Question

Part II

We will use the “Twins Data” tab in the workbook.

1) Single Variable

              a) Create a Scatterplot of “Wins” and “Runs” (You might need to rescale the axis for each)

              b) Run a Regression with “Wins” as y and “Runs” as x

c) What is your model? Slope t-value? F-Value? R squared?

2) Multivariable

              a) Traditional Stats

                             Run a regression with “Wins” as the y variable and both “Batting Average” and “ERA”

as the two x variables

What is your model? Slope t-values? F-Value? R squared?

              b) Moneyball Stats

                             Run a regression with “Wins” as the y variable and “OPS” and “WHIP” as the x variables

What is your model? Slope t-value? F-Value? R squared?

3) Of the 3 options which model do you feel works the best? Explain.

Base Data				Tradititonal			SABREmetric
Year	Wins	Runs		Batting Average	ERA		OPS	WHIP
2000	69	748		0.270	5.140		0.744	1.501
2001	85	771		0.272	4.510		0.770	1.345
2002	94	768		0.272	4.120		0.769	1.310
2003	90	801		0.277	4.410		0.772	1.319
2004	92	780		0.266	4.030		0.763	1.324
2005	83	688		0.259	3.710		0.714	1.233
2006	96	801		0.287	3.950		0.771	1.283
2007	79	718		0.264	4.150		0.721	1.340
2008	88	829		0.279	4.160		0.748	1.353
2009	87	817		0.274	4.500		0.774	1.382
2010	94	781		0.273	3.950		0.762	1.291
2011	63	619		0.247	4.580		0.666	1.438
2012	66	701		0.260	4.770		0.715	1.391
2013	66	614		0.242	4.550		0.692	1.413
2014	70	715		0.254	4.570		0.713	1.391
2015	83	696		0.247	4.070		0.704	1.330
2016	59	722		0.251	5.080		0.738	1.453

Answer 1

1) Here is the regression analysis for Win as response and Runs as predictor.

The model is Wins = -23.889 + 0.1408*Runs, slope t-value = 4.153, F-value = 17.25, R-squared = 53.48%

2)a) Here is the regression analysis

The model is Wins = -14.3375 + 549.75*Batting Average - 18.186*ERA, slope t-values = 6.131(Batting Average), -6.315(ERA), F-value = 54.37, R-squared = 88.59%

b) Moneyball Stats:

The model is Wins = -90.1849+ 190.828*OPS - 110.896*WHIP, slope t-values = 4.784(OPS), -5.732(WHIP), F-value = 41.97, R-squared = 85.71%

3) Based on R-square I think Traditional model is most useful.