered interchangeable in the 4, 5, and 6 spots. There seems to be a drop off from those systems to the next 3, and it drops off again to the next 3-7 systems. Baltimore seems to be a cut-off point for the upper half of farm systems, though it’s not the exact halfway point.

The bottom half starts with five teams that are pretty close together, and the next 3 after that follow suit.  Milwaukee, Chicago (AL), and Detroit are a bit worse than the others, and the Angels come in dead last by a substantial margin. The Angels grabbed the bottom ranking on 4 of 6 lists with Chicago and Detroit getting the others.

Putting Those Against the Averaged Grades and Risks

Team	RANK	Sys Gd	Sys Rs	50+	50+ Gd	50+ Rs	60+	60+ Gd	60+ Rs
St Louis	1	51.000	2.333	20	54.250	2.550	4	66.250	1.500
Texas	2	53.417	2.833	27	54.537	2.926	6	64.167	3.000
Seattle	3	51.250	2.500	22	53.523	2.727	3	65.000	3.000
Tampa Bay	4	53.250	2.800	29	53.534	2.828	3	63.333	2.000
Minnesota	5	51.583	2.867	23	53.587	2.957	4	64.375	2.875
Pittsburgh	6	51.129	2.677	22	53.864	2.955	6	63.750	2.833
Houston	7	51.694	2.645	20	55.375	2.900	4	66.875	2.875
San Diego	8	50.500	2.633	22	52.500	2.818	2	63.750	3.000
Chicago (NL)	9	51.333	2.700	21	54.048	3.048	4	63.750	3.000
Miami	10	51.083	2.750	20	54.000	2.950	4	63.125	2.250
Boston	11	53.103	2.828	27	53.611	2.889	5	62.500	2.400
New York (NL)	12	50.530	2.788	25	52.300	3.000	3	64.167	2.167
New York (AL)	13	52.000	2.767	23	54.130	3.000	4	61.250	3.125
Arizona	14	50.152	2.727	22	52.614	2.818	2	66.250	2.000
Kansas City	15	52.667	3.100	27	53.519	3.222	4	65.625	3.250
Cincinnati	16	51.034	2.707	22	53.182	3.068	3	63.333	3.000
Baltimore	17	48.583	2.617	15	53.167	2.900	2	71.250	1.750
Los Angeles (NL)	18	50.000	2.467	19	52.895	2.842	1	62.500	3.500
Toronto	19	52.222	3.130	22	53.864	3.295	5	61.000	3.800
Cleveland	20	50.081	2.758	20	52.750	3.025	3	64.167	3.000
Washington	21	50.645	2.823	23	52.826	3.065	3	65.000	3.333
Colorado	22	49.917	2.717	22	51.932	3.023	2	62.500	3.250
Atlanta	23	50.268	2.982	20	52.375	3.025	1	60.000	2.000
Oakland	24	49.907	2.833	17	52.941	3.059	1	65.000	2.500
Philadelphia	25	50.000	2.767	19	53.158	3.000	1	60.000	3.500
San Francisco	26	49.167	2.617	18	52.222	2.750	1	62.500	3.000
Milwaukee	27	50.083	2.700	21	52.500	2.857	1	60.000	2.000
Chicago (AL)	28	47.742	2.710	14	51.071	3.071	1	60.000	3.000
Detroit	29	48.250	2.900	16	51.563	3.188	1	62.500	2.000
Los Angeles (AL)	30	47.333	2.817	16	51.094	3.188	1	62.500	3.000
Average		50.664	2.750	21.1	53.098	2.965	2.8	63.547	2.730

This is the part about which I was very curious. The publications rank the systems, but how do they look when broken down? Again, the grades and risks are only based off of two lists – Baseball America and Baseball Prospectus – but they give us a decent overall look. Just for giggles, let’s look at how each column correlates with the overall ranking from the above section.

	Sys Gd	Sys Rs	50+	50+ Gd	50+ Rs	60+	60+ Gd	60+ Rs
Correlations	-0.731	0.288	-0.609	-0.758	0.507	-0.744	-0.440	0.145

When looking at the “grade” and number of prospect columns, we expect to see negative correlations because, as the ranking goes up, we expect to see lower grades and number of prospects. When looking at the “risk” columns, we expect to see positive correlations because the risk will theoretically grow as the ranking grows.

The highest correlation belongs to the 50+ grade – the averages of all 50+ grade prospects in a system. This grade looks at the quality of the prospects who are more likely to be average major-league players, and these are theoretically the most likely to make the majors in some capacity.

The next highest was a bit of a surprise to me – the number of 60+ grade prospects. What evaluators seem to be valuing is simply having these higher ceiling players in the system, and as you can see, the risk of these players doesn’t seem to matter that much as it nets the lowest correlation. Higher ceiling prospects are so valuable because they are the players around which a team can build a franchise. Some will certainly bust, but the ones that make it through the attrition can become core contributors, while the lower-ceiling prospects likely won’t have the same impact on the organization. Having more is obviously a good thing as those AT&T commercials tell me.

It is important to remember what the correlations are comparing. They are comparing what the publications seem to favor when ranking the overall farm systems, not the effectiveness of each column. Other research will need to be done into whether or not they favor the correct things, and that future research is one of the reasons I started this project. It’s a starting point. What really is more valuable between ceiling and risk? What types of systems flourish? Should a team sell out for ceiling, or is there a certain degree of risk that’s not worth taking? And do teams have a particular strategy, and how effective is it?

I don’t have the answers to these questions at the moment, but I hope that I have begun the process in a valuable way. Hopefully, more lists will begin to note the overall grade and risk of each prospect because more grades will help even out the outliers and give us a better measurement. If anyone has any comments or suggestions on how to improve this project for the future, please leave a comment.