Bautista at Second in Lineup?

but I think my point comes across.

it’s really conjecture at this point. While it’s certainly possible that dWAR overestimates JPA’s defensive abilities (I happen to think that, too), it’s also possible that the eye test underestimates JPA’s defensive contributions (though I think less likely than the former scenario). For example, it may be that the bad things JPA does don’t really matter all that much in the grand scheme of things, or that all catchers do a similar amount of bad things and JPA isn’t really very bad after all. Again, probably not likely, but we don’t really know.

Until we come up with better metrics, it’s really all conjecture. One may have a basis for their conjecture, but it’s still conjecture. And since this thread is largely trying to argue against conjecture-without-evidence, we should probably avoid stuff like that in this forum

your idea is an interesting one. I am trying to point out that you have independently come up with an idea (ideal lineup construction) that has been thought of before. Indeed, it has been thought of enough times that someone decided to test all the competing theories about lineup construction and gather some data to show which one is likely correct. They did that, and the data showed that your method is not the optimal method.

Looking at GIDP% comes into play when comparing similarly skilled batters, not when looking at whether a low-OBA guy (JPA) should be batting higher in the order in place of a higher OBA guy (say, EE or Lawrie, to use the original post as an example).

Your idea is a valid one that is worthy of further exploration. Lucky for us, that exploration has already been performed. Unfortunately, your idea does not really hold up to that scrutiny.

the difference between a totally optimized lineup and a commonly-used lineup (e.g. the one the Jays used last year, or the one proposed in the OP) has been estimated at 5 to 15 runs over a 162 game schedule. So, if my lineup played your lineup 162 times, my lineup would be expected to go 82-80. It may not happen every time we played a 162-game series, but it’s the expected outcome and would happen more often than not if we kept repeating the sample.

That’s not a huge difference at all, but it’s a quite easy one to achieve since you don’t have to go about acquiring new players to earn that extra win.

The essence of statistics is to determine the most likely outcome from a sample of data. This outcome won’t actually happen every time in real life, but decisions like lineup construction need to be made a priori, not in hindsight, so the best way to make such a decision is by looking at what the evidence shows is the best predictable outcome and going with that.

do you believe in results of studies produced in medical journals? Because studies that looked into treatment benefits from medical interventions (drugs, procedures, etc) use the same statistical manipulations to demonstrate their findings as many of the studies looking into baseball-related questions.

For a more relevant example to lineup construction, do you accept that astronomy is a valid science with valid results? Because many astronomical models are based on Monte Carlo studies – where starting and operating parameters are input into a simulation system and run thousands of times to determine the likely outcomes.

I don’t mean to come off as sounding gruff or condescending, I mean more to demonstrate that baseball statistical methods are not fundamentally different than statistics used in other areas that people would never really question. So if you accept that stats is a valid contribution to medical/scientific inquiry, you pretty much have to accept that it’s a valid contribution to baseball inquiry too (at least if you wish to remain consitent).