motnahp
Posts: 1837
Joined: 8/22/2005 Status: offline
|
I went to my end of 1979 NL season backup file and ran a test assn, after changing all teams' Hooks to "Extremely Patient". I did not make any of the roster moves during the "test 1980" season, so the players aren't EXACTLY where they were, but I was able to compare it to my just-completed 1980 season, in which I made all the player moves and allowed the AI to make roster adjustments at least twice monthly. Here's the data: Total assn. complete games: With hook set to patient: 79 With hook extremely patient: 44 Pitchers with 200+ IP Patient: 28 Extremely Patient: 16 I know this seems backward, but that's what I got. One weird (and unacceptable) thing that happened was that in my "extremely patient" test assn., RELIEF PITCHERS ended up with 7 of the top 11 spots in victories! In my "patient" assn., there was only one reliever in the top 10. Based on this completely un-scientific study, I am baffled. I am tempted to do an assn. without the closer option just to compare. I'll need time to set up all the ballparks, so this won't be done tonight. I do have one possible solution, and it's one I've mentioned a few times on this board. No one else has chimed in that they agree with me, but I'll repeat it again in this thread. I'll keep fishing and see if I get any bites. I KNOW the bait is good and tasty. I feel that a good addition to this program (optional, of course) would be for the AI to recognize batters' real-life AB or plate appearances and pitchers' IP for a particular season. Once a batter has reached his real-life # of AB or PA, or a pitcher has reached his real-life # of IP, that player should suffer a drastic ratings hit and/or have an increased tendency to tire/become injured. I'll use Houston's J. R. Richard as an example. Richard suffered a stroke in mid-1980 and never pitched again. Someone using Richard in the 1980 season should not (IMO) be able to have him available for the entire year. In my scenario, he would have a limited number of innings available that season, and once he reached that limit, his ratings would decline sharply and his propensity to become injured would increase sharply. This accounting was used in another game I own, Lance Haffner Baseball. The user was required to enter player percentages of a team's total for the season. Entering this was a very tedious process, but the data, once entered, gave the game's AI something to draw from when deciding whom to pitch, play, or pinch-hit with. I never stuck with that program long enough to see how well it worked, as the game itself was entirely text-based and got a little boring. I just know that the idea of it is well-intentioned and would add accuracy to season replays from any season. Any comments on this would be appreciated.
|