Doing the Splits, SOM-Style
Subchances Key to Forecasting Batters' Hits
This iss one in a series of articles on forecasting Strat-O-Matic baseball (~ards. Bruce Bundy has been at it since 1968 and says he achieves up to 95 percent accuracy. But keep in mind that only the game coin company has the correct formulas, that many of them rely on statistics not readily available and that some ratings are subjective.
By Bruce Bundy
Theory and trial and error are the tools of the trade when creating math formulas that reflect reality. In SOM Baseball, for example, the three generations of X-Charts reflect 30 years of such formulating prowess. The first X-Chart addressed the theory that each defensive position and each player's defensive ability has a different impact. That chart quickly filled the need while other formulas were honed. The second X-Chart more accurately reflected the amount of errors a fielder would make. But, in accomplishing its goal, the X-Chart lost the ability to deliver a hit and an error on the same play. It was trial and error that led to a true separation of range and fielding and the third X-Chart.
Trial and error also reared its ugly in Part 1: Walks. Please correct the last formula in last month's issue (for league average walks) to read:
Al WALK = (((W -1W) * 216)1 (AB + (W -IW)))I 2
Advanced Notes: Hit by pitch (HBP) is used with walk in many instances.
FORMULA #2: HIT BY PITCH
HBP= ((HBP * 216)/(AB + HBP+ (W -1W)))
SOM HBP equals ((HBP times 216) divided by (At Bats plus HBP plus (Walks minus Intentional Walks)))
How it works: Because HBP only appears on the hitter's card, there is no pitcher's reduction, resulting in a pure proportioning of MLB PAs to SOM PAs. The result is rounded to the nearest whole number. If your computer won't round numbers off, add 0.5 to the end of the formula.
Now apply the walk formula to the advanced side. You will need stats for Lefty AL, righty AL, lefty NL, and righty NL totals, as well as individual lefty-righty totals. Actual HBP should be added (AB + (W-IW) + HBP))) to the walk formula's plate appearances when making advanced cards.
The theory of subchances results from a need to be more precise than 216 chances will accommodate. There are many forms of subchances in SOM Baseball. The 1-20 Split Deck is the foundation for many of them.
By dividing each chance into 20 subchances, 4,320 different results can be obtained (20 * 216 = 4320). SOM doesn't use subchances as often as it could, because rolling dice and drawing split cards every at-bat slows playing time and affects the game's appeal. Walks, for example, are not subchanced, but easily could be. Trial and error has proven that walk results stay fairly realistic despite not being subchanced. Theory suggests that the need for simplicity sometimes outweighs the good of making the result more precise.
Each subchance in the 1-20 split deck has the value of 1/20th, or ".05." Examples: A hatter's 3 column on "8-" reads: HOMERUN
2 chan~e~3 ch~nc'~4 chance~
1-4 DOUBLE 5-20. To determine how much the homerun is worth, multiply the 4 subchances by .05 for a result of .20. Because there are 5 chances to roll an "8-," multiply .20 by 5. The result is 1.00, 1 chance, or 20 subchances. The double's worth is 16 subchances, multiplied by 5 (80 subchances), and dividing by 20 (4 chances). Another example would he a pitcher's 6 column on "9-" reads:
SINGLE* 1-17 lineout (2h) 18-20. To determine the single's worth, multiply the chance value of "9-," which is 4, by the 17 single subchances, to equal 68 subchances, or, 3.40 chances.
Hits are subchanced. This is logical, since batting averages are based on I ,000 at-bats.
FORMULA #3: BATTER'S HIT
SOM Batter's Hit equals (((Batting Average minus .265) plus Batting Average) times (108 minus SOM Walk))
How it works: Notice that Formula #1 (and Formula #2 in advanced) must be done prior to Formula #3. The SOM WALK (and SOM HBP in advanced) must be formulated first, as these results are incorporated into Formula #3. The "hit" formula is a theory in two parts. The first theory is the hitter's card must compensate for the average pitcher's card. When a pitcher has 108 chances and a batter has 108 chances, if a pitcher surrenders a .250 batting average and the batter is supposed to hit .260, the batter's card must be able to hit .270 to offset the pitcher's limiting him to .250 half the time
The .265 used in the formula reflects a general league average. This number should be recalculated every year for each league. To correctly adjust, use the league's total batting average and add ".011." The ".011" added to the average reflects the frequency that Groundball A++ will also become hits. Twenty years of trial and error (the author's) is why the number is .011 and not .009. This year's AL Batting Average was .260, NL .252. When making 1991 AL batters, replace the .265 with .271, 1991 NL with .263.
Since walk chances are not at-bats, subtract the walk chances from the 108 card chances. Finally, the result is to be rounded to hundredths.
Advanced Notes: Use each AUNL lefty-righty league total to replace .265. Hit By Pitch result must he subtracted from 108 in this manner:
(((BA -.271) + BA) * (108 -(HBP + (W-IW))))
Continued on Next Page
1SF) STRAT FAN, February 1992
Forecasting Hitters' Singles, Extra-Base Hits
Continued from Previous Page
Cal Ripken hit .323 in 1991. His SOM walk is 3. The formula would read thusly:
(((.323 - .271) + .323) * (108-3)) = 39.35 or (((.375)) * (105)) = 39.35 Cal Ripken will have 42.35 on-base chances, 39.35 hit chances out
of 105 at-bat chances (minus the 3 walk). In another 1991 AL example, Cal's brother Billy played hurt and only hit .216. His SOM walk is 2. The formula would read thusly:
(((.216 - .271) + .216) * (108 - 2)) = 16.97 or (((.161)) * (106)) =
Poor Billy will have only 16.97 hit chances and only 18.97 total on-base chances on his card unusable but for his good glove.
The formula for hit chances has a 90 percent-plus accuracy rating.
Subchances make SOM click. Without them you may as well be playing APBA. From the final statistics of the 1991 AL season, SOM can give a different amount of bits to batters ranging from a lowly .136 to a violent .606. That's only part of the power of 4,320 subchances.
Imagine a batter who hits 1 triple in 287 plate appearances. If he faces the same pitchers, SOM can numerically exact the same result. SOM has the ability to take into account all of the pitchers faced by the batter, and, despite the low ratio of triples to plate appearances, a triple will still appear on the batters card. An amazing feat, considering there is 0.3 percent (0.0034843) chance for the triple to occur.
This is accomplished by proportioning, or sharing, the result. Most of the things that can happen in an at-bat can occur on both the pitcher's card and hitter's card. This is how players can keep their numeric identity. Have there ever been two completely identical SOM cards in history?
More amazing, SOM has unlimited ability to translate card formulas to computer. Today's impurities come from the data available and the computers themselves. Be assured that in the future's data-rich computer environment, this game can exact the result of every hitter's every plate appearance and every pitcher's every batter-faced.
While the computer's potential may be limitless, the cards have a mystique that the computer cannot provide. The SOM manager physically removes an ineffective pitcher, studies the bench for the right card, even rips up the second baseman who chokes in the clutch. Face-to-face in the big game, managers forget about wood and ink, it's flesh and blood. That's the strength of SOM. It comes from the power of subchances. Time to see how more subchances work.
FORMULA #4: BATTER DOUBLE
Double=((4320 * Double)I(AB+(W-IW)))-90
Batter's double equals ((4,320 times Double) divided by (At Bats plus (Walks minus Intentional Walks))) minus 90.
FORMULA #5: BATTER'S TRIPLE
Triple = ((4320 * Triple) I(AB + (W-IW))) - 15
Batter's Triple equals (4,320 times Triple) divided by (At Bats plus (Walks minus Intentional Walks))) minus 15.
FORMULA #6: BATTER'S HOMERUN
Homerun=((4320 * HR)/(AB+(W-IW)))-50
Batter's Homerun equals ((4,320 times Home Run) divided by (At Bats plus (Walks minus Intentional Walks))) minus 50.
Row it works: Extra base hits (XB H) have the same basic formula. The batter's XBH per plate appearance is proportioned to the 4,320 subchances. Again, the -90,-IS and -50 represent the amount of double, triple and homerun subchances that should appear on the average pitcher's card. About 20 percent of your calculations will result in negative numbers. Replace those with zero.
For more precise readings, the -90,-IS and -50 should be recalculated annually and by AL/NL. The formula to determine these variables vary slightly:
The +6 reflects the doubles that appear on the X-Chart. Rounded oft the AL effect is -100, the NL -92.
The 0.75 not only compensates for the X-Chart reduction, but for SOM' 5 apparent practice of having two-thirds of all triples occur off pitchers' cards.
No adjustments needed for HR.
Advanced notes: The formulas above apply, but remember to calculate annually, by AL/NL, and by LIR with lefty-righty data and league totals. Also include HBPs to the plate appearances:
MYP Cal Ripken tipped the scale at 55 with all his XBHs. His elementary side XBR total subchances should read: D=190; T=22; HR=161. Brother Billy's subchances are D=58; T=5; HR=-50, or 0. Remember to zero all negative results.
With a hitter's SOM hits, walks, (HBPs), doubles, triples and homeruns now formulated, a hitter's card begins to take shape. Next month's issue will dissect how SOM determines which card design each hitter receives.
Any comments? Write:
Bruce Bundy, 4474 Outlook Dr., Brooklyn, OH 44144.
About the STRAT FAN Cards In This Issue
This month marks STRAT FAN's second effort to supplement the 24-man rosters that have been a Strat-O-Matic characteristic for about 25 card sets.
In October we offered 36 pitchers to fill out the 1978 staffs that needed the arms for proper play with the computer manager. This time we offer hitters and pitchers - 48 of them - for the 1984 American League, the year of the awesome Detroit Tigers, who went wire to~ wire for the AL East crown, then crushed Kansas City and San Diego
in the postseason. -,
There are 96 more cards for 1984 ahead in future issues of STRAT ~.:
FAN-National League pitchers and hitters and individual hitfing cards for NL pitchers. The computer manager needs the extra pitchers
for 1984, too. STRAT FAN intends to follow suit for as many past 4 seasons as possible.
These additions boost the 1984 set to more than 770 cards - an especially valuable tool for full-season replays, which are becon'ing much more common now that the computer game speeds play and stat-keeping.
The statistics for creating these cards were compiled by Luke Icraemer of Beaverton, OR, who is making his research available for purchase.
FOR SALE: Never before published: Leftyirighty stats for all 1984 SOM and STRAT FAN baseball players. $12. Luke Kraemer, 625 NW waterhouse Ave., Beaverton, OR 97006.