Methodology to Create Combo Cards

• Tally chances on each card (including endurance, ball park, weather, clutch, and injury) along with miscellaneous data (stats, fielding, bunting, etc.).
• Using Baseball Reference, find out how many PA's (or BF for pitchers) the player had for each league (also confirm actual combined stats). Determine what percentage of PA or BF was played against AL and NL, respectively.
• Calculate the overall chances vs L and vs R, for each type of outcome, based on those percentages.
  
  For example, 1973 Jim Fregosi had the following for SINGLE(LF):
  
  

  League	PA's	Pct.	SI(LF) chances vs L	SI(LF) chances vs R
  AL	173	.545	2.0	5.0
  NL	144	.455	9.0	0.0

   
  (2.0 * .545) + (9.0 * .455) = 5.185
  (5.0 * .545) + (0.0 * .455) = 2.725
  

  
  Roundest to the nearest .10, Fregosi would have 5.20 SI(LF) vs L and 2.70 SI(LF) vs R on his MLB card.

Total chances usually hit 108, but sometimes a tad over or under. In those cases, add or subtract lineouts to get to 108. OB always remains the same. K's, BB's, and HBP's need to be whole numbers. If they are not, for K's adjust lineouts; for BB's and HBP's, adjust SI*. Again, keeping the OB the same. Some chances may be consolidated that do not impact the game, such as lineouts and popouts.

Because the data is pretty precise, there are more splits than normal, but never more than ten. Splits are realistic, meaning you will not see something like HR 1-4, lo(ss) 5-20.

For the sake of making some chances fit, occasionally small adjustments are made: a .1 or -.1 here or there. For example, SI(lf) -.1, SI(cf) +.1 (and sometimes a little larger to get the BP single 2 chance to fit). LO's, PO's, GB A's, GB B's, GB C's, FB B's, and FB C's may not exactly match position (i.e. P,C,1B,2B,3B,SS,LF,CF,RF) calculations, but attempts are made to get them as close as possible. Note: There are less hit split opportunities for batters and pitchers with hit chances and low AB and IP, respectively. For them, larger adjustments need to be made to the type of outs to fit the hit chances. There are no splitting of outs. Splits of a DOUBLE(of) are not accepted by the game engine. Some of these may be converted to DOUBLE. The game engine does a good job of letting you know if an input is incorrect.

The same percentage logic is used to determine the MLB miscellaneous data too, like fielding range, error rating, running, bunting, balks, wild pitches, etc. If a player was C bunt in AL, and B bunt in NL, use the rating where he played the most. Sometimes you have to make a judgment call. Of course you are free to update the player yourself.

All players are placed on team "MLB".

Note: For players with very few PA's or IP's in one league, the above methodology is bypassed. Instead, minor adjustments are made to the larger used card.

Does this methodology work? After doing many simulations, normal, expected results are seen. Results fluctuate from simulation to simulation. Really no different than other simulations with game provided cards. The key is the methodology takes into effect the work SOM puts into making the cards realistic to actual results. For example, groundball pitchers will have the appropriate groundball chances. Ditto for flyball pitchers, groundball hitters, etc. This data cannot be surmised by just creating a combo card based on stats alone.

Feedback and comments are most definitely welcomed, and encouraged.

Thanks
Jim

P.S. Little tidbit... did you know that the Sac Fly % number, listed on the card image, does not include flyball B splits? I started freaking out thinking the game was not capturing my flyball B splits (there are more due to the methodology). No worries, I have tested and the game engine sees them. Try it yourself!