(disclaimer: this is written from the perspective of H2H leagues. I assume the strategy for Roto or Points leagues would be different)
Like many of you, I'm not content to only play one fantasy league and instead find myself with quite a few this year. Inevitably, you'll have a wide range of players but also some overlap. You'll be inclined to pick up those talented FA in multiple leagues, so you'll have three teams with Lastings Milledge. Or you'll target a specific player after a certain round, so you'll have Dustin McGowan in all your leagues because he dropped to the 18th or something.
So here's the thing. On any given day, looking at all your leagues, you'll have something like this: I have Jake Peavy on two teams, and somewhere on my teams I have a combination of Holliday, Hawpe, Atkins, Tulo, Taveras, and Helton. Peavy's pitching against the Rockies tonight. So what's your strategy? Do you bench Peavy and hope he gets shelled, and start all your Rockies? Do you start Peavy but bench those Rockies who have bad career numbers against Peavy? Or do you start everyone, knowing that it's very likely that either Peavy will do well, or one of those Rockies will?
To put it another way, when you have players in different leagues (or even on the same team) facing each other, do you try and predict what will happen and bench accordingly, or start everyone and hope the good outweighs the bad?
Looking at the above example, here's what could happen:
If I try to predict what happens (Peavy does great / gets shelled) and I'm right, I win big time. If Peavy has a complete game shutout with 12 K's I get those stats, and I avoid all the 0/4's from my Rockies hitters. If I guess the other way and I'm right, it's also a huge victory. I bench Peavy and avoid his 8 ER in 5 innings, and get lots of juicy batting stats from my rockies hitters.
On the other hand, if I try to predict what happens and I'm WRONG, I get absolutely hosed. I either miss out on great hitting stats and get stuck with Peavy's horrid stats, or I lose Peavy's stats and get stuck with a bunch of 0/4's. To recap, I either win huge or lose huge (there's also the disturbing possibility of everyone doing so-so and busting middle, but that's less likely than one of the above alternatives)
Let's see what could happen if I start everyone and hope for the best. Well, the absolute best case is that Peavy is so-so and the Padres defense makes a bunch of errors leading to no ER. Peavy puts up so-so ratios and a few K's, and doesn't get the win. He goes five innings, and the Rockies tag the Padres bullpen for about 8 runs. The absolute worst case is that Peavy gets shelled, but the damage is done by the few Rockies players you don't have in one of your leagues, leading to an only somewhat better version of the bad-outcome version of the above "prediction" method.
What happens more often than those extreme cases is you get a mixed bag. This situation isn't quite a zero sum game, because in theory a string of errors can lead to your hitters giving you some counting stats in the form of SB and Runs without negatively affecting your pitcher, but it might as well be: for all practical purposes, if your hitters do well your pitcher's numbers will suffer - and vice versa. You're likely to have either your BA or your pitching ratios take a hit that day, but most of the time you'll earn enough counting stats to make up for it.
Let's dig a little deeper. Starting everyone can't negatively affect your counting stats. Even if all your hitters go 0/4 that day, the worst you've done is put yourself in a position to lose AVG this week. You haven't helped yourself in counting stats, but you haven't hurt yourself either. And on top of that, your pitching ratios, and most likely wins and strikeouts were significantly improved. Likewise, a horrible start by your pitcher only harms your pitching ratios, with 4/5 hitting categories getting an almost certain bump. This is complicated by the fact that usually these players will be spread across different leagues, spreading around the good or bad stats for the hitters and concentrating them for the pitcher. That is, if the pitcher gets shelled it's likely a bad outcome for a single league and a good to moderate one for multiple ones.
The point is this: starting pitchers against a lineup containing many players you own can be an interesting and effective way of hedging your bets.
On a related note, consider this: when you start pushing eight teams, you'll inevitably have teams that have the same PAIR of players. Lastings Milledge and Edwin Encarnacion. Hedeki Matsui and Matt Kemp. Hedging Your Bets 2.0: on any given night, do you start Kemp over Matsui in one league, and Matsui over Kemp in another? It's the same basic principle as what I've outlined above. If you start Kemp in both leagues, have the opportunity to win big if Kemp has a big night and Matsui gets the golden sombrero. Here's the kicker, though - this is what makes this situation here different than the one above. Since the players aren't in direct competition with each other, they could possibly BOTH have great nights, which gives you one more scenario where you win big.
Let's break it down: (again, I realize I'm making it simpler than it actually is, since players don't simply have "good" and "bad" nights; it's a spectrum rather than just two options) If you start only Kemp, you have one scenario that's great: Kemp has a huge night. If you start both, you have a situation where them both having a huge night is a win in both leagues, but either of them having a huge night is a win in AT LEAST ONE league. Again, I see hedging your bets when it comes to starting players as a viable strategy.
I'd like to hear people's thoughts on the matter.