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 |
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.
Get Answers For Free
Most questions answered within 1 hours.