Help "reading" the cards for new mystery game managers

HAL does know the card that's being used and probably would act appropriately but you need to know if the pitcher has been really tough on left hand hitters, were there runners in scoring position with 2 outs, the park that the game is played in etc. to make a decision on the cards. In Royals or the Astrodome with two out and runners in scoring position HAL might pinch hit with Garcia's 81 card vs. 3 of Oliva's cards especially if the left-handed pitcher is tougher vs. left-handed hitters. Maybe even 1971 but I doubt it because of the natural HR mark.

Here's an extreme example: I'm in an 80s Mystery League. With Mitch Williams (L) on the mound, HAL pinch-hits for Ken Singleton (S) with Pat Sheridan (L).

Now, Sheridan is my back-of-the-bench 5th outfielder, and absolutely putrid against left-handed pitching on 4 of his 5 cards. He's so-so against lefties on the 5th. Singleton I'm sure needs no introduction, but he is somewhat weak against lefties on two of his five cards.

Mitch, of course, is tougher against lefties than righties, and Sheridan is a lefty. Singleton, as a switch-hitter, obviously would have faced him as a righty, but would have used the vs. L half of his card.

From this result, it seems like I can confirm both that I have the only good Sheridan card and one of the two bad Singleton cards. The only other conclusion is that HAL is really bad at this game. Thoughts?

If Sheridan is hitting .254 against left handers and .209 against right handers Hal is less inclined to pinch hit for Sheridan against Williams no matter which mystery year is assigned to Sheridan. I say this because I've seen HAL pinch hit for Ripken in the finals even though Ripken had a spectacular regular season, hitting .300 against both sides. The problem is that he was hitting .167 in the postseason and HAL brought in a pinch hitter and seemingly didn't care which mystery year was assigned. I'm not saying HAL will always ignore the mystery year when making decisions. I don't know. I am saying the decision logic is not as simple as it might seem and there may be all sorts of black box goings-ons that can't easily be explained by one assumption.