HAL AI

[AdamKatz]

I have read on these boards that HAL knows what card your players have. Does this mean, for example, that if Niedenfeur has his amazing car, that the computer will put him in all the time over other relievers that are usually better and more expensive?

Similarly, does this hold true with hitters? I have Dykstra and G.Young. If Dykstra has his card that can't hit lefties and Young has his one good year, will the computer pinch hit for Dykstra often? or will the computer never ph a 1mill guy for a 5 mill guy unless you set those directions on the individual cards?

[Paul5757]

The answer is "maybe."

Basically, I wouldn't make any moves based on what HAL does once or twice. HAL might not consider a given situation important enough to use a PH. You also have to consider what pitcher is being used--for example, HAL may not use a PH for Dykstra vs. a "reverse lefty." Or HAL might be having a bad day.

Instead, look for patterns of usage. In my opinion, this is a better indicator than replay statistics. For example, if HAL continually uses Niedenfuer in nearly all save situations, it means you either have one of Niedenfuer's better years (despite possible bad stats) or the rest of your bp really sucks (and it fairly unlikely you'd draw 4 other bad cards).

[bjs73]

You need to usurp the control away from HAL as much as possible in my opinion. There are many who disagree with me though.

For instance, you could have a fairly decent W. Hernandez (maybe not his lights out year) but at least one of his heavy LH killer cards. If you don't set him up the way he's supposed to be used (a lefty specialist) then you will be misled into thinking that you've got his worst card when looking at his stats in a small sample size.

I guarantee that HAL will be glad to use W. Hernandez against both righties and lefties unless you direct him not to do so. Others say they like to let HAL lead them to their best reliever. I used to do that in my early days too. However, I've changed all that in search of balance and matchups. If you've got a crew of guys with both strengths [i:ea53c1aac9]and[/i:ea53c1aac9] weaknesses in the pen, you can work with that by setting their bullpen tick marks correctly.

That being said, Tom N. is mostly a righty specialist. He actually has a serviceable card in 2 or 3 years because of it. If you get his lights out card you'll know fairly quickly anyway. If you let him take on lefties, you might be misled into thinking you've got his worst year when you don't.

As far as hitters go, HAL is an idiot there also. HAL tries to use the matchups in determining who to play. He'll start a 3L hitter over a 2R hitter when making lineups occasionally but that kind of information isn't always correct either.

I've used the "let HAL decide" method on my position players when I have a stud player that is doing horribly in the first 39 games. When I do this, I usually pick up a suitable replacement player, take my "stud" out of the lineup for a series, and see which player HAL puts in the games that night.

I had the bad Mike Schmidt one time and I went ahead and picked up Gaetti at the first cut line. HAL put Gaetti in the lineup all 3 games that night. In my opinion, it confirmed what I already knew was true.

[Outta Leftfield]

Maybe is right. Exactly what HAL does and doesn't know is a topic of much debate. I've sometimes dropped a hitter fairly early because HAL kept pinch hitting for him in situations where, with a good card, he should have stayed in. I'm not sure these decisions were right or not--usually, though, I had some other evidence to indicate that the player was not in a good year. The pinch hitting might have pushed me to pull the trigger a little sooner, though.

On the bullpen, I still tend to let HAL manage the bullpen for the first few games. I'm not sure whether he's making knowlegeable choices or merely working at random, but the reason I do it so that--even if it's random--I begin to get info on my relievers from the very start. It could be that if I set up all the relievers roles from the beginning, I might overlook some guy who had a dominant card. HAL spread the work around enough so that I begin to get an idea pretty quickly. Also, I tend to use cheap relievers, so I don't begin the year with an obvious ace.

BUT, once the data starts to come in and I start getting an idea of what each RP's capabilities are, I start to use the manager and card settings, and as the year progresses I've pretty much set everyone's card in some way and have picked out a closer and/or setup man. The L/R settings are particularly useful. If a guy has a platoon weakness, you can really guide HAL to use him only in situations where he's likely to succeed. Sometimes I've got an RP in a bad year and none of the FA options look appealling so I just fence him off. Set him, say, to face lefties only, mop up only, and quick hook. With settings like that, the guy might go 30 games without every pitching.

Just to show what HAL will do when unchecked--I have T. Burgmeier in my bullpen. He's a lefty killer but usually has trouble with righties. HAL treated him as the go to guy, pitching him soley against RH, and he got hammered in several early games. After 15 AB he'd given up 4 ER and 7 hits while facing 14 RH and only one LH. Once I figured out what HAL was doing, I set Burgmeier to "Avoid right-handed batters," and since then he's been much more effective. Still, RH are hitting .375 against him and that wouldn't stop HAL from throwing him at another RH without my settings.

[rookssa1958]

You can' trust HAL too much, you have to use your common sense. Some examples, I have a team that is a stinker, but I do not have many options because of league rules. So I set only my lead off hitter and let Mr. HAL set the rest. He made some moves which seemed strange to me, playing Righties vs RH pitching over Lefties I had available. But I set my lineup like he did and wen ton a modest winning streak.

In another league, I had Mr. Eckersley who finished the regular season with 53 saves. I can't remember if I set him as closer or not, as I like to be able to stop a rally early in the game if needed....
Anyway, in game 7 of the world series, with 2 on and 2 out, HAL brings in Chris Welsh (a 75 cent LH who I had marked as mop up, avoid LH) to face Kirk Gibson........ anyway, Welsh got the out, but it is not a move a sane manager would have made, and shows that HAL will not identify your stud reliever even if you give him the chance......

[Paul5757]

Take a look at Welsh's '85 card vs. LHP (ironically, his worst year WHIP-wise). Welsh has a 1-7 triple on a 4-4 roll. Otherwise, just a couple of singles here and there. Considering Eck's propensity to give up HR vs. LH hitters in most years, the move actually makes a little bit of sense.

[rookssa1958]

Did I forget to mention that over spot starts, Welsh was 6 - 1 for me with lefties hitting over .350 and righties hitting under .240?

I only spotted him when I thought a reverse lefty would thrive??????

[Outta Leftfield]

[quote:0f9b992122]Did I forget to mention that over spot starts, Welsh was 6 - 1 for me with lefties hitting over .350 and righties hitting under .240?[/quote:0f9b992122]

Wow! This suggests that HAL really doesn't have a clue. How many IP did Welsh have?

[Paul5757]

How many BF for both Eckersley and Welsh? Did you know which card Gibson was?

[rookssa1958]

Playoffs W L S IP H ER BB K ERA WHIP
Welsh, C. 0 0 1 .1 0 0 0 0 .00 .00
Eckersley, D. 2 0 2 11.1 6 1 1 14 .79 .62

And remember for the playoffs, I stuck Welsh in the rotation so HAL would not use him, but released him for game 7, just in case and set him to avoid LH?

Reg Season W L S IP H ER BB K ERA WHIP
Welsh, C. 6 1 0 60.2 58 28 32 16 4.15 1.48
Eckersley, D. 5 3 53 122.2 88 19 23 125 1.39 .91

Royals Stadium for all games.

Gibby
AB R 2B HR RBI BB K E BA SLG OBP
276 51 6 18 43 35 11 6 .236 .475 .329