BronxBomber Tribune

The relationships between RPS and Movement of Balls

This article is for english users, Korean users go to this link.

I picked 22starters for my analysis, and if there was no pfx logs, I excluded them.

(for example, there is no pfx logs in Tyoko Dome Opener)

The criteria for these 22 starters are,

1. Starters who has the most Ks

2. Starters who showed impressive perfomance in April 2008.

3. Starters who was the Cy-young nominees(or got Cy-young award)

I got 9818 sample balls from these starters and 5965 of them were fastballs, and 5800 balls were remained when I excluded datas that was located out of the standard deviation(understable ranges, not the theoritical sigma range). And These are the results of their fastball average values.

Name	Team	SPEED(KMH)	STDEV	DIFF	PFX	BRK	RPS	FA#	Balls#	FA%
J. Weaver	ANA	90.4(145.46)	1.83	8.48	13.01	2.93	41.02	169	497	34.00%
D. Haren	ARI	90.34(145.36)	1.74	7.82	12.99	4.48	41.79	203	416	48.80%
D. Cabrera	BAL	93.94(151.15)	1.62	9.10	12.34	4.62	40.70	399	489	81.60%
D. Matsuzaka	BOS	91.07(146.53)	1.46	9.37	12.51	4.29	39.83	207	394	52.54%
J. Beckett	BOS	95.17(153.13)	1.90	9.01	12.52	4.91	42.09	207	268	77.24%
C. Zambrano	CHC	90.34(145.36)	2.05	6.44	11.30	5.77	36.95	346	471	73.46%
J. Cueto	CIN	93.04(149.7)	1.39	7.30	11.23	3.09	37.63	264	439	60.14%
A. Harang	CIN	88.59(142.54)	1.97	6.95	12.81	3.84	40.72	282	460	61.30%
C. Hamels	PHI	88.26(142.02)	2.08	6.63	12.93	3.74	41.19	281	525	53.52%
C.C. Sabathia	CLE	93.6(150.61)	1.28	7.93	11.77	4.34	39.27	320	458	69.87%
C. Lee	CLE	89.95(144.73)	1.35	7.76	13.82	4.15	44.24	349	412	84.71%
J. Verlander	DET	93.31(150.14)	2.00	8.99	15.58	5.59	51.15	195	489	39.88%
R. Oswalt	HOU	92.19(148.33)	1.35	7.98	10.80	4.53	35.47	274	467	58.67%
B. Sheets	MIL	92.16(148.28)	1.19	7.36	12.11	3.28	40.11	217	378	57.41%
J. Santana	NYN	90.64(145.84)	1.56	7.29	11.23	4.63	36.47	246	440	55.91%
J. Peavy	SDN	93.24(150.02)	1.27	9.10	12.48	4.78	40.86	253	547	46.25%
F. Hernandez	SEA	94.89(152.69)	1.35	8.74	11.14	4.80	37.35	319	542	58.86%
M. Cain	SFN	92.91(149.5)	1.64	8.62	12.89	3.16	42.31	346	474	73.00%
J. Sanchez	SFN	90.12(145)	2.26	7.97	12.13	5.16	38.83	319	383	83.29%
T. Lincecum	SFN	95.47(153.62)	1.72	8.72	12.73	2.93	42.99	335	493	67.95%
R. Halladay	TOR	92.08(148.16)	1.73	8.11	10.41	6.23	34.04	222	403	55.09%
D. McGowan	TOR	94.89(152.68)	1.66	9.48	12.77	3.84	42.43	212	373	56.84%

Speed is described in MPH and numbers in ( ) is the the unit of KMH(Kilometers per Hour, the Korean unit). Stdev is the standard deviation of start speed, DIFF is the diffrerence between average speed of start and end speed. And maybe you already know what pfx and brk(break) means. FA# is the number of fastballs, and Balls# is the number of all balls in their log. FA% is the portion of fastballs. RPS is the Rotation per Second(not RPM, R per minute)

I tried to get relationships between start speed and RPS of fastballs. My hypothesis was,

"The Faster the ball is, the more spin it has to be"

But I found it was silly when I saw my results. See the graph below.

Show the value of R square. The tendency line(I don't know the exact terms to decribe this, maybe trend line?) has no meaning to this graph.

BECAUSE, each pitcher has their unique movement in their fastball and each pitcher has their unique RPS in their fastball(regardless of the speed)

For example, there is the pitcher who has average start speed of 93 miles and has RPS of high 30s(the case of Cueto), there is pitcher like average speed of 93 MPH and RPS over 50(case of Verlander)

Let's see this in picture.

As you see, there is many pitchers located in speed of 90 to 95 mph, and 35 to 45 in rps. What I want to say is, the rps of ball IS NOT AN ABSOLUTE CRITERIA FOR SPEED, there's somewhat other varibles effect on the value of rps.

So, I merger 5800 fastballs in to one excel sheet and separated them to 1 mile criteria.(85 to 97 in unit of 1mile per hour)

TO 85 to 97 MPH, there was 12 intervals and 5500 samples, and their relationship was very interesting to me. Let's see that.

The value of Y axis is RPS, ans value of x-axis is decribed in the picture.

As you see above, the value of R square has very siginificnat meaning compared to speed-rps relationships. And the relationship was decribed in every 12 intervals.

MOVEMENT - RPS Relationship
speed range	Formula	R^2	Sample#
85-86	3.1525x-3.67	0.8348	101
86-87	3.0805x-2.7342	0.9030	115
87-88	3.3946x-6.5701	0.9138	192
88-89	3.2884x-4.5288	0.9187	370
89-90	3.1882x-2.662	0.9244	486
90-91	3.249x-3.0762	0.9329	571
91-92	3.3829x-4.3772	0.9146	745
92-93	3.3303x-3.2576	0.9054	803
93-94	3.3855x-3.6724	0.9192	741
94-95	3.2175x-1.1906	0.9486	672
95-96	3.0155x+0.9232	0.9696	448
96-97	3.2262x-0.315	0.9582	270

I got this formulas but I don't know how to apply these into speed-rps-movement relationship.

But I got one important lessons from today's analysis.

"RPS is not an absolute guideline for ballspeed. But we can expect that the more rps it has, the more movement it has to be."

And in datas above, horizontal movement value has the big portion on them, so I have to correct or revise on sinking 2-seamers like fastballs by Halladay, Wang.

Thank you for reading.

Mingu, Song.

If you have any questions about this article, contact me landor82 at gmail.com or write comments below in the blue box(click the blue button when you finished writing your comment.)