SOM HR (in)accuracy
Posted: Sun Jun 01, 2008 10:55 pmI can still remember the ad in a baseball magazine from the pre-season of 1969: “Manage your own big league team! Players perform just like they did in real life!” I could not imagine how a game could recreate statistical accuracy… until I received those cards and played the seasons myself. I remember replaying the 1970 Reds against the National League (18 games vs. division rivals and 12 games vs. the East). Rose hit .315 for me, one point below his real life average; Tolan hit .316 on the button. Johnny Bench hit 48 homers (up from 45); Perez hit 40 to match his real life totals. The Reds even matched their real life record of 102-60. The most exciting thing about this replay for me was how uncannily accurate the statistics were.
Over the years, SOM added increasing complexity—the company said “realism”. But as I played the game, I found it less and less statistically accurate with each addition. Suddenly, lefty-righty balances could be exploited; one could stack a lineup with righties against a righty who did poorly against righties, even though teams used predominately left-handed hitters against him in real life. Soon, ballpark homers and ballpark singles were born—an idea that sounded good to me until I learned that a ballpark’s influence on homers could vary from year to year (without any changes to the dimension of the field)!
Suddenly, players produced ridiculous numbers in replays, taking away the amazement of watching SOM stats parallel those of real life.
And then I found SOM on-line! I was EXCITED! I was looking forward to statistically accurate seasons with my favorite players of all time.
Now, after completing 11 seasons with the 1969 teams, I have some observations/reservations:
In 1969, the NL hit 1470 home runs and the AL hit 1649 (an average of 1560). One would expect TSN/SOM on-line stats to be similar for 12 team leagues. Instead, here are the league home run totals for the 11 leagues I have participated in (The percentages following the HR totals are the percent over the actual average of the 1969 AL/NL):
1949 25%
2301 48%
2281 46%
2229 43%
2391 53%
2514 61%
2275 DH 46%
2505 100M 61%
2202 DH 41%
2586 66%
1545 -1%
All of these seasons have been 80M leagues, except the one indicated. Only two of them have been DH leagues.
What accounts for this absurd inaccuracy in home run totals in what is supposed to be the gaming world’s most accurate game? Could it be that everyone is playing in Fulton? Not so. I checked the leagues and there is some tendency for managers to select HR ballparks, but not enough to warrant this inaccuracy. In fact, three of the leagues above had unique stadium rules, allowing only one Fulton, one Wrigley, etc. into the league.
And so, I suggested an experiment: put everyone in the league into Municipal Stadium (HR 1-3 vs. LH and RH). The league was conducted through www.somdraft.com (a live draft engine). Lo and behold, something amazing happened: hitting statistics were accurate. In fact, the final league listed above (1545 HRs) is this league in which everyone played in this “extreme pitcher’s park”.
Want more accuracy? The team that led the league in runs (802) nearly matched both the AL leader (790) and the NL leader (798). More importantly, individual batters performed statistically like they did in real life. Here is a list of some of the leading HR hitters from the Municipal Stadium League. I left out certain hitters because they were traded (Petrocelli, Killebrew) or because they were used far more than in real life (Held, McAuliffe).
Player Homers in MSL (Homers in real life)
McCovey 44 (45)
May 33 (38)
Wynn 32 (33)
Jackson 44 (47)
Stargell 34 (29)
Allen 35 (32)
Aaron 26 (44)
Bonds 37 (32)
Epstein 29 (30)
Bando 29 (31)
Howard 37 (48)
Blair 25 (26)
F Robby 27 (32)
Perez 22 (37)
Colbert 27 (24)
Powell 34 (37)
Yaz 29 (40)
Reese 17 (16)
Oliver 21 (17)
Freehan 15 (16)
Torre 13 (18)
Santo 18 (29)
Oliva 14 (24)
Northrup 17 (25)
Smith 12 (25)
Clemente 18 (19)
Tolan 14 (21)
Davis 10 (11)
Melton 18 (23)
Rose 18 (16)
Bench 23 (26)
Jones 9 (12)
While there are some anomalous numbers (Aaron, Perez, Yaz), the majority are quite accurate: no 75 homers for McCovey, no 53 for Powell, no 62 for Howard.
I have concluded that SOM has devolved into a very (statistically) inaccurate game, and that the main culprit is ballpark homers (although I suspect that lefty-righty cards play a role as well). What can be done? Nothing… except perhaps playing all our leagues in Municipal, which I don’t think anyone is going to do.
This little study is far from being scientific or complete, but it was fun.
John E.
Over the years, SOM added increasing complexity—the company said “realism”. But as I played the game, I found it less and less statistically accurate with each addition. Suddenly, lefty-righty balances could be exploited; one could stack a lineup with righties against a righty who did poorly against righties, even though teams used predominately left-handed hitters against him in real life. Soon, ballpark homers and ballpark singles were born—an idea that sounded good to me until I learned that a ballpark’s influence on homers could vary from year to year (without any changes to the dimension of the field)!
Suddenly, players produced ridiculous numbers in replays, taking away the amazement of watching SOM stats parallel those of real life.
And then I found SOM on-line! I was EXCITED! I was looking forward to statistically accurate seasons with my favorite players of all time.
Now, after completing 11 seasons with the 1969 teams, I have some observations/reservations:
In 1969, the NL hit 1470 home runs and the AL hit 1649 (an average of 1560). One would expect TSN/SOM on-line stats to be similar for 12 team leagues. Instead, here are the league home run totals for the 11 leagues I have participated in (The percentages following the HR totals are the percent over the actual average of the 1969 AL/NL):
1949 25%
2301 48%
2281 46%
2229 43%
2391 53%
2514 61%
2275 DH 46%
2505 100M 61%
2202 DH 41%
2586 66%
1545 -1%
All of these seasons have been 80M leagues, except the one indicated. Only two of them have been DH leagues.
What accounts for this absurd inaccuracy in home run totals in what is supposed to be the gaming world’s most accurate game? Could it be that everyone is playing in Fulton? Not so. I checked the leagues and there is some tendency for managers to select HR ballparks, but not enough to warrant this inaccuracy. In fact, three of the leagues above had unique stadium rules, allowing only one Fulton, one Wrigley, etc. into the league.
And so, I suggested an experiment: put everyone in the league into Municipal Stadium (HR 1-3 vs. LH and RH). The league was conducted through www.somdraft.com (a live draft engine). Lo and behold, something amazing happened: hitting statistics were accurate. In fact, the final league listed above (1545 HRs) is this league in which everyone played in this “extreme pitcher’s park”.
Want more accuracy? The team that led the league in runs (802) nearly matched both the AL leader (790) and the NL leader (798). More importantly, individual batters performed statistically like they did in real life. Here is a list of some of the leading HR hitters from the Municipal Stadium League. I left out certain hitters because they were traded (Petrocelli, Killebrew) or because they were used far more than in real life (Held, McAuliffe).
Player Homers in MSL (Homers in real life)
McCovey 44 (45)
May 33 (38)
Wynn 32 (33)
Jackson 44 (47)
Stargell 34 (29)
Allen 35 (32)
Aaron 26 (44)
Bonds 37 (32)
Epstein 29 (30)
Bando 29 (31)
Howard 37 (48)
Blair 25 (26)
F Robby 27 (32)
Perez 22 (37)
Colbert 27 (24)
Powell 34 (37)
Yaz 29 (40)
Reese 17 (16)
Oliver 21 (17)
Freehan 15 (16)
Torre 13 (18)
Santo 18 (29)
Oliva 14 (24)
Northrup 17 (25)
Smith 12 (25)
Clemente 18 (19)
Tolan 14 (21)
Davis 10 (11)
Melton 18 (23)
Rose 18 (16)
Bench 23 (26)
Jones 9 (12)
While there are some anomalous numbers (Aaron, Perez, Yaz), the majority are quite accurate: no 75 homers for McCovey, no 53 for Powell, no 62 for Howard.
I have concluded that SOM has devolved into a very (statistically) inaccurate game, and that the main culprit is ballpark homers (although I suspect that lefty-righty cards play a role as well). What can be done? Nothing… except perhaps playing all our leagues in Municipal, which I don’t think anyone is going to do.
This little study is far from being scientific or complete, but it was fun.
John E.