[joefromchicago]

THE SIMULATION (SIM2)

As mentioned before, the OOTP pitcher settings for 1922 are five-man rotations and nine-man staffs. Maybe these could actually yield more realistic results. To test that hypothesis, I again ran the Tigers. I still ranked the starters according to the number of starts they had historically, which gave me the following staff:

Herman Pillette (SP)
Howard Ehmke (SP)
Red Oldham (SP)
Hooks Dauss (SP)
Ole Olsen (SP)
Syl Johnson (RP)
Bert Cole (RP)
Ken Holloway (RP)
Carl Holling (RP)

In this sim, Lil Stoner was the odd man out. He didn't make the opening-day roster, for reasons that I can't remember (probably because he stinks), although he did get called up to the bigs twice because of injuries to Holloway and Holling. In real life, Holling and Holloway combined for 10.1 innings pitched. Once again, I used "set it and forget it" for the starting rotation, and once again none of my starters suffered any serious injuries. As a result, the stats for this sim (sim2), and their comparison with the previous sim (sim1) and real life, were as follows:

Hermann Pilette (sim2): 40 G, 38 GS, 316.1 IP
Hermann Pilette (sim1): 38 G, 32 GS, 256.2 IP
Hermann Pilette (RL): 40 G, 37 GS, 274.2 IP

Howard Ehmke (sim2): 41 G, 33 GS, 292 IP
Howard Ehmke (sim1): 35 G, 31 GS, 254 IP
Howard Ehmke (RL): 45 G, 29 GS, 279.2 IP

Red Oldham (sim2): 39 G, 31 GS, 253 IP
Red Oldham (sim1): 34 G, 29 GS, 226 IP
Red Oldham (RL): 43 G, 28 GS, 212 IP

Hooks Dauss (sim2): 33 G, 29 GS, 242 IP
Hooks Dauss (sim1): 38 G, 22 GS, 206 IP
Hooks Dauss (RL): 39 G, 25 GS, 218.2 IP

Ole Olsen (sim2): 27 G, 23 GS, 157 IP
Ole Olsen (sim1): 27 G, 26 GS, 184 IP
Ole Olsen (RL): 37 G, 15 GS, 137 IP

Syl Johnson (sim2): 5 G, 0 GS, 7 IP
Syl Johnson (sim1): 15 G, 14 GS, 87 IP
Syl Johnson (RL): 29 G, 8 GS, 97 IP

Bert Cole (sim2): 50 G, 0 GS, 77.1 IP
Bert Cole (sim1): 42 G, 0 GS, 74.1 IP
Bert Cole (RL): 23 G, 5 GS, 79.1 IP

Lil Stoner (sim2): 11 G, 0 GS, 25.2 IP
Lil Stoner (sim1): 46 G, 0 GS, 83.1 IP
Lil Stoner (RL): 17 G, 7 GS, 62.2 IP

Carl Holling (sim2): 12 G, 0 GS, 15.2 IP
Carl Holling (sim1): did not play
Carl Holling (RL): 5 G, 1 GS, 9.1 IP

Ken Holloway (sim2): 6 G, 0 GS, 9.2 IP
Ken Holloway (sim1): did not play
Ken Holloway (RL): 1 G, 0 GS, 1 IP

Were the results for sim2 any more realistic than those from sim1? Well, it could be said that the top three pitchers compiled too many innings in sim2, but then it could also be said that they compiled too few innings in sim1. I was actually surprised by the number of relief appearances made by the top starters in sim2. With four relievers in the bullpen, I would have thought that the AI would use the starters in relief only sparingly. That, however, wasn't the case. Oldham and Olsen, for instance, were used more often in relief in sim2 than in sim1.

The starts, of course, were not as widely distributed in sim2 as they were in sim1. The biggest loser in this regard was Johnson, who went from 14 GS and 83.1 IP in sim1 to zero GS and 7 IP in sim2. That sort of result should be expected. In the game, Johnson isn't really good enough to be either a regular starter or a regular reliever. With a six-man rotation, he gets some IP just because he's the sixth man in the rotation. With a five-man rotation, however, he doesn't really have a role to play. His innings, then, largely went to the top four men in the rotation, who collectively accumulated 17 more starts in sim2 than they did in sim1.

The five starters also accumulated more complete games in sim2 (82) than the six starters did in sim1 (72) and more than the 1922 Tigers did in real life (67) - and this despite having four relievers in the bullpen. That may have had something to do with the strategy pre-sets, which I didn't change from the out-of-the-box settings (neutral for everything). Would there have been fewer CGs in sim2 if the manager was Pat Moran, as was the case in sim1? I don't know, but it seems odd that the AI would actually use relievers more often by employing the strategy of a 1920s manager. Something's off here, and I'm not sure what it is.

One other stat stands out. Bert Cole dominated the bullpen in sim2. In sim1, Cole shared bullpen duties with Stoner, but in sim2 Cole had more IP than the other four relievers put together. Cole finished 32 games and saved 8. Not only were these numbers unrealistic, but they indicate that the AI was using Cole either as a stopper or a closer, roles that didn't really exist in 1922. So that's unrealistic on a couple of levels. Indeed, despite the fact that the game setting for "use of closers" is "very rarely," some relievers on other teams (though not on Detroit) were designated as closers in sim2. That just shouldn't be happening.

CONTINUED