The art of lineup construction

[supertyphoon]

I read a somewhat interesting LA Times newspaper article about lineups that I'd like to pass along:

http://www.latimes.com/sports/la-sp-0524-batting-order-20110524,0,2051071.story

http://www.latimes.com/sports/la-sp-0524-batting-order-box-20110524,0,1230223.story

[quote:38509bcfd9]Mattingly ... considers how each batter is going to affect those hitting in front and behind of him. "You're always trying to protect," he said. [/quote:38509bcfd9]

I don't think that real-life consideration is a factor here. At all. It's just dice rolls and probability. Our players aren't "affected" by those hitting in front or behind them. They don't hit differently if there's a speedy runner on first or try to hit a SF if there's a runner on third, nor are they extra patient when hitting in front of the pitcher or leading off an inning.

[quote:38509bcfd9]Scioscia...studies ways to group hitters to maximize their effectiveness. "You look at the two guys in front of a guy and the two guys after him to make sure that lineup just has some continuity and flow," he said. "You try to juggle to keep those holes in your lineup at a minimum so you're hopefully pressuring teams every inning."[/quote:38509bcfd9]

I think he has a very good approach. Two table setters followed by a power hitter, then your two best hitters at 4 and 5 followed by another power hitter at 6. Speed and effective "small ball" hitters at the bottom of the order.

Some thoughts:
I (almost) always put my highest OBP in the first two spots of the lineup, especially if they have negative clutch. That "second cleanup hitter" #6 spot is the perfect place for my best cluth hitter.

In high-cap and live draft leagues where your have great hitters from top to bottom, I think an effective strategy is to bat someone like Ruth, Bonds or Williams leadoff, just to give them as many plate appearances as possible.

Usually my reasoning for who bats 8 or 9 in the order is which player has the greatest injury risk. If I have a top catcher with a 2 injury risk and no suitable backup, I try to limit his exposure with fewer ABs and defensive replacement with a lead.

I've always considered Tony LaRussa one of the better ones at managing a team, which includes a willingness to think outside the box when it comes to lineups. I grew up in SoCal in the 60s, and I have to say that Walt Alston had probably the least effective lineups I've ever seen. Luckily they didn't have to score many runs with that pitching staff and ballpark. Earl Weaver was a genius at using his roster and getting the most possible out of his lineup. He gets dismissed nowadays "waiting for a 3-run homer" to score runs, but it worked for him, and he was way ahead of his time in platooning to maximze his team's OBP and SLG.

[PotKettleBlack]

* runners get held, decreasing the range of the defender holding them, increasing (slightly) hit chances. So, fast guys, or at least running threats DO affect the guys batting behind them.

Similarly, relative clutch seems to be a major concern for HAL in making the IBB decision. So, having a clutch batter 5th with a negative clutch guy sixth will generally get the clutch guy walked and the 6th guy grounding out. So, there is some protection, albeit small.

Here's the rub. When you work Markov-chain based simulations or if you use historical data on the value of events, the tweaking of lineups from optimal to worst is generally pretty small, holding to the same bats in the lineup. The baseball musings site brings it home, but with MLB style rosters (rather than ATG), it's even tighter. About 10-20 runs a season, or 1-2 wins. Much is made of lineups, but in the real world, it's pretty small optimizations.

PS- Earl Weaver would routinely have a very lousy bat in the 2 hole to sac the leadoff guy over. Hardly genius considering how counterproductive that was, even given the cultural constraints of the time. I'm a TLR guy. Like what TLR tries to do... really, he should commit to the pitcher batting 8th full time.

[macnole]

I agree on the lineups in SOMO having limited spread in performance. However, in simulations, whether Monte Carlo, Markov, or historical based--they are seeded by and bounded by the same overall thing: the results that already occurred. Based on those results, independent player probabilities are used to simulate play with the events in any order, weighted by what actually happened.

In SOMO--this is a sound approach as it's already a statistically based replay.

The limitation in real life, however, is there is no way to simulate the different pitches or approaches a hitter would get based on who would have been hitting in front or behind him--the human factor is missing.

In that sense, lineups can play a significant non-linear role in outcomes.

But for SOMO; absolutely agree that's an assertion based in fact--and I've let HAL manage lineups for half the season, only to find the team performed about as well as it did in the first half.

[paige91]

I don't know much about the "best" way to have your lineup, but it seems like OBP is extremely important, especially at $100 million and below. I have a team that just finished the season first in runs scored but only 56 HR playing in Forbes 65. The team was poor defensively, but that didn't seem to matter as much as a very high team OBP.

[url=http://fantasygames.sportingnews.com/stratomatic/team/team_other.html?user_id=306759]Oklahoma Outlaws[/url]

My lineup against LHP was
Hamilton
McGraw
Appling
Brouthers
J Alou
Clines
Doyle
Youngs
Cochrane

The lineup against RHP was
Hamilton
McGraw
Mazzilli
Brouthers
Youngs
Weintraub
Appling
Doyle
Cochrane

[sdajr76]

i have had many small ball teams finish at or near the bottom in home runs and win championships.

http://fantasygames.sportingnews.com/stratomatic/team/team_other.html?user_id=54699

http://fantasygames.sportingnews.com/stratomatic/team/team_other.html?user_id=266363

-steven