Hmmm....that's a tough question to put a definite number on, but let's try, if only for the Ramirez-Lee situation.
Now, Ramirez and Lee both have a vast majority of their plate appearances coming from the 3rd and 4th spots in the lineup, respectively, so let's assume (both for the sake of argument and for the sake of my sanity, because I really don't want to go through and account for days they hit fifth for some reason or other and the like) that all of their respective PA's have come from the 3rd and 4th spots in the lineup. Let's also assume that Lou "should have" realized Lee and Ramirez were struggling mightily early on in the season, pretty much 10 games into the season for both hitters (that's around 80 PA's total that we can't count against Lou, and we'll round it up to 100 to account for other variable "stuff").
Now, what "should" have Lou done? Well, it's not as simple as benching them in favor of a better hitter, and really we can't fault Lou for not benching two of the biggest stars on the squad, so let's say he should have moved them down to the 7 and 8 spots in the lineup to work out their problems. And last year in the NL, the 3-4 spots in the lineup received about 10% more PA's than the 7-8 spots did
Lee and Ramirez have a total of 412 PA's so far this season, and have a combined VORPr (Value Over Replacement Player tuned to per game rate) of -.363, meaning they are costing the team about .36 runs per game they play.
This is where things get tricky and the math gets fudgy: I have to turn VORPr from a per game rate to a per plate appearance rate, and I have no idea how many mathematical gods I am offending by doing this, but stick with me, because the premise is more important than the actual result.
Combined, Lee and Ramirez have appeared in 95 games, across which they accumulated those 412 PA's (and their VORPr), that's around 8.66 PA's per game (had to double the per game PA rate since we are looking at two players) for that -.363 VORPr. Applying that to VORPr, we get that Lee and Ramirez are costing the team, on average, .0419 runs per plate appearance (or .029 runs per PA individually).
Okay, so accounting for the 100 PA's we can't reasonably punish Lou for, that leaves us with 312 PA's garnered by Lee and Ramirez in the 3-4 spots that we can punish Lou for. Assuming that Lee and Ramirez would have seen 10% fewer PA's by this time had they been moved down in the line up past the artificial point we set earlier, this means that the Lee-Ramirez combo has seen roughly 31 PA's it shouldn't have, or at least wouldn't have had the combo been moved to the 7-8 spots, and those 31 PA's (for the duo as a whole) are worth about -1.3 runs.
So really, they wouldn't be any noticeably better thus far in the season had Lou (or another manager inserted the moment Lou decided to stick with the Lee-Ramirez combo past 100 PAs) moved them down in the lineup. However, across an entire season, that amount of run-costing can easily add-up to actual wins in the standings (probably between 1 and 2 marginal wins).
Though not much, that's just a small area of the game in which Lou has conceivably cost this team runs. Then we have to consider his other lineup decisions, bullpen usage, 1-run strategy uses, etc.
