High cap leagues roster construction

I've been doing pretty well in Leagues of $140m or less, but once I get to about $200m I start getting crushed. The small sample size I'm working with imply these lessons:

1. The pitching and defense strategy doesn't work in high cap leagues.

2. Going for home field advantage with heavily lefty- or righty-laden teams works better, but still doesn't work.

3. Having most players in the lineup platoon makes sense from a maximize-starting-lineup-value perspective, but since Hal doesn't sub them after a pitching change I'm often left hanging with my platoon players facing the wrong kind of pitcher for the last several innings, which kills me.

4. I suck at high cap leagues.

Any feedback on these issues or other factors that affect high cap rosters? Thanks!

I still would like to hear how folks approach high salary cap leagues - any input is appreciated!

I guess you play in ATG.

Never played in ATG 200M cap league, but I did look at last year's results in the ATG Barnstour:

See Event 4 here: http://365.strat-o-matic.com/index.php/tournament/22

Here is my analysis from comparing successful teams vs losing teams
1-With a few exceptions, all teams had spent 200M on their roster. So if differences exist between losing teams and winnigest teams, it's not about what money was spent, but how it was spent.

2-Successful teams seem to have more platoons at work. Most teams with 98 wins or more had at least two working platoons, whereas losing teams were more relying on nine regular starters. On the contrary, losing teams are spending 40M-50M on bench players that are seldom used.

3-Many losing teams had very bad stats coming from the bullpen, so it seems to me that many losing teams don't manage carefully their bullpen.

One golden rule I think is you MUST NOT, NEVER EVER, have a pitcher, whoever he is, on the mound in the fatigue mode in 200M leagues. The way the fatigue mode works, when a pitcher goes below F9, there is additional rolls on the offensive side (instead of getting 50% of rolls on offensive, it shifts towards 52%-54% on the offensive side and up increasing as the F gets lower). And simply put, there is just too much offense on 200M teams to let that happen. Relievers are more vulnerable to get on the fatigued mode than starting pitchers (especially S9 starting pitchers), so I would definitively put them on quick hook. I would also make sure that I have sufficient bullpen so to never have one on the mound that Hal leaves on the fatigued mode because the bullpen is too thin. On this regard, I seriously question the strategy most owners use of going with a roster of 10 pitchers, especially if there is on the roster a few specialists or extra starting pitchers who can't relieve. It seems to me that I would rather have the maximum, 12 pitchers, with a R5 at the back-end in case the game goes in overtime, to make sure I never get caught off guard on the mound.

Another issue is that many owners use their 5th/6th best starting pitcher a S/R as their go-to-reliever, but I don't think this is the best strategy. Your 5th/6th best SP is typically a 8.5M starting pitcher, and usually they have more than 10% on-base on each side, with some power, so they are roughly equivalent to a R3 4.5M relief pitcher, but with more stamina. In other words, when you're using a 8.5M S/R as your main reliever, you are NOT using one of the 30th best reliever in the set. In a 12-team league, you should aim to have, at minimum, one of the 12 best relievers in the set and two of the 24 best relievers in the set. So unless you are lucky enough to have Christy Mathewson as your 5th best *SP, you should rely on, say, Sutter and Nen, to set up the game and close it. As told above, though, they should be on quick hook and removed before their fatigued mode gets on.