Building A Projection System

Rk Name Pos Act Pec My E P E My 1. Rodriguez 3b .340 .319 .310 .021 .030 2. Ramirez ss .315 .277 .302 .038 .013 3. Renteria ss .297 .262 .259 .035 .038 4. Rollins ss .290 .274 .272 .016 .018 5. Jeter ss .285 .305 .277 .020 .008 6. Guillen ss .283 .306 .291 .023 .008 7. Reyes ss .278 .276 .266 .002 .012 8. Tejada ss .271 .296 .289 .025 .018 9. Young ss .270 .286 .275 .016 .005 10. Wilson ss .269 .247 .246 .022 .023 11. Eckstein ss .266 .247 .257 .019 .009 12. Greene ss .263 .272 .266 .009 .003 13. Hardy ss .261 .254 .241 .007 .020 14. Sea Bass ss .260 .247 .233 .013 .027 15. Cabrera ss .260 .260 .253 .000 .007 16. Peralta ss .259 .281 .265 .022 .006 17. Loretta ss .254 .252 .250 .002 .004 18. Bartlett ss .253 .269 .252 .016 .001 19. Betancourt ss .248 .251 .244 .003 .004 20. Scutaro ss .246 .253 .246 .007 .000 21. Furcal ss .244 .268 .278 .024 .034 22. Lopez ss .239 .272 .268 .033 .029 23. Drew ss .236 .276 .289 .040 .053 24. Durham 2b .227 .295 .269 .068 .042 25. Lugo ss .225 .269 .261 .044 .036 26. Uribe ss .222 .263 .228 .041 .006 27. Vizquel ss .221 .264 .242 .043 .021 28. Crosby ss .219 .265 .247 .046 .028 29. McDonald ss .211 .215 .215 .004 .004 30. Izturis ss .210 .234 .221 .024 .011 Average ss .257 .269 .260 .023 .017

As I sit here working on a simple projection system to evaluate translations from Japan to the United States, I beta ran one of the simple methods I came up with. The method is based on Marcel and I was only trying to project Equivalent Average. I looked at most middle infielders from the 1990s and developed a simplistic general age curve for all of them. Fitted that using a similar weighted season process that Marcel uses. I then looked at the set of 2007 SSs with a large amount of PAs and compared the projections versus the actual results for PECOTA and the simplistic method I came up with. Surprisingly the method I devised was more accurate. Weird. In case you're interested, the results are to the right.